Arxiv:0805.2201V2 [Astro-Ph] 17 May 2008 A97031;Mojzatobree.D

arXiv:0805.2201v2 [astro-ph] 17 May 2008 A97031;mojzatobree.d . [email protected] 94720-3411; CA tet abig,M,02138. MA, Cambridge, Street, at abr,C 93117. CA Barbara, Santa rd,Budr O80309-0389. CO Boulder, orado, LEFIT) po 8 -89 ilnead aCanada la de Villanueva E-28691 78, Spain. (Madrid), Apdo. (LAEFF-INTA), 80 aLgn,Tnrf,Spain. Tenerife, Laguna, La 38205 RMSOKBEKU OPA N EOD XESV PANCHROMA EXTENSIVE BEYOND: AND PEAK TO BREAKOUT SHOCK FROM H SHRCLTP BSPROA20DASCAE IHSWIF WITH ASSOCIATED 2008D SUPERNOVA IB TYPE ASPHERICAL THE 2 4 3 5 6 7 8 1 rpittpstuigL using typeset Preprint ApJ to Submitted .Liebert J. .Modjaz M. .Dufour P. .P Kirshner P. R. .BraoyNavascues y Barrado D. avr-mtsna etrfrAtohsc,6 Garden 60 Astrophysics, for Center Harvard-Smithsonian a ube lblTlsoeNtok 70CroaDr., Cortona 6740 Network, Telescope Global Cumbres Las Fellow. Research Sloan etrfrAtohsc pc srnm,Uiest fCo of University Astronomy, Space & Astrophysics for Center aoaoi eAtosc sailyFsc Fundamental Fisica y Espacial Astrofisica de Laboratorio nttt eAtosc eCnra,C i ataSN E- S/N, Lactea Via C/ Canarias, de Astrofisica de Instituto eateto srnm,Uiest fCiao 60South 5640 Chicago, of University Astronomy, of Department eateto srnm,Uiest fClfri,Berkel California, of University Astronomy, of Department -a rninsi rdcal osqec fftr sen future rest-fram of and consequence predictable thresholds a detector though is in ratio, transients duratio fold flux X-ray long X-ray carefully a to as to 080109 optical such XRT need high properties, cores. very Ib observed and SN of luminosity, of number terms depa a in with April of ejecta flashes associate 2008 the we on for that seen obtained been lines Spectra oxygen stars. double-peaked WN show early-type of those hc with which lo h ealdsuyo h al tgsi uenv exp supernova in stages headings: early Subject fashion. the routine a of in study properties, detailed the allow yr tipdevlp N n oetmt h tla radius stellar measure the reliable estimate a to fo construct and constraints to SN, powerful us provide stripped-envelope allow Ib, data SN any comprehensive for reported ones nrrd( infrared h hrceitcsaeo hc raot(∆ breakout shock of shape characteristic the olwdby followed agtsoty(∆ shortly caught − . n . or eoeteosto R 819adorearly- our and 080109 XRT of onset the stripped-envelop before regular hours in 4.5 sup than in and component faster 2.8 breakout subside shock to the that appears find we sample, small ogrs ie(8dy)adamds pia ekluminosity peak optical c modest showing a (i.e., and Ib days) SN (18 pre We normal time differ. spectroscopically r rise that a time long aspects is a key 2008D dr span several we SN find data throughout and Our progenitor; anal literature its detailed the and 2770. and supernova transient o NGC the X-ray of galaxy, data the nature same of X-ray detection the as the in in after well discovered 2007uy as serendipitously Ib 2008D was SN which (SN) 080109 supernova (XRT) on data infrared) .Smith N. 16 epeetetnieerypooerc( photometric early extensive present We − 1 . .Culdwt ai ut-aeeghflo-pobservat follow-up multi-wavelength rapid with Coupled ). 11 3 1 11 ± .H Marion H. G. , .Li W. , .Falco E. , 0 1 . JHK a,atrcretn for correcting after mag, 4 .Starr D. , 2 .Kocevski D. , R 56 A ⋆ T 1 idcy(∆ decay Ni .Butler N. , E 1 = s tl mltajv 10/09/06 v. emulateapj style X eetos hc t08 n .1dy fe ne r h e the are onset after days 0.71 and 0.84 at which detections, ) 2 aais itne n esit uenve eea s — general supernovae: — redshifts and distances galaxies: t . S 08)—glxe:idvda NC2770) (NGC individual galaxies: — 2008D) (SN .Friedman A. , 1 ∼ ± 1 6 .N Steele N. T. , .H Blake H. C. , 14,15 1 a)atrsokbeku eel eldffrne mn th among differences real reveals breakout shock after day) 1 . R 0 .Lee N. , 1 .Poznanski D. , > t 1 .Chornock R. , ⊙ ssalrta yia ai fW tr,admrial con marginally and stars, WN of radii typical than smaller is as.Crflcmaio ihohrsrpe-neoes stripped-envelope other with comparison Careful days). 5 2 1 .Ganeshalingam M. , .S Olivier S. S. , 2 1 E .Bouy H. , .Stockton A. , ( B − 1 .Hicken M. , 1 umte oApJ to Submitted V .Perley D. , B Rr UBV ABSTRACT ) ey, Host t 1,7 l- ≈ .Brown W. , 16 0 = 0 .Olszewski E. , . 17 1 ′ Ii li vne hcg,I 60637. IL Chicago, Avenue, Ellis ai,C 95616. CA Davis, 00Es vne iemr,C 94550. CA Livermore, Avenue, East 7000 hryAe usn Z85721. AZ Tucson, Ave. Cherry iyo or ae or ae N46556. IN Dame, Notre Dame, Notre of sity nvriyo aiona at rz A95064. CA Cruz, Santa California, of University utn X78712. TX Austin, .G Williams G. G. , − . 2 6 1 9 10 11 12 13 14 15 ′ .J Foley J. R. , 1 .Garnavich P. , JHK .Blondin S. , as otpoone ntebu bands) blue the in pronounced most days, 5 nvriyo aiona eateto hsc,1Shields 1 Physics, of Department California, of University ± Swift nttt o epyisadPaeayPyis LL L-41 LLNL, Physics, Planetary and Geophysics for Institute eateto srnm,Uiest fAioa 3 N. 933 Arizona, of University Astronomy, of Department eateto hsc,25NewadSineHl,Univer- Hall, Science Nieuwland 225 Physics, of Department nvriyo aionaOsraois-Lc Observatory Lick - Observatories California of University srnm eatet nvriyo ea tAustin, at Texas of University Department, Astronomy hsc eatet ea tt nvriy-SnMarcos, San - University State Texas Department, Physics R ssalwdu odrv osrit nthe on constraints derive to us allowed ysis iiealsyXrymsin ( missions X-ray all-sky sitive 0 rpris eso htfidn such finding that show We properties. e N u nqeuprlmt obtained limits upper unique Our SN. e ⋆ . wcmaioswt eut rsne in presented results with comparisons aw 2 oeso h raoteiso.Our emission. breakout the of models r ,sfeigXrysetu,lwX-ray low spectrum, X-ray softening n, a) h ih uvso h Nshow SN the of curves light The mag). 1 s etordt htfis salse that established first that data our sent .Challis P. , eto h ooercotu o this for output bolometric the of ment tde noeigterpyia origin physical their uncovering studies fispoal ofRytprogenitor, Wolf-Rayet probable its of trsfo peia ymtya has as symmetry spherical from rtures n pcrsoi otcladnear- and (optical spectroscopic and ) aadrn olwu bevtosof observations follow-up during data oin,a ela fterprogenitor their of as well as losions, npcosH ie)wt relatively a with lines) He onspicuous os hc r rca,teedt will data these crucial, are which ions, 93 otsatrmxmmlight, maximum after months 3 29.3, ( roa eae ogmarybursts gamma-ray to related ernovae M pasdffrn rmcasclX-ray classical from different ppears 11 .X Prochaska X. J. , V h soitdXrytransient X-ray associated the n ieotcl( optical time 1 2 .S Stringfellow S. G. , .S Bloom S. J. , = 18 12 n .M Wood-Vasey M. W. and , − neo or o19days 109 to hours 2 of ange .Holden B. , 2 17 .Chen H. , -A RNIN 080109 TRANSIENT X-RAY T . 0 ± proa:individual upernovae: ris ground-based arliest 0 I BEVTOSOF OBSERVATIONS TIC . Vr BV a and mag 3 1,3,4 8 13 .H eVries de H. W. , 13 m rmour From em. .V Filippenko V. A. , .Illingworth G. ; ′ .M Silverman M. J. , i ′ ∼ n near- and ) itn with sistent upernovae 5 10 .Berlind P. , 2 M − B 10 2 = 3 9,10 13 2 1 1 , Ave, , ; , , 3, , 2

for SN 2007uy in the same galaxy. Spectra and photome- 1. INTRODUCTION try were obtained on a nightly basis (weather permitting) starting 2008 Jan. 10, sometimes with multiple observa- Massive stars (initial M & 8 M⊙) die violently, and the ensuing core-collapse supernovae (SNe) are classified tions per night. In addition, we reduced and analyzed into different spectroscopic SN subtypes. The classifica- Swift/XRT and Swift/UVOT data, as well as Chandra tion depends on the presence or absence of hydrogen and X-ray observations. Furthermore, we obtained optical helium in the SN spectra, and its sequence is set by the spectra at the Apache Point Observatory through the amount and kind of outer envelopes the progenitor re- awarding of Director’s Discretionary Time. tained before explosion: Type II SNe show prominent H, The exact mechanism for removing progressively in- while Type IIb SNe show He and weaker H, Type Ib SNe creasing amounts of the H and He layers in the pro- lack H and conspicuously show He, and Type Ic SNe lack genitors of stripped-envelope SNe is not fully under- both H and He (see Filippenko 1997 for a review). SNe stood. The origin could either lie in strong winds of IIb, Ib, and Ic are collectively called stripped-envelope the very massive progenitor (& 30 M⊙; Woosley et al. SNe (Clocchiatti et al. 1996). 1993), sudden eruptions (Smith & Owocki 2006), inter- While the number of well-observed, normal, stripped- actions with a binary companion star (e.g., Nomoto et al. envelope SNe has grown (see Matheson et al. 2001; 1995; Podsiadlowski et al. 2004), or the interplay of all Richardson et al. 2006; Modjaz 2007 and references the above. Thus, observational clues that could help dis- therein), the pool remains small where detailed multi- tinguish between the different mechanisms or that can wavelength observations shortly after explosion trace the set a limit to the progenitor size are crucial. emergence of the shock from the stellar envelopes. Re- Finally, the recent connection between long-duration cently, the Swift satellite (Gehrels et al. 2004) with its gamma-ray bursts (GRBs) and specifically broad- rapid slewing capabilities, space-based nature, and X- lined SNe Ic (SNe Ic-bl) has piqued interest (see ray and optical/ultraviolet telescopes has opened a new Woosley & Bloom 2006 for a recent review) in stripped- window for SN observations soon after their massive pro- envelope supernovae (SNe) and the attempt to constrain genitor explode. their progenitors. Nevertheless, while a few SNe con- A marvelous opportunity to closely study the evolution nected with GRBs have been fairly well observed, there and shock breakout of a nearby SN is offered by X-ray exist little early-time and high-energy data on the com- transient (XRT) 080109/SN 2008D. During Swift’s rou- parison set of core-collapse SNe. Thus, as the sample tine follow-up observations of SN 2007uy (Nakano et al. size increases, we can hope to use the full set of SN data to uncover trends between the different properties, such 2008; Blondin et al. 2008c) in NGC 2770, XRT 080109 56 was discovered serendipitously (Berger & Soderberg as progenitor radii, ejecta mass, and amount of Ni pro- 2008; Kong & Maccarone 2008; Soderberg et al. 2008) in duced. the same galaxy in data obtained by the Swift X-ray Tele- In §2 we present X-ray data obtained with Swift/XRT scope (Swift/XRT), starting on 2008 Jan. 9.56 (UT dates and Chandra of XRT 080109 associated with SN 2008D, are used throughout this paper). and in §3 and §5 we present our extensive optical and After the announcement of the XRT detection, nu- near-infrared (NIR) photometry and spectroscopy on merous groups, including our efforts at Mt. Hopkins SN 2008D. Moreover, we present pre-explosion optical (AZ) and at Gemini-South, obtained data on the op- and NIR images obtained during the follow-up obser- tical counterpart (see Page et al. 2008, Soderberg et al. vations of SN 2007uy that were obtained 2.8 and 4.5 2008 and Malesani et al. 2008a for a detailed account), hours before the onset of XRT 080109 that yield stringent which revealed it to be a SN. Malesani et al. (2008b), upper limits. We discuss the photometric properties of § Blondin et al. (2008a), and Valenti et al. (2008b) origi- SN 2008D related to the two phases of evolution in 4 and § nally classified it as a SN Ic, possibly with broad spectral construct bolometric light curves in 6, after carefully es- § lines, and we (Modjaz et al. 2008b) were the first to sug- timating the extinction. 7 compares it with the rest of § gest a classification as a spectroscopically ordinary SN Ib, the known SNe with observed shock breakout, and 8 which is supported by subsequent reports (Valenti et al. presents nebular spectra that reveal double-peaked oxy- § 2008a; Soderberg et al. 2008; Malesani et al. 2008a). gen lines, indicative of global ejecta asphericities. In 9 SN 2008D was independently found by the Katzman we discuss in detail the X-ray properties of XRT 080109 Automatic Imaging Telescope (KAIT; Li & Filippenko compared with other kinds of X-ray events and pre- 2008) during the regular course of the optical Lick dict X-ray detections by future all-sky X-ray surveys. § Observatory Supernova Survey (Filippenko et al. 2001; We conclude with 10. We note that Soderberg et al. Filippenko 2005) that was monitoring NGC 2770 as part (2008) presented a similar and independent analysis on of the target galaxy sample. We promptly established XRT 080109/SN 2008D using the same Swift data and our extensive monitoring program of SN 2008D with fa- independent optical data, with which we compare some cilities at Mt. Hopkins and at Lick Observatory, extend- of our results. Unique aspects of our work include the ing our observing campaign that was already underway very early-time (< 1 day after outburst) optical and NIR data on SN 2008D, with which we can test the predicted early-time light-curve models of Soderberg et al. (2008), 601 University Drive, San Marcos, TX 78666. 16 Lawrence Livermore National Laboratory, 7000 East Avenue, the late-time spectra that show asphericities in the SN Livermore, CA 94550. core ejecta, and NIR spectra. 17 Institute for Astronomy, University of Hawaii, 2680 Wood- lawn Drive, Honolulu, HI 96822. 18 Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721. 2. X-RAY PHOTOMETRY AND SPECTROSCOPY 3

105 060218/SN2006aj 104 104 103 103 102 102 0.3-10 keV] 1 -0.8 ± 0.2 -1 t 10 101 100

erg s 0

42 10 10-1 -1 10 -3.4 ± 0.6 t 10-2 10-2 10-3 -3 10 080109 XRT Rate [0.3-10 keV] 10-4 Luminosity [10 Chandra 10-4 Optical/NIR Epochs 10-5 101 102 103 104 105 106 107 Time Since Observation Start [s] Fig. 1.— Observed temporal evolution of the X-ray luminosity of XRT 080109/SN 2008D (black points) in the XRT bands (0.3−10 keV), since t0= 2008-01-09 13:32:49, including the Chandra observation from 2008 Jan. 19. Depending on when one defines the time of onset, a α range of power-law slopes (green line) can be fit with α = −3.4 ± 0.6 (where LX ∝ t ) for data after ∼ 300 s. The count rate to luminosity conversion adopted here is 3.5 × 1042 erg s−1 per count s−1. The epochs on which optical and NIR data were obtained are indicated. See text for details. For comparison, we also plot the Swift X-ray light curve of XRF 060218/SN 2006aj (blue points, Campana et al. 2006; Butler et al. 2006), which Swift/XRT started observing ∼ 100 s after the gamma-ray trigger. We downloaded the Swift/BAT and Swift/XRT tα). The index increases if we assume a start time (Burrows et al. 2005) data from the Swift Archive19 and for the transient before the beginning of the observa- quick-look data site20. The XRT data were processed tion. For comparison, we plot in Figure 1 the X-ray with version 0.11.4 of the xrtpipeline reduction script light curve of XRF 060218/SN 2006aj, the most re- from the HEAsoft 6.3.121 software release. We employed cent GRB/XRF-SN connection (Campana et al. 2006; the latest (4 December 2007) XRT calibration files. Our Mirabal et al. 2006; Modjaz et al. 2006; Pian et al. 2006; reduction of XRT data from cleaned event lists output Sollerman et al. 2006), which was also discovered and ob- by xrtpipeline to science-ready light curves and spec- served by Swift/XRT (Campana et al. 2006; Butler et al. tra is described in detail by Butler & Kocevski (2007a). 2006). While the expression “X-ray flash” (XRF) is In particular, our custom IDL tools account on a frame- used in the literature when referring to the high-energy by-frame basis for pileup and source flux lost due to hot event 060218, the transient 060218 was qualitatively dif- or bad pixels and also perform rejection of pixels con- ferent from classical X-ray flashes and quite peculiar. In taminated by nearby field sources. § 9 we discuss in detail the properties of the peculiar In order to best account for contamination from nearby XRF 060218 and of XRT 080109 compared to other, clas- sources, we extract source flux in circular regions of two sical XRFs. sizes: 16 and 4.4 pixels radius, corresponding to 90% and The peak X-ray luminosity of XRT 080109 is ∼ 700 50% source containment, respectively. The spectra were times lower than that of XRF 060218 associated with fit using ISIS22. Quoted uncertainties below correspond SN 2006aj. No other SN besides XRF 060218/SN 2006aj to 90% confidence intervals. has been observed at such early time in X-rays, with which we could compare XRT 080109/SN 2008D. 2.1. X-ray Photometry: A 520 s Transient For the remainder of the paper, we adopt t0= 2008-01- The X-ray transient (Berger & Soderberg 2008) rises 09 13:32:49 (± 5 s) as the time of onset of the XRT — in a time 50 ± 30s, and has a duration T90 = 470 ± 30 i.e., the start of the Swift/XRT observations (Ofek et al. s or ∆tFWHM = 80 ± 30 s. The fluence of the X-ray 2008; Modjaz et al. 2008a). While we cannot fully rule transient is about 3.0 × 10−8 erg cm−2 in the 2−10 keV out an earlier onset of X-ray emission (Page et al. 2008; band (3.7 × 10−8 erg cm−2 in the 2−18 keV band). Xu et al. 2008), a generic Fast Rise Exponential Decay In Figure 1 we plot the X-ray light curve of (FRED) profile fit to the X-ray light curve of XRT 080109 SN 2008D, after adequately subtracting the other three constrains t0 to be less than 10 s before the first Swift X-ray field sources as resolved by Chandra observa- X-ray detection. The uncertainty of ∼ 10 s for the on- tions (Pooley & Soderberg 2008; see below). Prior to set of XRT 080109 influences its power-law decay indices 1 ksec but after the peak time, the X-ray luminosity very slightly, as indicated above, and does not affect our optical and NIR data analysis of SN 2008D. An indepen- (LX ) decay can be fit by a broken power law in time with late-time index α = −3.4 ± 0.6 (where L ∝ dent constraint comes from our pre-explosion images in X the optical bands. From the non-detection of the optical 19 ftp://legacy.gsfc.nasa.gov/swift/data . counterpart of XRT 080109 on those images (see § 3), 20 http://swift.gsfc.nasa.gov/cgi-bin/sdc/ql . we obtain an upper limit of 3 hours (i.e., 10 ksec) be- 21 http://heasarc.gsfc.nasa.gov/docs/software/lheasoft/. fore the start of the XRT observations. In other words, 22 http://space.mit.edu/CXC/ISIS/ . 4 with the reasonable assumption of a simultaneous on- 1990) at 5.1σ confidence (∆χ2 = 26.06 for 1 additional set of XRT 080109/SN 2008D, the onset of XRT 080109 degree of freedom), while the BB fit does not require ab- was less than 3 hours before the Swift/XRT observations, sorption in addition to that expected from our Galaxy 21 −2 which is fully consistent with the independent FRED fit (NH < 1.5 × 10 cm ) at 90% confidence. We return to XRT 080109. to the NH considerations and additional factors for the Chandra observations (Pooley & Soderberg 2008) have X-ray model selection in § 9. For completeness, we de- shown that three bright field sources lie near the X- rive relevant parameters for both models in the following ray transient and likely contribute substantial counts at section and summarize them in Table 1. late times even for small Swift/XRT extraction regions. The X-ray transient exhibits a declining spectral hard- These sources (X1, X2, and X3; see Pooley & Soderberg ness with time. To quantify this effect, we divide the 2008) are 3′′. 8, 1′′. 4, and2′′. 6 (9.0, 3.2, and 6.1 XRT pixels), counts into two time regions 0−140s and 140−520 s, each respectively, from the X-ray transient. Using the Chan- containing ∼ 170 source counts. Assuming no variation dra spectral fitting results and the Swift/XRT point- in NH with time, we find that the power-law photon in- spread function (PSF) from the calibration database, dex Γ (where N(E) ∝ EΓ) and BB temperature kT both we determine a field source contamination count rate of evolve at 4.2σ confidence (∆χ2 = 17.96 for 1 additional × −3 × −4 +0.5 +0.7 2 10 cps (6.9 10 ) cps for the 16 (4.4) pixel ex- degree of freedom) from Γ1 = 1.7−0.4 to Γ2 = 2.8−0.5 traction region. or kT1 = 0.90 ± 0.10 keV to kT2 = 0.56 ± 0.08 keV. In Correcting to the true source flux level using the PSF other words, either the power-law spectrum steepens or model, this shows that the flux from XRT 080109 (Fig. 1) the BB cools between the two time intervals. These can is dominated by the field source flux for any observations be compared to the values from the time-integrated fits +0.3 that start 5 ksec after t0 (i.e., the data point at 16 ksec) above, Γ = 2.1−0.4 or kT =0.75 ± 0.07 keV. and is likely completely overwhelmed by the field source The unabsorbed and time-averaged luminosity (as- flux after 1 day. We therefore use the 16 pixel extrac- suming a distance of D = 31 ± 2 Mpc; see § 4) is tion region for the temporal/spectral analysis prior to L = (12+5) × 1042 erg s−1 (0.3−10.0 keV) for the 5 ksec, while we use the 4.4 pixel extraction region to X,iso −2 ± × 42 −1 measure or limit the flux at later times. Based on the power-law model and LX,iso = (4.6 0.6) 10 erg s flux level implied by the Chandra observation and pos- for the BB model (bolometric). The peak luminosity for ± sibly indicated by the Swift/XRT data alone, the flux both models is a factor 2.9 0.6 times greater than the decay must become shallow at late times or an addi- time-averaged luminosity. tional emission component must dominate. At the dates The total X-ray energy released during the X-ray transient event is E = 5.8+2 × 1045 erg (0.3−10.0 keV) of our optical and NIR photometry (at t − to = ∆t = X,iso −1 46 0.84 and 0.71 days, i.e., ∼ 70 and 60 ksec), the X-ray for the power-law model and EX,iso = (2.4 ± 0.3) × 10 luminosity is between ∼ 2 × 1038−39 erg s−1. The es- erg s−1 for the BB model (bolometric). timated X-ray luminosity spans a large range (and may From the BB X-ray luminosity and flux, the BB radius X 9 only be an upper limit) since it is not clear if and when is RBB = (1.0±0.2)×10 cm, and this does not appear to the additional emission component dominates or when +4.8 vary significantly in time over the 520 s (∆R =1.1−4.3 × the decay becomes shallow. In any case, the X-ray lu- 108 cm). The combination of no change in radius and minosity (for 0.3−10.0 keV) is 3–4 orders of magnitude decreasing temperature is consistent (within the large lower than the luminosity inferred from the optical ob- uncertainties in temperature) with the observed decline servations (assuming a blackbody [BB] photosphere; see in X-ray luminosity (by a factor of ∼4) over the 520 s § ≈ −3 − −4 6.2), thus, LX /Lopt 10 10 at 70 ksec for duration, if the emission arises from a pure BB. XRT 080109/SN 2008D. Our results above are broadly consistent with the 2.2. X-ray Spectral Fitting: Power Law vs. Blackbody analysis performed by various groups (Page et al. 2008; Xu et al. 2008; Soderberg et al. 2008, Li 2008), except We restrict our spectral fitting to the time period for the late-time Swift/XRT data and the interpretation 0−520 s after the onset of XRT 080109, where the con- in Xu et al. (2008). Xu et al. (2008) do not take properly tamination from X-ray field sources is negligible. To ex- into account the fluxes from the contaminating sources plore whether the emission is thermal or nonthermal in that are fully revealed only by the Chandra observations, nature, we test two models. The data are best fit by an and this makes a difference for data at ∆t > 103 s. We absorbed power law (χ2/ν = 12.73/24) but also accept- furthermore note that while our derived power-law fit ably fit by an absorbed BB (χ2/ν = 28.56/25), as shown parameters are consistent within the uncertainties with in Figure 2. We note that the BB fit also has χ2/ν ≈ 1, those of Soderberg et al. (2008), our derived peak X-ray so it cannot be confidently excluded purely on statistical luminosity and EX are a factor of ∼2 and ∼ 4 lower, grounds as suggested by Soderberg et al. (2008). respectively, than their corresponding values, despite a The amount of soft X-ray absorption required for each similar adopted count rate to luminosity conversion fac- model impacts model selection. In the following we con- tor (scaled to the same distance), and this disagreement vert to an H-equivalent column density assuming the so- remains unexplained. lar abundances from Anders & Grevesse (1989). We note We note that the light curve of XRT 080109 is quite that recent abundance estimates (e.g., Lodders 2003) similar to that observed for XRF 060218/SN 2006aj, if result in a 50% larger value of NH . The power-law we allow a time stretch by a factor of ∼ 0.1 (Fig. 1). fit implies an additional H-equivalent column density This motivates seeking insight into the spectrum of +2.1 21 −2 NH = 5.2−1.8 × 10 cm , greater than that expected XRT 080109 by direct comparison with the X-ray spec- from the Galaxy (1.7 × 1020 cm−2; Dickey & Lockman trum of XRF 060218. The high signal-to-noise ratio 5 1 -1 keV -1 0.1 Counts s

4 2 0

∆χ [σ] -2 -4 0.8 1 2 3 4 5 6 Energy [keV]

Fig. 2.— Observed X-ray spectrum of XRT 080109 (crosses) in the Swift/XRT bands (0.3−10 keV) averaged over the 520 s duration of the event. Top: Best-fit blackbody (BB; solid line) and power-law (dashed line) models. Bottom: Data residuals (crosses) from the BB model fit (solid line) and from the power-law model fit (dashed line). The value for the deviation of the respective model fit from the data is the distance between the crosses and the respective lines. (S/N) X-ray spectrum of XRF 060218 during the pro- 3. OPTICAL PHOTOMETRY posed shock breakout phase at t < 6 ksec was predom- Optical photometry was obtained with KAIT inantly nonthermal but also required at high confidence (Filippenko et al. 2001)) and the 1-m Nickel telescope, a soft (kT = 0.1 − 0.2 keV) BB in addition to a power both at Lick Observatory, and the 1.2-m telescope at law (see Campana et al. 2006; also, Butler et al. 2006). the Fred Lawrence Whipple Observatory (FLWO). If we downsample the spectrum of XRF 060218 for t < 6 ksec to contain 380 counts as in XRT 080109, we 3.1. Photometric Calibrations find that the XRF 060218 spectrum is fit in a strikingly similar fashion to the XRT 080109 spectrum: a BB fit For photometric calibrations, the field of SN 2008D was +0.9 21 −2 has kT =0.72−0.5 keV and local NH < 3.6 × 10 cm observed in UBVRI during photometric nights on 2008 (χ2/ν = 35.22/23), while a power-law fit has Γ = 1.7±0.3 Jan. 18, 19, and 20 with KAIT, in BVRI on 2008Jan. 12 ′ ′ +1.8 21 −2 2 with the Nickel telescope, and in BV r i on 2008 Jan. 18 and local NH = 2.2 × 10 cm (χ /ν = 9.25/23). −1.5 with the FLWO 1.2-m telescope. About a dozen Landolt From the low χ2/ν, we see that the power-law model with (1992) standard-star fields were observed at different air- N in excess of Galactic also effectively overfits the soft H masses throughout each photometric night. Photometric spectral excess in XRF 060218, which requires a power solutions to the Landolt standard stars yield a scatter of law plus BB given all the counts. ∼ 0.02 mag for all the filters for the Nickel telescope, and It is quite reasonable that the spectrum of XRT 080109 about 0.03 mag for KAIT. The SN 2008D field was also may in fact be a superposition of a soft, thermal excess observed for several sets of UBVRI images with different and a harder, power-law continuum. The XRT 080109 depth in the photometric nights. The photometric solu- data alone do not constrain well this combined model. 2 tions are used to calibrate a set of local standard stars However, we note that we can find excellent fits (χ /ν = in the SN 2008D field as listed in Table 2, and a finder X 11 12.57/24) requiring RBB = 10 cm and assuming the chart is shown in Figure 3. We shall refer to the calibra- same (factor unity) ratio of power law to BB flux as in tion based on the Lick images as the “Lick calibration” XRF 060218 (Butler et al. 2006), without increasing NH throughout the rest of the paper. The calibration based relative to that in the pure power-law fit. The BB tem- on the FLWO 1.2-m telescope observations (“FLWO cal- perature in this case is kT =0.10 ± 0.01 keV. Hence, the ibration”) is consistent with the Lick calibration (within X apparent low RBB from the BB fit alone may be an ar- the uncertainties) in the filters that they have in common tifact of fitting a low S/N composite BB plus power-law (BV ). spectrum with an incorrect pure BB model. We discuss the implications and interpretation of the 3.2. KAIT Data Reduction X-ray data in § 9.1 . KAIT followed SN 2008D in BVRI and unfiltered mode nightly (weather permitting) after its discovery. 6

The data were automatically processed with bias subtraction, dark-current subtraction, and flat fielding. As SN 2008D occured on a spiral arm of NGC 2770, we use image subtraction to remove the contamination of the host-galaxy emission. Unfortunately, there are no multi- color KAIT template images for the field of NGC 2770 taken before SN 2008D. Fortunately, such images were obtained with the FLWO 1.2-m telescope during the course of follow-up photometry for SN 2007uy, which also occurred in NGC 2770 (as shown in Figure 3). For KAIT, the image subtraction and subsequent photometry reduction are performed with the KAIT photometry pipeline (M. Ganeshalingam et al., in preparation). Two independent image-subtraction routines were employed in the pipeline. One routine is based on the ISIS package (Alard & Lupton 1998) as modified by Brian Schmidt for the High-z Supernova Search Team, and the other is based on the IRAF23 task PSFMATCH (Phillips & Davis 1995). PSF fitting photometry is per- Fig. 3.— A finder chart for the local standard stars in the field of formed on the subtracted images, and the results from SN 2008D showing both SN 2008D and SN 2007uy in NGC 2770. The field of view is 10′×11′. North is up and east is to the left. the two routines are averaged whenever appropriate. Ar- The displayed image is the r′-band image taken with the FLWO tificial stars are injected into the original KAIT images 1.2-m telescope on 2008 Jan. 29. and extracted from the subtracted images to estimate the scatter of the measurements. To convert the KAIT instrumental magnitudes into the standard Johnson BV 3.4. FLWO 1.2-m Telescope Data Reduction and Cousins RI system, the color terms for the KAIT filters as determined from many photometric calibrations The field of SN 2007uy in NGC 2770 was monitored are used in the conversion (as detailed in Modjaz et al. by the FLWO 1.2-m telescope in the context of the CfA 2001 and Li et al. 2001). The final error bars for the SN monitoring24 efforts when SN 2008D was discov- magnitudes are the scatter in the artificial star simula- ered, which provide us with valuable pre-explosion im- tion and the calibration error added in quadrature. ages. We followed SN 2008D on a nightly basis (weather ′ ′ We note that the FLWO 1.2-m telescope template im- permitting) in BV r i . For performing photometry of ages were taken with the Johnson BV and Sloan r′i′ FLWO 1.2-m telescope images, we adopted the image- filters, while the KAIT observations of SN 2008D were analysis pipeline of the SuperMACHO and ESSENCE taken with the Johnson BV and Cousins RI filters. The collaborations (see Rest et al. 2005, Garg et al. 2007, difference between the r′i′ and RI filter transmission and Miknaitis et al. 2007 for details). The operations curves raises the possibility of a systematic uncertainty of this optical photometry pipeline will be discussed in during the image-subtraction process. We investigated depth by M. Hicken et al. (in preparation). In brief, this issue with the artificial star simulations and con- we employed differential photometry by measuring the cluded that the systematic uncertainty is smaller than brightness of the SN with respect to a set of stars (see ∼ 0.03 mag (which is not reported in the final uncertain- Figure 3) in the SN field and we employed the DoPHOT ties). photometry package (Schechter et al. 1993) to measure The final photometry of SN 2008D from the KAIT ob- the flux of the SN and its comparison stars. We per- servations is listed in Table 3. formed image subtraction with the robust algorithm of Alard & Lupton (Alard & Lupton 1998; Alard 2000), using as template images the 1.2-m images taken on 2008 3.3. Nickel Telescope Data Reduction Jan. 08 of the field. The “FLWO calibration” and the color terms for Keplercam (Modjaz 2007) were used to SN 2008D was remotely observed with the Lick Obser- convert the instrumental magnitudes into the standard vatory 1-m Nickel telescope in BVRI on 2008 Jan. 11 system. and 12. These data were processed with proper bias and We present the 1.2-m BV r′i′ photometry in Table 4. flat-field images, and the image subtraction and photom- For multiple observations on the same night we report the etry were performed with the KAIT photometry pipeline weighted mean of those observations. Intra-night obser- (but modified to deal with the Nickel telescope images). vations did not reveal statistically significant variations The FLWO 1.2-m telescope template images were used for SN 2008D. We furthermore derive upper limits for to generate the subtracted images, and the Lick calibra- SN 2008D based on 2008 Jan. 9.44 images of the field of tion and the proper color terms for the Nickel telescope SN 2007uy, three hours before the onset of XRT 080109 filters are used to convert the instrumental magnitudes (§ 2) and present them in Table 4. The upper limits were into the standard system. The final photometry is also estimated by placing fake stars with a range of magni- listed in Table 3. tudes at the position of SN 2008D on the template images and then determining when they are detected at varying 23 IRAF is distributed by the National Optical Astronomy Ob- servatory, which is operated by the Association of Universities for thresholds above background. Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation. 24 See http://www.cfa.harvard.edu/supernova/index.html . 7

2MASS Dec 2, 1999 PAIRITEL Jan 9, 2008 PAIRITEL Jan 12, 2008

SN 2008D

SN 2007uy

Fig. 4.— Montage of three composite NIR (J,H,KS) images for the field of NGC 2770, before it hosted SN 2007uy and SN 2008D (using an archival 2MASS image, left), on 2008 Jan. 9.35 (PAIRITEL, middle) just before SN 2008D was visible (i.e., 4.5 hours before the onset of XRT 080109), and on 2008 Jan. 12 (PAIRITEL, right), when both SNe are visible. 3.5. UVOT Data Reduction ters Automated Infrared Telescope (PAIRITEL)25 located at FLWO. PAIRITEL is the world’s largest NIR The Swift/UVOT has followed SN 2008D intensively imaging telescope dedicated to time-domain astronomy (Kong et al. 2008b, Immler et al. 2008a, Li et al. 2008, (Bloom et al. 2006). The telescope and camera were and Immler et al. 2008b, Soderberg et al. 2008). part of the Two Micron All Sky Survey (2MASS; We retrieved the level-2 UVOT (Roming et al. 2005) Skrutskie et al. 2006) project. J-, H- and Ks-band (1.2, data for SN 2008D from the Swift data archive. Al- 1.6, and 2.2 µm; Cohen et al. 2003) images were acquired though UVOT observed the SN with all available filters, simultaneously with the three NICMOS3 arrays with in- there are only very few reported detections (Immler et al. dividual exposure times of 7.8 s. Each image consists of 2008a; Soderberg et al. 2008) for the UV filters (includ- a 256×256 pixel array with a scale of 2′′ pixel−1, yield- ing uvw1, uvm2, and uvw2), so we focus our reduction ing an individual field of view (FOV) of 8.′5 × 8.′5. These to the U, B, and V filters. Although photometric cali- individual raw images were dithered after every third ex- bration for these filters have been published by Li et al. posure in order to compensate for bad pixels and the (2006) and Poole et al. (2008), these calibrations deal bright NIR sky. with isolated point sources, and not objects with a com- The PAIRITEL reduction pipeline software plicated background such as SN 2008D. (Bloom et al. 2006) estimated the sky background Template images for the SN 2008D field in the UVOT from a star-masked median stack of the SN raw images. UBV filters exist due to the follow-up observations of After sky subtraction, the pipeline cross-correlated, SN 2007uy. However, the image-subtraction technique stacked, and subsampled the processed images in order for the UVOT data is complicated by the fact that the to produce the final image with an effective scale of UVOT detector is a photon-counting device that suffers 1′′ pixel−1 and an effective FOV of 10′×10′. The effec- from coincidence loss (Roming et al. 2005); the flux of an tive exposure times for the final “mosaic” images ranged object is not linearly proportional to its brightness. For- from 15 to 20 minutes. Multi-epoch observations over tunately, SN 2008D is sufficiently faint (U > 19.0 mag, the course of the night (totaling between 40 min to 4 B > 18.4 mag, and V > 17.4 mag, see below) that the hours) were obtained to search for intranight variations, corrections are very small (< 0.01 mag), much smaller but none were found at a statistically significant level. than the photometric calibration scatter of ∼ 0.04−0.05 For the NIR photometry, we used the image-analysis mag (Li et al. 2006). pipeline of the ESSENCE and SuperMACHO projects We used the UVOT observations of SN 2007uy on 2008 (Rest et al. 2005; Garg et al. 2007; Miknaitis et al. Jan. 6 as the template images to reduce the data of 2007). This pipeline performs PSF photometry with SN 2008D (a total exposure time of 808 s in each of U, B, the DoPHOT photometry package (Schechter et al. and V filters). We employed the photometric calibration 1993). We performed difference photometry (following as described by Li et al. (2006) to measure the photom- Wood-Vasey et al. 2007) using the image-subtraction al- etry of SN 2008D on the subtracted images. Many tests gorithm of Alard (2000). As a template image for the have shown that the two calibrations by Li et al. (2006) field of SN 2008D we used our PAIRITEL images of and Poole et al. (2008) are consistent with each other. SN 2007uy taken on 2008 Jan. 9. Thus, we also de- The UVOT photometry of SN 2008D is reported in rive NIR upper limits for SN 2008D based on images of Table 5, and the overall optical photometry presented in the field of SN 2007uy, taken on 2008 Jan. 9 08:26:19, 4.5 § Figure 5 with respect to start of XRT 080109 ( 2). The hours before t0, the onset of XRT 080109 (see §2), and UVOT BV values reported here are fully consistent with present them in Table 6. The upper limits were estimated those in Soderberg et al. (2008), while our U photometry in the same fashion as for the optical data, namely by is systematically fainter by 0.2 mag. placing fake stars with a range of magnitudes at the position of SN 2008D on the template images and then determining when they are detected at varying thresholds above background. For illustration, we show in Figure 4 3.6. Near-Infrared Photometry various NIR images that compare the field of NGC 2770

We obtained NIR (JHKs) photometry measurements 25 See http://www.pairitel.org/ . with the refurbished and fully automated 1.3-m Pe- 8

K −10 5 s H−8 J−6 10

i’−4

15 r’−2

V 20 B+2 Standard Magnitude + Offset U+4

25 S08: i’ S08: r’ S08:V S08:B S08:U

−1.0 0.1 1.0 10.0 100.0 Days after Outburst

Fig. 5.— Observed optical and NIR light curve of SN 2008D after the onset (right panel) of XRT 080109 at t0 = 2008-01-09 13:32:49 (± 5 s), which we take as the time of shock breakout. The filled circles show UBV data from Swift/UVOT, the empty circles are BV data from ′ ′ KAIT (RI data are not shown for sake of clarity), while the empty squares are BV r i data from the FLWO 1.2-m telescope. JHKs data (filled stars, triangles, and squares) are from PAIRITEL. Note the very early optical data points (∆t =0.84 days after shock breakout) from the FLWO 1.2-m telescope, as well as NIR data (at ∆t =0.71 days) from PAIRITEL. Swift/UVOT upper limits are indicated by the arrows. We also plot the pre-explosion upper limits derived from the 1.2-m CfA (at ∆t = −2.8 hr) and the PAIRITEL (at ∆t = −4.5hr) data (left panel). The data have not been corrected for extinction. We note that our earliest ground-based BV r′i′ data points are consistent with the light curves predicted by the cooling envelope BB emission model (dotted lines) of Soderberg et al. (2008, S08). before the appearance of the two SNe using an archival 0.1 days after burst, and going out to 109 days. They 2MASS image (left) with more recent images taken with show the typical two-component light curve (see also PAIRITEL: just before SN 2008D was visible on 2008 Soderberg et al. 2008): the first due to shock breakout Jan. 9.35 (middle), and 3 days after SN 2008D appeared that is most expressed in the blue bands (∆t = 0.1−4 (on 2008 Jan. 12, right). days) and the second due to the standard radioactive de- We performed the final calibration onto the 2MASS cay of 56Ni and 56Co (∆t & 5 days) with half-life times of system by using as reference stars the field stars from the 6.6 days and 77.3 days, respectively, which lead to energy 2MASS catalog, of which there were 20−30 in the FOV. deposition behind the photosphere in SN. No color-term corrections were required since our natural The most remarkable aspect is the very early optical system is already in the 2MASS system. We extensively and NIR data points (at ∆t = 0.84 and 0.71 days, re- tested the accuracy and precision of both the PAIRITEL spectively), which are the earliest broad-band observa- reduction and our photometry pipeline by comparing our tions reported of a SN Ib. They were obtained as part photometry of 2MASS stars in the SN observations to of our observing campaign of SN Ib 2007uy in the same that in the 2MASS catalogue (Modjaz 2007). We present galaxy. Here all four (BV r′i′) filters indicate a first peak our PAIRITEL photometry in Table 6. at 0.8−1.8 days. This behavior appears like a plateau phase in Figure 5 since time is plotted in log space. We 4. EARLY-TIME BROAD-BAND PHOTOMETRY are confident in our photometry at early phases, because the SN is clearly detected in the subtracted frames and Figure 5 shows the optical (UBVr′i′) and NIR (JHK ) s the later photometry is in excellent agreement with both photometric evolution of SN 2008D, starting as early as 9 the Swift/UVOT and the KAIT photometry in the same Extraction of the low-resolution spectra were done us- filters. We note that these earliest BV r′i′ data points ing the optimal weighting algorithm (Horne 1986), and are consistent with the light curves predicted by the cool- wavelength calibration was accomplished with HeNeAr ing envelope BB emission model of Waxman et al. (2007) lamps taken at the position of the targets. Small-scale used by Soderberg et al. (2008), whose first optical data adjustments derived from night-sky lines in the observed points in BV RI start later, at ∆t = 1.9 days. frames were also applied. The spectra were either taken Our V -band light curve of SN 2008D peaks on 2008 at the parallactic angle (Filippenko 1982), at low airmass, Jan. 27.9 ± 0.5 (i.e., at ∆t = 18.3 ± 0.5 days) at ap- or with an atmospheric dispersion corrector, in order to parent magnitude mV = 17.33 ± 0.05 mag. The B-band minimize differential light loss produced by atmospheric rise time (between shock breakout and maximum light) dispersion, except for the GMOS-South spectrum. Tel- is slightly shorter than in the V band, namely 16.8 ± 0.5 luric lines were removed following a procedure similar days, peaking on 2008 Jan. 26.4 at mB = 18.5 ± 0.05 to that of Wade & Horne (1988) and Matheson et al. mag. These rise times are at the long end of the range (2000b). The final flux calibrations were derived from of those observed for stripped-envelope SNe, and similar observations of spectrophotometric standard stars. Since to SN 1999ex (Stritzinger et al. 2002). After correcting the MMT and FAST spectrographs suffered from second- for a Galactic extinction and host-galaxy reddening of order light contamination with the grating used during E(B − V )Host = 0.6 ± 0.1 mag (see § 6.1), these val- a few of our observations, we observed standard stars of ues correspond to a peak absolute magnitude of MV = different colors. By carefully cross calibrating the data −17.0±0.3 mag and MB = −16.3±0.4 mag for SN 2008D, with blue standards for the blue wavelength range and using a distance modulus of µ = 32.46 ± 0.15 mag (which with red standards for the red, we believe we minimized corresponds to a distance of D = 31 ± 2 Mpc) . This dis- the impact of the second-order light. This is supported tance modulus is based on the measured heliocentric ve- by the fact that the MMT+Blue Channel spectra and locity of NGC 2770, corrected for the effect of the Virgo those from the Lick+Kast Double Spectrograph (that cluster and the great attractor by using the local velocity uses an order-blocking dichroic on the red CCD) from field model given by Mould et al. (2000). These are mod- the same night are consistent. est peak luminosities compared to those of other known Three NIR spectra were obtained on 2008 Jan. 14.45, SNe Ib and SNe Ic, in particular about 1 mag fainter than 21.51, and 28.51 using the 3.0-m telescope at the the mean of the SN Ib sample in Richardson et al. (2006), NASA Infrared Telescope Facility (IRTF) with the SpeX but within the range of observed values. SN 2008D is medium-resolution spectrograph in prism (LRS) mode fainter than SN 1999ex (by ∼0.9 mag and 1.0 mag in (Rayner et al. 2003). The SpeX instrument provides V and B, respectively) and SN 1993J (by ∼ 0.4 mag). single-exposure coverage of the wavelength region from Since the peak luminosity is expected to be set by the 0.8−2.5 µm in either cross-dispersed or prism mode. The amount of synthesized 56Ni, SN 2008D likely produced data were reduced and calibrated using a package of less than MNi . 0.07 M⊙, barring any strong asphericity IDL routines specifically designed for the reduction of during the explosion which would modify the estimated SpeX data (Spextool v3.2; Cushing et al. 2004). Cor- mass (Höflich 1991) (but see § 8). rections for telluric absorption were performed using the extracted spectrum of an A0V star and a specially designed IDL package developed by Vacca et al. (2003). 5. SPECTROSCOPIC OBSERVATIONS Additional details of the data acquisition and reduction Optical spectra were obtained with a variety of in- methods are given by Marion et al. (2008). struments: Gemini-South via queue-scheduled observations (GS-2007B-Q-2, PI: Chen), the 6.5-m MMT, 5.1. High-Resolution Spectroscopy: Probing the the 10-m Keck I and Keck II telescopes, the 3-m Interstellar Medium Shane telescope at Lick Observatory, the Astrophys- We obtained one high-resolution spectrum with Keck ical Research Consortium 3.5-m telescope at Apache I on 2008 Jan. 12.5 (i.e., at ∆t =2.95 days), over the Point Observatory (APO), and the FLWO 1.5-m tele- ˚ scope. The spectrographs utilized were the GMOS- observed wavelength range 3440−6300 A. We detect underlying H II-region emission lines (Balmer lines and South (Hook et al. 2003) at Gemini, the Blue Channel II I II (Schmidt et al. 1989) at the MMT, the LRIS (Oke et al. [O ] λ3727), as well as Na and Ca H&K absorption lines in the host galaxy (see also Soderberg et al. (2008); 1995) and HIRES (Vogt et al. 1994) at Keck I, the I II DEIMOS (Faber et al. 2003) at Keck II, the Kast Dou- Malesani et al. (2008a)). The Na and Ca H&K absorption lines have identical redshifts as the centroid of ble Spectrograph (Miller & Stone 1993) at Lick, the DIS II spectrograph at the APO, and the FAST spectrograph the [O ] emission lines, z = 0.00700 ± 0.00005. Since (Fabricant et al. 1998) on the FLWO 1.5-m telescope. both emission and absorption lines show the same velocity, we consider the Na I and Ca II H&K absorption Almost all optical spectra were reduced and calibrated lines to have interstellar origin (as opposed to circum- employing standard techniques (see, e.g., Modjaz et al. −1 2001) in IRAF and our own IDL routines for flux cali- stellar) and attribute the offset of ∼160 km s from bration. In addition, the HIRES spectrum was reduced the measured redshift of the nucleus of the host galaxy using HIRedux26, an IDL package developed by Jason X. NGC 2770 (znuc =0.006451±0.000087; Falco et al. 1999) Prochaska for the HIRES mosaic, and the GMOS-S and to the galaxy’s rotation curve. Given that NGC 2770 DEIMOS spectra were reduced using slightly different is an inclined, edge-on, typical spiral (Sb/Sc) galaxy techniques (see Foley et al. 2006, 2007). (Soderberg et al. 2008), rotation on the order of 160 km s−1 at the distance of the SN from the nucleus (∼ 26 http://www.ucolick.org∼xavier/HIRedux/index.html . 9 kpc) is plausible; the full rotation curve of NGC 2770 10

Valenti et al. 2008b). In § 5.3 below, we discuss the 2.0 ﬂeeting, double-absorption feature around 4000 A˚ which is only seen in our spectrum at ∆t = 1.84 days 1.5 (Blondin et al. 2008a).

λ Figure 7 shows that the optical helium lines 1.0 (blueshifted He I λ4471, He I λ5876, He I λ6678, and He I λ7065) become apparent starting at ∆t ≈ 0.5 10−12 days (8−6 days before V -band maximum)

Normalized f and grow stronger over time, allowing this SN to be classified as a SN Ib (Modjaz et al. 2008b; Valenti et al. 0.0 2008a; Soderberg et al. 2008; Malesani et al. 2008a). This is in stark contrast to all SNe associated with −0.5 5930 5932 5934 5936 5938 5940 5942 GRBs that have been exclusively broad-lined SN Observed Wavelength (Å) Ic (e.g., Galama et al. 1998; Stanek et al. 2003; Hjorth et al. 2003; Malesani et al. 2004; Modjaz et al. Fig. 6.— Observed high-resolution Keck spectrum of the host- 2006; Mirabal et al. 2006; Campana et al. 2006; galaxy Na I D1 and D2 absorption lines in SN 2008D, taken on Pian et al. 2006). 2008 Jan. 12.5 (∆t = 2.95 days). Note that two Na ID lines are saturated, as are the Ca H&K lines (not shown here). We derive z = 0.00700 ± 0.00005. 5.3. The Double-Absorption Feature around 4000 A˚ extends to ∼400 km s−1 (Haynes et al. 1997). Perhaps the most intriguing aspect of the very early- In Figure 6, we display the observed and normalized time spectra of SN 2008D is the transient, prominent, high-resolution spectrum of SN 2008D around the host- double-absorption feature around 4000 A˚ (Fig. 8). It is galaxy Na D and D absorption lines. Since the ab- 1 2 only present in our MMT+Blue Channel spectrum taken sorption lines are saturated, we can only place lower on Jan 11.41 (∆t = 1.84 days), and was first noted limits to the column densities of Na I and Ca II H&K: −2 −2 by Blondin et al. (2008a). Multiple tests have shown N > 1.0 × 1013 cm and N > 1.4 × 1013 cm . Na I Ca II that this feature is not an artifact of the data reduction We measure an equivalent width (EW) for Na I D of 2 or flux calibration processes, but is real. Furthermore, ˚ II 0.67 ± 0.04 A (at rest), and for Ca K of 0.55 ± 0.04 Malesani et al. (2008a) obtained spectra of SN 2008D A˚ (at rest). In § 6.1 we discuss how we use the EW at nearly the same epoch, Jan. 11.3 (i.e., 2 hours ear- measurements to constrain the reddening for the SN in lier than our spectrum) with wavelength range extending this system. Unfortunately, our HIRES spectrum does down to 3700 A,˚ and note that their spectra show the not cover Hα in SN 2008D, in which case we could have same double-absorption feature. searched for any shock-ionized hydrogen from the cir- The double-absorption feature is apparently absent cumstellar medium. from our Gemini+GMOS spectrum taken on Jan 11.26, but the spectrum is extremely noisy below ∼ 4500 A˚ 5.2. Low-Resolution Spectroscopy due to the particular setup used. The spectrum of In Figure 7, we present the sequence of early-time op- Soderberg et al. (2008) on the same and subsequent tical spectra after subtracting the SN recession veloc- nights apparently does not extend sufficiently far to the ity, dereddened by E(B − V )Host = 0.6 ± 0.1 mag blue to include this feature at ∼4000 A˚ (their Fig. 3). (§ 6.1). Furthermore, the spectra are labeled accord- The double-absorption line is no longer visible in spectra ing to age (∆t) with respect to date of shock breakout taken & 1 day later (i.e., ∆t & 2.99 days) and the two (t0 = 2008-01-09 13:32:49) and epoch (tVmax), defined as components appear to blend into a single, broad (FWHM days relative to V -band maximum (2008 Jan. 27.9 = JD ≈ 23, 000km s−1) absorption feature (maximum absorp- 2454493.4, see § 4). Knowing both reference dates is cru- tion at rest-frame ∼ 4100 A),˚ as clearly seen in the Jan. cial for interpreting the spectra and their time evolution. 13.30 MMT+Blue Channel spectrum (∆t =3.74 days) in These are the earliest spectra presented of a normal Figure 8. Also, spectra taken by Malesani et al. (2008a) SN Ib (starting 1.70 days after shock breakout, which do not show the feature in spectra taken after Jan. 13.07, means 17.0 days before V -band maximum) and provide i.e., ∆t & 3.51 days. insights into the different layers of the SN ejecta. As the The double-absorption feature appears to consist of photosphere recedes with time, it reveals lower-velocity two overlapping P-Cygni profiles, with rest-frame full material since the SN ejecta are in homologous expan- width at half-maximum (FWHM) of ∼ 11, 500km s−1 sion (where v ∝ R). Moreover, we obtained a late-time − and ∼ 11, 000km s 1 for the bluer (maximum absorption spectrum at t = +91 days that shows the advent of Vmax ˚ the nebular epoch, in which the spectrum is marked by at rest-frame ∼ 3980 A) and redder (maximum absorp- strong forbidden emission lines of intermediate-mass ele- tion at rest-frame ∼ 4240 A)˚ absorptions, respectively. ments [O I] and [Ca II] (not shown in Fig. 7). We discuss Both absorption components are thus significantly nar- the late-time spectrum in § 8. rower than other absorption features in the same spec- −1 At early times our densely time-sampled series shows trum, whose FWHM ranges between ∼ 15, 000km s −1 the dereddened spectra to consist of a blue, nearly fea- and ∼ 30, 000km s . tureless continuum (except around 4000 A),˚ with super- In the following section we attempt to address the imposed characteristic broad SN lines not attributed to line identification for this feature. While Malesani et al. hydrogen (Malesani et al. 2008b; Blondin et al. 2008a; (2008a) cannot firmly identify this double-absorption fea- 11

SN 2008D (dered by E(B−V)=0.6 mag) ∆ t = tVmax = 1.84 d −16.6 d 2.80 d −15.6 d 20 2.99 d −15.4 d 3.74 d −14.7 d 3.82 d −14.6 d 4.79 d −13.6 d 4.84 d −13.6 d 15 5.75 d −12.7 d 5.80 d −12.6 d 6.77 d −11.7 d 9.00 d −9.4 d + Constant

λ 11.66 d −6.8 d 22.78 d +4.3 d 10 23.88 d +5.4 d

Scaled f 30.56 d +12.1 d 32.61 d +14.2 d 33.87 d +15.4 d 37.79 d +19.4 d 5 49.81 d +31.4 d

60.67 d +42.2 d

HeI HeI HeI HeI CaII 0 4000 6000 8000 10000 Rest Wavelength (Å)

Fig. 7.— Spectral evolution of SN 2008D, dereddened by E(B − V )Host =0.6 mag and labeled with respect to date of shock breakout, t0 = 2008–01-09 13:32:49 (indicated by ∆t), and to date of V -band maximum, 2008 Jan. 27.9 (indicated by tVmax). The first few are the earliest spectra presented of a SN Ib. Note the fleeting double-absorption feature around 4000 A˚ in our early spectrum at ∆t = 1.84 days, which is discussed in § 5.3. Some spectra have been binned to more clearly visualize their effective S/N. Unfortunately, we have a gap in our spectroscopic coverage around maximum light. The characteristic optical He I lines (due to blueshifted He I λ4471, He I λ5876, He I λ6678, and He I λ7065) become visible starting ∆t ≈ 12 days or tVmax ≈ −6 days. Note that not all spectra are presented in this figure. For the late-time spectrum, see § 8. ture, they ascribe it to Fe II, which can produce many other weaker features, as shown in Figure 9 (using a con- lines in the blue part of the spectrum. Amongst the pub- tinuum temperature of ∼13,000 K, consistent with the lished supernova spectra, the closest match to the “W” BB temperature as derived from our photometry data, feature that we can find is in the earliest spectrum of SN Table 7). SYNOW did not produce any other strong 2005ap (Quimby et al. 2007) from 2005 March 7 (∼ 1 features that are associated with these ions, matching week before maximum light), which shows a similar fea- the full extent of the observed spectrum. We tried to fit ture at slightly redder wavelengths (by ∼200 A).˚ Quimby the spectrum with Fe II (as suggested by Malesani et al. et al. used the parameterized supernova spectrum syn- 2008a) but were unable to find a good fit, since Fe II thesis code SYNOW (Fisher et al. 1999; Branch et al. produces features to the red of the “W” feature that are 2002) to fit that spectrum with four ions. They ascribed not observed. the “W” feature to a combination of C III, N III, and O While our SYNOW fit produces a good match to III with a photospheric velocity of 21,000 km s−1. We the data, we note that SYNOW fits do not produce downloaded the spectrum of SN 2005ap from the SUS- unique line identifications and the suggested identifi- PECT database27, generated SYNOW models using only cation here needs to be confirmed by detailed spec- those three ions, and were able to satisfactorily repro- tral modeling (Mazzali & Lucy 1998; Baron et al. 1999; duce the “W” feature in SN 2005ap using similar fitting Dessart & Hillier 2007). Furthermore, while the “W” parameters to those of Quimby et al. We then increased feature is no longer visible in spectra of SN 2008D taken what SYNOW calls the “photospheric velocity” (vphot) of ∼ 1 day later (see also Malesani et al. 2008a), it seemed the fit to match the observed wavelengths of the “W” fea- to linger for ∼ 1 week in the spectrum of SN 2005ap −1 ture in SN 2008D and found that vphot ≈ 30,000 km s (at slightly lower velocities). In addition, SN 2008D and produced a good match to the “W” as well as to some SN 2005ap are different in many other ways: SN 2005ap was much more luminous than SN 2008D (by ∼4 mag, 27 http://bruford.nhn.ou.edu/ suspect/index1.html . i.e., a factor of ∼40), and never developed He I lines; 12

3 O III SN 2005ap + constant λ N III

2 O III

Scaled f C III/ N III 1 SN 2008D

3500 4000 4500 5000 5500 Rest Wavelength (Å)

Fig. 9.— The reddening-corrected early-time spectrum of SN 2008D (black) at ∆t = 1.84 days (2008 Jan 11.41) and for comparison, the spectrum of SN 2005ap (Quimby et al. 2007). The fit via the parameterized spectral synthesis code SYNOW is shown in red: the “W,” along with some other weaker features, can be reproduced with a combination of O III and CIII/NIII. asphericity is involved, then existing spectral synthesis codes (such as SYNOW which assumes spherical symmetry) are challenged to accommodate that aspect. Thus, we encourage theoretical efforts to verify our suggested spectral identification for this seemingly important and transient feature. The MMT spectra (including the one displaying the double-absorption feature) will be made publicly available through the CfA Supernova Archive29 .

Fig. 8.— Early-time spectra of SN 2008D in the range 3500- 6. BOLOMETRIC LIGHT CURVES 4700 A,˚ from ∆t = 1.70 to ∆t = 4.79 days past the X-ray outburst, not corrected for reddening. The transient double-absorption fea- 6.1. Constraining the Amount of Extinction ture is marked with arrows in the spectrum at ∆t = 1.84 days. Many lines of evidence indicate significant host galaxy indeed, its spectral SN classification and the exact ori- extinction toward SN 2008D. The high-resolution spec- gin for its large luminosity are uncertain (Quimby et al. trum (§ 5.1) shows strong interstellar Na I D and Ca II 2007). lines at the recession velocity of the host galaxy, while At face value, an identification of the “W” feature with low-resolution spectra show in addition diffuse interstel- C III, N III, and O III at high velocities seems plausi- lar bands (Herbig 1975) that indicate a significant col- ble. We expect the early supernova spectrum to have umn of gas along the line of sight in the host galaxy. Fur- high velocities based on an extrapolation of the observed thermore, the observed spectra and colors are much red- velocities at late times back to the earliest observations. der than those of other SNe Ib at comparable epochs. In In addition, what SYNOW designates as the “excitation this section we estimate the extinction toward SN 2008D temperature” is high (Tex ≈ 17,000 K) at these early from various independent methods and derive E(B − V ) times28 and many of the elements that will contribute = 0.6 ± 0.1 mag, consistent with the value adopted to the later spectra are more highly ionized. The tran- in Soderberg et al. (2008). The Milky-Way reddening sient nature of the feature might then be explainable along the line of sight to SN 2008D is E(B − V )MW = because the elements involved will recombine to lower 0.0226 (Schlegel et al. 1998). This is negligible compared ionization stages as the ejecta cool (at ∼15,000 K where to the host galaxy extinction and thus we ignore the MW most of doubly ionized elements recombine to singly ion- contribution for our discussion of the total extinction for ized species). SN 2008D. Alternatively, the reason for the short-lived nature of While the presence of Na I D absorption in SN spec- this “W” feature could be asphericity in the SN ejecta. tra usually indicates some extinction along the line of We have evidence from late-time spectra for aspheric- sight, it is usually difficult to determine the exact value ity in at least the inner SN core (see § 8), which may of E(B − V )Host based on EW(Na I D) alone. This diffi- be connected to some inhomogeneity in the SN ejecta culty is due to a variety of reasons: saturation effects of density or chemical abundance profile at the outermost the Na I D absorption lines, possible depletion of sodium radii that the early-time spectra are probing. If indeed in high-density environments, and varying gas-to-dust

28 ratios (see detailed discussions in Leonard & Filippenko The apparent incongruency between the excitation and 2001 and Turatto et al. 2003). With these warnings continuum temperatures probably arises from the LTE (local thermodynamic equilibrium) assumption in the SYNOW code in mind, we follow Munari & Zwitter (1997) and derive (Branch et al. 2002), or because we underestimated the reddening, but this is less likely. 29 http://www.cfa.harvard.edu/supernova/SNarchive.html . 13

E(B − V )NaID = 0.5 mag using EW(Na I D) = 0.7 A,˚ For our first method, we perform a least-squares fit which should be used as a lower limit to the reddening of a BB spectrum to the SED of each epoch and con- because the lines are saturated. In addition, comparison strain the derived BB temperature (TBB), BB radius of spectra and B −V colors with those of other SNe Ib at (RBB), and BB luminosity (LBB) as a function of age comparable epochs, such as SN 1999ex (Stritzinger et al. (see Table 7 and Fig. 10). The error bars include the 2002; Hamuy et al. 2002), yields E(B − V ) = 0.6 − 0.8 uncertainties from the formal fits as well as uncertain- mag. ties introduced by the uncertain reddening (±0.1 mag), This value for the reddening is consistent with fitting which affect the derived BB fit parameters differently, de- an absorbed BB spectrum to the spectral energy distribu- pending on the underlying SED. The reddening-induced tion (SED) of SN 2008D constructed from our simultane- uncertainties for the BB temperatures are the largest at ous broad-band (UBVRIJHKs) photometry (see § 6.2). early times, where temperatures are high such that a lot Perhaps the most stringent constraint comes from com- of the flux is in the shortest wavelength points, which in paring SN 2008D with SN 2005hg, a normal SN Ib that turn are very sensitive to small changes in reddening. We suffered very little host-galaxy extinction (Modjaz 2007). do not include the systematic uncertainty in distance (of Comparison of the spectra (Fig. 15) and the broad-band order ∼20%). At later times (& 8−10 days) the assump- ′ ′ tion of a BB continuum is expected to break down, based SEDs constructed from our UBVr i JKHs photometry for both SNe indicates a host-galaxy extinction for on the optical spectra where distinct absorption lines de- 2 SN 2008D of E(B − V ) ≈ 0.65 mag, if both SNe have velop (see Fig. 7). While the χ values of the fits are similar evolution. strictly not acceptable except at early times (1 < ∆t < 5 days), the overall fits are reasonable overall even be- We therefore adopt E(B − V )Host = 0.6 ± 0.1 mag for SN 2008D, and a Milky Way extinction law yond this time with no obvious residual trends, and we (Cardelli et al. 1989). This is the value used to dered- consider the BB measured temperatures, radii, and lumi- den all of our spectra (see Fig. 7). nosities to still be generally quite reliable. For clarifica- tion, we note that the optical/NIR data to which we fit a 5 6.2. Constructing Bolometric Light Curves BB were obtained much later (∆t & 10 sec) than when the main X-ray emission period of XRT 080109 occured Our broad-band data set (spanning 3,000A˚ − 22,000 (∆t . 103 sec, § 2). Thus, the Xray and optical/NIR BB A)˚ lends itself uniquely to computing the bolometric fits were not and cannot be performed simultaneously. output of SN 2008D. We determine it in two differ- For our crucial first ground-based data point (at ∆t = ′ ′ ent ways: (a) by fitting a BB to the optical and NIR 0.71−0.84 days), we use our ground-based BV r i JHKs broad-band photometry and deriving the equivalent BB data to construct the SED. The BB fit to the SED opt +3.3 opt luminosity (LBB) by analytically computing the corre- 13 yields RBB = (27.9−3.8) × 10 cm, and TBB = sponding Planck function, and (b) by performing a di- +3300 ′ ′ (11, 7000−2000)K (Table 7). We note that, in order for rect integration of the SN UBVRr Ii JHKS broad-band material to be located at the derived RBB for the first magnitudes (LU→Ks ). In both cases, our NIR data are data point at 0.84 days, the physical expansion speed crucial, as they provide a factor of two in wavelength for that material would have to be very high (∼ 40,000 coverage and leverage for the SEDs. While theoretical km s−1), if launched from the derived progenitor stellar considerations might justify using a simple BB approxi- 10 mation for the SN emission shortly after shock breakout radius (R⋆ ≈ 8×10 cm, see § 6.3). This estimate for the (Waxman et al. 2007), we do note that without detailed photospheric velocity is consistent with the line widths observed in the optical spectra at slightly later times (i.e., models of SN Ib/c atmospheres, it proves difficult to say −1 how well BB fits approximate the true SN luminosity 2 < ∆t < 5 days), which are ∼20,000−30,000 km s over the full light curve of SN 2008D. We address this (but see § 5.3). Thus, if indeed such fairly high ejecta question more in the rest of the section, when compar- velocities (∼0.13c) are involved at very early times, then ing the bolometric luminosities computed via BB fits and the BB emission, if it correctly describes the SED of the those via direct integration. nascent SN, is expected to be Doppler boosted and the opt We start by interpolating the light curves by actual temperature is slightly smaller, TBB/ [1+(v/c)] = Swift/UVOT (in UBV ), KAIT (in RI), FLW0 1.2 m 11,700 K/1.13 = 10,300 K at ∆t = 0.84 days. ′ ′ (in r i ), and PAIRITEL (in JHKs) onto a grid spaced We present the resultant BB luminosity evolution regularly in time, as the observing times were different in Figure 10 (right, top panel) and include that of for different telescopes. We then sampled those light SN 1999ex for comparison. Both SNe show the initial curves at intervals closest to the actual observing time dip due to shock breakout also in the bolometric lumi- of Swift/UVOT and the FLWO telescopes. After remov- nosity evolution, with SN 2008D showing a much less ing extinction (E(B − V )Host = 0.6 ± 0.1 mag and pronounced contrast between shock breakout and the assuming a Milky Way extinction law), we converted peak due to heating of 56Ni than in SN 1999ex. The our photometry into monochromatic fluxes at the effec- BB fit light curve for SN 2008D bears strong qualitative tive wavelengths of the broad-band filters, using zero- resemblance to those computed by Ensman & Woosley points from Fukugita et al. (1995) (for UBVr′i′, since (1988) for core-collapse SNe Ib resulting from hydrogen- the UVOT UBV photometric system is close to the John- free Wolf-Rayet star progenitors. In their models, the son UBV with Vega defining the zeropoint; Poole et al. first peak due to shock breakout is visible, while the ex- 2008) and Cohen et al. (2003) (for JHKs). We in- act time at which the dip occurs and the strength of the cluded 0.05 mag of systematic error in the UVOT fil- breakout peak depend in a complex fashion on the ki- ters (Li et al. 2006) and 4% conversion error when using netic energy, amount of 56Ni, and mass of progenitor in Fukugita et al. (1995). their models. 14

51 11 50/3 The second method for computing the bolometric light 1 ×10 (R⋆/10 cm) erg, based on our optical and output is by direct integration of the observed broad- NIR data from the first four epochs (0.8 < ∆t < 3.8 band photometry. Here we are in the favorable position days), where the Waxman model is valid. However, the of having NIR data, as the NIR contribution to the bolo- extreme sensitivity to the assumed radius makes these metric output is usually unknown for SNe Ib/c and can exact values extremely uncertain. be quite large (30−50%) for certain SNe (Tominaga et al. Alternatively, we can use our early-time light-curve 2005; Tomita et al. 2006; Modjaz 2007). fitting results to provide constraints on computing R⋆ We integrate the total optical and NIR monochromatic if the ratio Ek/Mej can be measured independently. fluxes in the region between the effective wavelengths of Analytically, we can rewrite equations [18] and [19] of the U and Ks filters via the trapezoid approximation. Waxman et al. (2007) such that For this purpose, we extrapolate the SED to zero at the ˚ ˚ − blue U (3000 A) and red Ks (24,210 A) edges of the R E 0.24 total set of filters. We plot the resultant bolometric light ⋆ 4 0.4 1.68 K,51 11 =0.286 TBB,ev RBB,14 ∆td , 10 cm Mej,⊙ curve (LU→Ks ) in Figure 10 (right, bottom panel) and, for comparison, also that of SN 1999ex (based on U → z (1) integration). where TBB,ev is the BB temperature in eV, RBB = 14 We find that the bolometric luminosity based on direct 10 RBB,14 cm, ∆td time in days after shock breakout, 51 U → Ks integration is quite similar to that based on the EK = 10 EK,51 erg, and Mej = Mej,⊙ M⊙. Thus, the BB fits (within . 0.05 dex or . 10%), except for the derived BB fits to the early-time data should, in prin- first three data point at ∆t = 0.8 days, 1.8 days, and 2.8 ciple, yield independent measurements of R⋆ for each days (different by a factor of 2.1, 1.5, and 1.2, respec- epoch, for a given Ek/Mej. tively). At ∆t = 0.84 days, the peak of the SED is at The ratio Ek/Mej can be measured independently wavelengths bluer than the observed wavelength cover- from the SN spectra (since v ∝ (E /M )1/2). Using ∼ ˚ ph k ej age (shorter than 4400 A for the B-band effective wave- the value estimated by Soderberg et al. (2008), namely length) and explains the missing flux in the direct inte- 17 −1 Ek/Mej = 4.0 × 10 erg g , and our Equation 1, we gration. It also explains the fact that LU→Ks has a less- compute R⋆ for each of the first four epochs (0.8 < ∆t< pronounced upturn at ∆t = 0.84 days than seen in LBB. 3.8 days), and for the range of adopted extinction valued The true bolometric luminosity (Lbol) at that epoch lies (E(B−V )Host =0.6 ± 0.1 mag). Taking an average over somewhere between the two values, since LU→Ks is the the epochs, we obtain, for each choice of (E(B−V )Host=( absolute lower limit and certainly underestimates Lbol. 10 0.5, 0.6, 0.7) mag, R⋆,0.5mag = (4.6 ± 2.0) × 10 cm, However, in general (for ∆t & 3 days) we find a 10 R⋆,0.6 mag = (7.9 ± 3.2) × 10 cm, and R⋆,0.7 mag = much better agreement between bolometric values de- 10 (14.3 ± 5.0) × 10 cm. Thus, our best estimate for R⋆ rived from BB fits and those based on direct integra- 10 is R = (8 ± 7) × 10 cm ≈ 0.1 − 2 R⊙ (where we tion (within ∼10%) than other studies of SNe Ib/c ⋆ have included the effect of the uncertain host galaxy ex- (20−30%; Stritzinger et al. 2002; Folatelli et al. 2006), rivaling the close agreement found for SN 1987A (within tinction). Given the very weak dependence of R⋆ on E /M , measurement uncertainties in E /M are ex- . 10%) which was observed across the optical and in- k ej k ej frared (U − M) bands (Suntzeff & Bouchet 1990). pected to not significantly influence the value of R⋆. We note that our estimate for R⋆ is consistent with that of 6.3. Soderberg et al. (2008), who obtain based on a similar Towards Constraining the Progenitor Size of 10 analysis, R ≈ 7 × 10 cm=1R⊙ (but do not ex- XRT 080109/SN 2008D ⋆,S08 plicitly take into account the effect of the uncertain host Now we compare our data to the predictions of the galaxy reddening). cooling BB envelope model of Waxman et al. (2007) as We can now directly compare our estimated R⋆ the source for the early-time emission, where the stel- of XRT 080109/SN 2008D to stellar radii of Wolf- lar envelope expands and cools after the passage and Rayet (WR) stars, which possess dense winds (see breakout of the shock wave. The time evolution of the Abbott & Conti 1987; Crowther 2007 for reviews), radius and the temperature of the cooling envelope are and are the proposed progenitors of stripped-envelope given in their equations [18] and [19] and are a function SNe, including that of XRT 080109/SN 2008D of, but do not depend sensitively on, the ejecta mass (Soderberg et al. 2008). Specifically, bonafide WN (Mej), ejecta kinetic energy (EK ), and radius of the pro- stars, i.e., the nitrogen-sequence spectral subtype of WR genitor before explosion (R⋆). In Figure 10 we overplot stars with their helium layer intact but without H (see the predicted values using the parameters put forth by Smith & Conti 2008 for the distinction between WN and 11 Soderberg et al. (2008) for SN 2008D, namely R⋆ = 10 WNH, i.e., H-rich WN stars), are the likely progen- 51 cm, Mej = 5 M⊙, and EK =2 × 10 erg (see their cap- itors of SNe Ib. Their radii are in the range 1−20 tion to Figure 3). Our values are consistent with their R⊙ (Herald et al. 2001; Hamann et al. 2006; Crowther predictions, within the uncertainties. We suggest that 2007). These estimates are based on recent advance- the simplifying assumption of Waxman et al. of the op- ments in computing WR model atmospheres that in- tical photons being on the Rayleigh-Jeans tail (see their clude line-blanketing effects by Fe-peak elements (e.g., Eq. [20]) might probably not be valid: the best-fit BB to Hillier & Miller 1998; Gräfener et al. 2002), and allow our optical-to-NIR data at ∆t =0.84 days has the rather for a more reliable determination of stellar parameters cool temperature of TBB ≈ 10,000−15,000 K. Using the (Hamann et al. 2006). Provided that those models and prescription from Waxman et al. (2007), we derive best- their data are accurate, our value of R⋆ ≈ 0.1−2 R⊙ is 11 50/4 fit values of Mej = 6 (R⋆/10 cm) M⊙ and EK = smaller by a factor of 5−10 than typical values for such 15

25 43.0 08D U−>Ks Blackbody Fit 99ex −18 20 S08 ] 42.5 −1

−17 [mag] K] 3 15 bol

[erg s 42.0 [10

BB −16 BB

T 10

log L 41.5 08D −15 Absolute M 5 99ex S08 41.0 −14 43.0 20 U−>Ks Integration

] −18

−1 42.5 15 [mag]

cm] −17 bol 14 [erg s 10 42.0

[10 −16 U−>Ks BB R 5 41.5 −15 log L Absolute M 0 41.0 −14 0.1 1.0 10.0 100.0 0.1 1.0 10.0 100.0 Days since Outburst Days since Outburst

Fig. 10.— Left: Temperature (top) and radius (bottom) derived from a BB fit to the optical and NIR SED of XRT 080109/SN 2008D (filled circles) as a function of time, assuming D = 31 ± 2 Mpc and corrected for reddening (E(B − V )Host = 0.6 ± 0.1 mag). The error bars include the uncertainties as reported from the formal fits, as well as the effect of uncertain extinction. For comparison, we plot the corresponding values obtained for SN 1999ex (open circles) by Stritzinger et al. (2002), which they derive from BB fits to their BVI data (their Table 8 and assuming their adopted values of E(B − V ) = 0.28 mag, distance = 51.16 Mpc, and explosion date of JD 2,451,480.5, as estimated from non-detection on that night, 1 night before the detection of SN 1999ex). We compare our data of SN 2008D to the predicted 11 values of the stellar cooling envelope model (dashed lines) in Soderberg et al. (2008, S08), using the same model parameters (R⋆ = 10 51 cm, Mej =5M⊙, EK = 2 ×10 erg), based on models of Waxman et al. (2007). The cooling envelope model is expected to only apply for ∆t . 3 − 4 days, before heating by 56Ni sets in. Their predictions are consistent with our data, within the uncertainties. Right: Bolometric light curves of XRT 080109/SN 2008D (filled circles) computed via two different techniques: fitting a BB to the reddening-corrected ′ ′ UBV Rr Ii JHKs fluxes (LBB, top) and via direct integration of the reddening-corrected broad-band photometry (LU→Ks , bottom). The error bars do not include the systematic uncertainty due to distance. The agreement between LBB and LU→Ks is very good (within ∼ 10% ), except for the first three data points at ∆t . 3 days, where direct integration underestimates the full radiative output, since the peak of the SED is at wavelengths shorter than the observed wavelength range. For comparison, we plot the same values obtained for SN 1999ex (open circles) by Stritzinger et al. (2002) as derived from their BB fits to their BVI data, and from direction integration of their UBV RIz data (their Table 8). The dashed line is the predicted BB light curve of SN 2008D using the model paramaters from Soderberg et al. (2008, S08), based on models of Waxman et al. (2007), as in the left panel. We note that, by definition, the value for the bolometric magnitude Mbol of SN 2008D will vary depending on the passband used over which different authors (S08, Malesani et al. 2008a) integrate its flux. WN radii, and it appears barely consistent with those of mag). While it is beyond the scope of our considerations early-type WN stars (e.g., radii of 1 − 2 R⊙ for WN4-s here to offer an alternative to the Waxman et al. model, and WN2-w types in Hamann et al. 2006).30 we note that our ability to assess the Waxman model is This result leaves us with two possible inter- hampered by the uncertain reddening, and, more impor- pretations: (a) the constrained progenitor size of tantly, by the paucity of early-time data. Our detailed XRT 080109/SN 2008D is inconsistent with what we analysis highlights the need for early-time (∆t < 24 expect for a typical WN star, and its progenitor was hours) broad-band data that cover such shock breakout actually much more compact, similar to an early- events, in order to render them useful for constraining the type WN star, or (b) if we assume that R⋆ of progenitor properties unambiguously. Independent from XRT 080109/SN 2008D has to be comparable to typi- the considerations above, Soderberg et al. (2008) infer cal WN radii, then the Waxman et al. description of a from their extensive radio observations that the progen- cooling stellar BB envelope might not be accurately de- itor had a steady wind with a mass loss rate consistent scribing the data. Underestimating the host extinction is with those of WR stars. probably not the dominant reason for the measured low 7. COMPARISON WITH OTHER SUPERNOVAE value of R⋆, since the additionally needed reddening for achieving R⋆ =10 R⊙ is too high (extra E(B − V ) ≈ 0.4 Here we investigate the properties of SN 2008D in the context of other SNe whose shock breakout has been ob- 30 Measurements of radii via spectroscopic WR binaries independently show the trend of smaller radii for early-type WN stars served in the hopes that we can learn about the respective (Cherepashchuk 1982). physical properties. The optical signature of shock breakout (UV excess 16

Fig. 11.— Comparison of the optical Swift/UVOT (filled circles) and ground-based (empty circles and triangles) light curves of SN 2008D with those of other normal SNe (left) and SNe associated with low-luminosity GRBs (right, but see § 9), for which shock breakout was observed or the time of explosion is well known. Data are from Richmond et al. (1994) for SN1993J, Stritzinger et al. (2002) for SN 1999ex, Galama et al. (1998) for SN 1998bw, and Campana et al. (2006) (UBV ) and Modjaz et al. (2006) (r′i′) for SN 2006aj. The offsets refer to the UBV RI light curves of SN 2008D, while its r′i′ light curves are offset by r′ − 2.6 and i′ − 4.3 to match RI. The rest of the SNe are offset to match the peak magnitude of SN 2008D in the respective bands. For SN 1998bw and SN 2006aj, the SN explosion date was taken as coincident with the onset of GRB980428 and XRF 060218, respectively (Galama et al. 1998; Campana et al. 2006). For SN 1993J and SN 1999ex, the explosion dates were estimated from optical non-detections on the night before (for SN 1999ex, JD 2,451,480.5; Stritzinger et al. 2002), and on the night during discovery (for SN 1993J, JD 2,449,074.5 ± 0.05; Wheeler et al. 1993). and dip in the first few days after explosion) has been SN 1999ex) reveals that while the light curves (and spec- observed in two stripped-envelope normal SNe: SN IIb tra) of SN 2008D and SN 1999ex are similar during the 1993J (Schmidt et al. 1993; Richmond et al. 1994) and 56Ni-decay phase, the light curves related to shock break- SN Ib 1999ex (Stritzinger et al. 2002). Other SNe with out are dramatically different among these three SNe. In early-time observations (at ∆t ∼1 day or earlier) in- particular, there are differences in when the “dip” be- clude the broad-lined SN Ic 2006aj connected with low- tween the two peaks occurs: the U-band dip in SN 2008D luminosity GRB/XRF 060218 (Campana et al. 2006, but occurs at ∆t ≈ 5 days, while that of SN 1999ex and see § 9), GRB 980425/SN 1998bw (for the V and R SN 1993J occur at ∆t ≈ 4 and 8 days, respectively. The filters, Galama et al. 1998), the famous SN II 1987A dip occurs the earliest in GRB 980425/SN 1998bw (∆t< (Hamuy et al. 1988), and most recently, two SN II 1 day), followed by XRF 060218/SN 2006aj (at ∆t ≈ 3 found by the Supernova Legacy Survey and detected days). While our sample sizes are very small, it appears with the UV-satellite GALEX (Schawinski et al. 2008; that the dip occurs earlier in broad-lined SNe Ic associ- Gezari et al. 2008). ated with GRBs/XRFs than in regular stripped-envelope SNe. In other words, the shock breakout component 7.1. Photometry: Shock Breakout and Radioactive in SN-GRBs appears to subside faster than in regular Decay as Power Sources stripped-envelope SNe, or alternatively, the 56Ni phase commences earlier in SN-GRBs presumably due to their Figure 11 compares the optical light curves of larger amount of synthesized nickel. SN 2008D to those of the other SNe with observed shock breakout, where the SNe are offset to match the Furthermore, we note differences in the dip-to-peak peak magnitude of SN 2008D in the respective bands. contrast amongst the shown SNe — i.e., the difference in magnitudes between the “dip” phase and the SN peak Careful comparison with other stripped-envelope SNe 56 caught shortly (∼ 1 day) after shock breakout (SN 1993J, powered by Ni. While the dip-to-peak contrast (CDP ) 17

SN 1999ex has a very red phase 1−5 days after explosion, where it rapidly changed color, while it resumed the typical SN color behavior after ∼ 8 days past explosion. We leave it as a future topic to model the diversity of the very early-time light curves of SNe 1993J, 1999ex, 2006aj, and 2008D in terms of progenitor radius, EK , and Mejecta, and to compare these values with those derived from other methods (e.g., light curve and nebular models). We note, however, that the temporal evolution of the inferred BB temperature and radius of SN 2008D are quite different from those derived for SN 1999ex (Fig. 10, left panel) using a similar analysis (BB fits to the SED, but only BV I for SN 1999ex) (Stritzinger et al. 2002). In SN 1999ex, TBB appears to decline steeply (for 1 < ∆ < 4 days, 12,000 K < TBB < 7,000 K) but then rise again (for 5 < ∆ < 10 days, 7,000 K < TBB < 10,000 K), explaining its peculiar color evolution (Fig. 13). Stritzinger et al. (2002) interpreted the declining temperature in SN 1999ex during the first four days (i.e., 1 < ∆t< 5 days) as indicating adiabatic cool- Fig. 12.— Same as in Fig. 11, but relative to date of V -band ing of the photosphere that followed shock breakout, and maximum for each SN. the rising temperature for ∆t> 6 days as heating via radioactive heating. In SN 2008D, however, the derived color temperature decreases continuously. If indeed the SED of SN 2008D can be well characterized by a BB (and we do not have good evidence for the contrary), then this observation indicates that the temperature of the photosphere does not change appreciably between the shock breakout and radioactive deposition phase. Comparing the light-curve shapes of the different SNe with respect to date of V -band maximum (Fig. 12), we find that SN 2008D declines more slowly than any of the other SNe and has a wider peak. One route of parame- terizing the SN light-curve shape is using the ∆m15(X) parameter, which is defined as the decline (in magnitudes) during the first 15 days after maximum brightness in the passband X (e.g., Phillips 1993). For SN 2008D, we measure ∆m15(U)=1.2 mag, ∆m15(B)=0.8 mag, and ∆m15(V )=0.6 mag.

7.2. Optical Spectroscopy: Emergence of He I Lines The very early spectra (∆t ≈ 3 − 5 days) actually re- semble those of SN 2006aj (see top two spectra in Fig- Fig. 13.— Comparison of the B − V color curve of SN 2008D ure 14), the SN connected with GRB/XRF 060218, which dereddened by E(B − V ) = 0.6 mag, with those of other SNe. The initially lead to the identification of this SN as a broad- B − V color of SN IIb 1993J (Richmond et al. 1994) is dereddened lined SN (Blondin et al. 2008b; Kong et al. 2008a) (but by E(B − V ) = 0.08 mag (we adopted the lower value derived by Richmond et al and not their higher alternative of E(B − V ) = see Malesani et al. 2008b). This spectral resemblance 0.28 mag), and that of SN Ib 1999ex by E(B − V ) = 0.3 mag suggests that in both systems, the material constitut- (Stritzinger et al. 2002). ing the photosphere at early times is at high velocity, as expected from homologous expansion. Indeed, the veloc- −1 is quite extreme in SN 1999ex (CDP (99ex)= 3.5 mag in ity we determined via SYNOW, vphot ∼30,000 km s at V ), the values for the other SNe cluster around 1 ± 0.2 ∆t =1.8 days in the spectrum with the double-absorption mag (CDP (93J) = 1.1 mag; CDP (06aj) = 0.8 mag), with feature (see §5.3) is similar to those of broad-lined SN Ic CDP (08D) = 1.2 mag in the V band (see also Fig. 12). at the same epoch measured in a slightly different man- Moreover, comparison with other stripped-envelope SNe ner (e.g, SN 2002ap, Mazzali et al. 2004). from the literature (Richardson et al. 2006) and from However, while SN 2006aj and other broad-lined SN Modjaz (2007) yields a lower limit to CDP of ∼ 1.7 mag Ic continue to have significant amount of mass at high −1 (as shock breakout, and therefore the beginning of the velocity (at least 0.1 M⊙ at 20,000 − 30,000 km s of dip, was not observed for those SN). the total of 2 M⊙ ejected mass as derived from models Furthermore, we note that the B − V color evolution in Mazzali et al. 2006), at ∆t ∼ 12 days, the receding of SN 2008D is very similar to that of SN 1999ex dur- photosphere in SN 2008D reveals the lower-velocity lay- ing the time of 56Ni-decay, but very different during the ers and the development of He I lines which SN 2006aj shock breakout phase (Figure 13). Unlike the other SNe, and other broad-lined SN Ic do not show (see bottom two 18 spectra in Figure 14). In particular, at ∆t ≈ 4 − 5 days, the broad features near 6000 A˚ and 7800 A˚ are at similar 10 velocities in both objects, but by ∆t = 12 days, the velocities of O Iλ 7774 are ∼ 20,000 km s−1 in SN 2006aj −1 8 ∆ and ∼ 10,000 km s in SN 2008D. 08D, t=3.74 d, tVmax=−14.7 d Unfortunately, no spectra shortly after explosion were obtained of SN 1999ex, the only other SN Ib photometri- 6 cally caught 1 day after shock breakout (Stritzinger et al. ∆ 06aj, t=4.95 d, tVmax=−5.0 d 2002; Hamuy et al. 2002), which could have been used for 4 comparison. On the other hand, such spectra exist for SN 1993J, the SN IIb that had retained some of its hydro-

+ Constant 2 ∼ ⊙ λ ∆ gen envelope ( 1 M , Elmhamdi et al. 2006). Spectra 08D, t=11.66 d, tVmax=−6.8 d obtained 2−3 days after explosion are nearly featureless (e.g., Filippenko et al. 1993; Baron et al. 1995), similar 0 HeI HeI HeI HeI to those of SN 2008D, and interestingly did not exhibit Scaled f strong Balmer lines, as would have been expected from −2 ∆ non-LTE models. 06aj, t=12.85 d, tVmax=+2.8 d No narrow emission lines of high-ionization species such as [Fe X], N V, or C IV are detected, nor −4 are narrow Hα and He II, as were seen in very early-time spectra (2−4 days after shock breakout) of SN 1993J (Garnavich & Ann 1994) and of SN II 2006bp 4000 5000 6000 7000 8000 9000 (Quimby et al. 2007), and interpreted as recombination Rest Wavelength (Å) of the circumstellar gas that had been ionized by the shock breakout (Garnavich & Ann 1994). The nonde- Fig. 14.— Comparison of the MMT and APO spectra of SN 2008D taken at ∆t = 3.74 days and 11.66 days, respectively, tections in SN 2008D may indicate a low-density circum- − with those of SN 2006aj at similar ages (∆t = 4.95 and 12.85 days, stellar environment (≪ 109 cm 3). Since high-density respectively). SN 2006aj was the SN associated with XRF 060218 lines such as [O III] λ 4363 are not present either, we (Modjaz et al. 2006). While the early-time spectra of SNe 20008D conclude that high density (larger than the respective and 2006aj are similar, indicating mass at high velocities for both systems, the spectrum of SN 2008D clearly develops moderately critical densities) is probably not the reason for the ab- wide He I lines at later times (bottom spectrum), while SN 2006aj sence of narrow CSM emission lines. continues to exhibit broad lines and no helium lines. As Figure 15 shows, the later spectral evolution of SN 2008D resembles that of a normal SN Ib, where the He I lines are as pronounced as in SN 2005hg and more 40 so than in SN 1999ex. 08D, t =−6.8 d, ∆t=11.66 d, dered [E(B−V)=0.6 mag] We now investigate the temporal behavior of the He I 30 Vmax lines, which are the spectral hallmarks of SNe Ib. In

SN 2008D and in a larger sample of SNe Ib, we deter- 20 He I HeI HeI HeI mine the blueshifts in the maximum absorption in the 05hg (Ib), t =−7 d He I lines via the robust technique of measuring SN 10 Vmax + Constant line profiles as presented by Blondin et al. (2006). This λ 99ex (Ib), t =−4.5 d, ∆t=16 d, dered [E(B−V)=0.3 mag] technique employs a spectral-smoothing algorithm that 0 Vmax incorporates the wavelength-dependent noise of ground- based spectra and reduces the impact of line overlap. We Scaled f −10 94I (Ic), t =−5 d, dered [E(B−V)=0.2 mag] measure the location of maximum absorption for the He I Vmax lines using their recommended procedure of spline inter- −20 polation and a smoothing factor of 0.002, and employ 4000 6000 8000 10000 the relativistic Doppler formula. We refer to our veloc- Rest Wavelength (Å) ity measurements as “He I absorption-line velocities” in the rest of this work. In Figure 16 (see also Table 8) we Fig. 15.— Comparison of the APO spectrum of SN 2008D taken I at 6.4 days before V maximum (i.e., ∆t = 11 days after shock present the measured values for He λ5876, the strongest breakout) with other early-time spectra of SNe Ib and Ic, dered- optical He I line in SN 2008D, along with absorption-line dened by the indicated amounts: SN 1999ex (Stritzinger et al. velocities of a sample of other SNe Ib (see Modjaz 2007), 2002; Hamuy et al. 2002), SN 2005hg (Modjaz 2007), and SN Ic 1994I (Filippenko et al. 1995). The optical series of He I lines, measured in the same fashion. I I defining SN 2008D as a SN Ib, are indicated (namely, He λ4471, In SN 2008D, the He lines become apparent as early He I λ5876, He I λ6678, and He I λ7065). We note the similarity as ∆t ≈ 6 days (i.e, tVmax ≈ −13 days) and are fully between SN 2008D and SNe 1999ex and 2005hg. ≈ ≈ − identified at ∆t 11.7 days (tVmax 6.9 days). As − − mentioned, SN 2005hg is spectroscopically similar to −14,000 km s 1 (for SN 2004gq) to ∼ −10,000 km s 1 SN 2008D and also exhibits a similar temporal He I line (for SN 2005hg), and seem to follow a power law decline. velocity evolution, with He I detected ∼12 days before its For comparison, Branch et al. (2002) showed that the V -band maximum (its explosion date is unknown). Fig- He I velocities in the SNe Ib of their limited sample fol- ure 16 reveals that the velocities span a large range for lowed a standard pattern and traced fairly closely (within normal SNe Ib; e.g., at maximum light, they range from 2000 km s−1 for the same epoch) the photospheric veloc- 19 ities derived from synthetic fits to Fe II lines, which we constraints for future theoretical investigations in SN overplot in Figure 15 (dashed line, using their power- Ib line formation via advanced radiative transfer codes −2/(n−1) law fit vphot ∝ t with n = 3.6). However, the such as CMFGEN (Dessart & Hillier 2007), PHOENIX SNe Ib in our sample show a larger departure (up to (Baron et al. 1999), and SEDONA (D. Kasen, private 6000 km s−1) from the corresponding standard photo- communication) that incorporate time-dependent, non- spheric velocity, larger than those of SN 1998dt, which LTE, and line-blanketing effects. Branch et al. (2002) had declared as the exception in Finally, we note that SN 2008D does not exhibit any of their sample. In addition, both SNe 2008D and 2005hg the peculiarities associated with the unique maverick SN show He I line velocities that are lower than the cor- Ib 2005bf, such as He I lines whose absorption-line veloc- responding photospheric velocity for dates before maxi- ities increase over time (see Fig. 16). Thus, any compar- mum light. This might mean that the line-emitting ma- isons with SN 2005bf or including exclusively SN 2005bf terial is located at radii smaller than the bulk of the in rate calculations of such X-ray transient events, as material that makes up the photosphere or that there done by Xu et al. (2008), should be regarded with cau- are optical-depth effects at play. tion. For SN 2005bf, the increasing He I line velocities might be due to nonthermal excitation of 56Ni, where the He I lines become optically thick at increasingly larger radii, and thus in higher-velocity material, since the de- Days after XRT 080109 Outburst position rate of 56Ni increases towards the outer layers ] −20 0 20 40 60 80 100 over time (Tominaga et al. 2005; Modjaz 2007). −1 −18 SN 2008D −16 SN 1998dt 7.3. NIR Spectroscopy as a Discriminant km s

3 SN 2004gq −14 SN 2005bf Figure 17 shows the three NIR spectra of SN 2008D SN 2005hg (taken with the IRTF) in which we detect ﬁve prominent −12 B02 blueshifted He I lines. We tabulate their velocities as −10 inferred from maximum absorption in Table 9. −8 −6 −4 He I 5876 Absorption Velocity [10 −50 0 50 100 Days since V−Band Maximum

Fig. 16.— Temporal velocity evolution of He I λ5876 in SN Ib: SN 1998dt (Matheson et al. 2001), SN 2005bf (Tominaga et al. 2005; Modjaz 2007), SN 2004gq and SN 2005hg (Modjaz 2007), and SN 2008D (this work). He I λ5876 is usually the strongest He I line. The absorption velocities decrease over time, except for the maverick SN Ib 2005bf (see text). The velocities span a large range for normal SNe Ib: at maximum light, they range from −14,000 km s−1 (for SN 2004gq) to ∼ −10,000 km s−1 (for SN 2005hg). The velocities for SN 2008D lie within the range of observed values, with the earliest measurements before maximum for a normal SN Ib. The thick dashed line is the functional form for the photospheric velocity as fit by Branch et al. (2002, B02) to SN Ib data, The thin dashed lines demarcate the extent in their uncertainty. Fig. 17.— See text for details. Three NIR spectra of SN 2008D taken with the IRTF. The dates of observation are indicated, along with SN ages (∆t) with respect to time of shock breakout (in days). The five detected It has been long suggested that the He I lines are due to He I lines are indicated by their rest wavelengths. The spectra are nonthermal excitation and ionization of gamma-ray pho- shown as both raw and Fourier-smoothed, and are normalized and 56 offset by arbitrary amounts for clarity. Note the strong presence tons produced during the radioactive decay of Ni and I I 56 of blueshifted He 1.0830 at ∼ 1.02−1.05 µm and blueshifted He Co (Harkness et al. 1987; Lucy 1991; Mazzali & Lucy 2.0581 µm at ∼ 1.97−1.98 µm. The broad absorption at 1.05 µm 1998). The fact that the gamma-ray photons need time probably includes an additional blend of other lines (mostly Fe II, to diffuse through the ejecta to reach the He layer has but also Mg II, C I and Si I). been invoked to explain the late emergence (after maximum light) of He I in some SNe Ib (e.g., Filippenko One of the outstanding questions in the study of 1997). A challenge for those who study the compli- stripped-envelope SNe is how much of the He envelope cated nonthermal excitation effects in SN line formation is present in the progenitor star at the time of explo- would be to self-consistently explain for SN 2008D its sion. While the presence of the optical He I lines is used early emergence of He I lines (∼ 5 − 6 days after explo- to define the class of helium-rich stripped-envelope SNe sion, and ∼ 12 − 13 before maximum light) by excitation (i.e., SNe Ib), there is a long-standing debate on the pres- of the relatively small amount of synthesized 56Ni (. ence of helium in SNe Ic as inferred from NIR spectra. 0.07 M⊙). Clearly, the densely time-sampled and early Determining unambiguously if indeed helium is present sequence of SN 2008D spectra will provide important in SNe normally classified as of type Ic (and in which 20

of the strong blueshifted He I 2.058 µm is very similar to the strong blueshifted He I 1.083 µm (see Fig. 18). For example, on 2008 Jan. 21.5 (∆t ≈ 12 days, and ∼ 6 days before maximum light), the velocities inferred from the maximum absorption of five detected NIR He I lines lie at ∼ 12,000-14,000 km s−1 and are all consistent with each other. These values are somewhat higher than those inferred from the optical lines at the same epoch (10,000 − 12,000 km s−1, Table 8), probably due to optical-depth effects, such that the NIR lines have an optical depth of unity (2/3) at larger radii, and therefore at higher velocities. In order to quantify how much He I is contributing to the absorption feature at ∼1.04 µm in SN 2008D, we attempted to fit the full extent of our optical and NIR spectra with SYNOW (Fisher et al. 1999). However, our attempts failed, as the current line list for SYNOW does not include the He I line at 2.058 µm and possibly other NIR lines of various ions. We do note, however, that we were able to reproduce the absorption line at 1.04 µm alone with a combination of Ca II,O I, Fe II, C I, C II, Si II, Ca I, and He I. We encourage implementation of updated NIR line lists in SYNOW to accommodate future spectral synthesis of NIR SN spectra. In conclusion, SN 2008D gives supporting evidence to the argument that the broad absorption at ∼1.04 µm is a poor diagnostic for helium, since it is most likely a blend of Fe II, C I, Ca II, and S I at the same wavelengths (Millard et al. 1999; Gerardy et al. 2004; Fig. 18.— The two main helium lines detected in NIR spectra Sauer et al. 2006) and He I in SN Ib. Rather, NIR spec- of SN 2008D in velocity space for two epochs: 2008 Jan. 21.51 I (∆t = 11.95 days; solid lines) and Jan. 28.51 (∆t = 18.95 days; tra of SNe Ic should be searched for He lines between dashed lines). The top plot shows the NIR spectra with respect to 1 and 2 µm to see if they are present and exhibit similar He I 1.083 µm as zero velocity, and the bottom one with respect line profiles. Thus, we recommend that in future search I to He 2.058 µm. The largest difference between the two lines is for helium in SNe Ic, both J-band and K-band spectra the effect of blending of other elements with the He I 1.083 µm line. The blueshifted He I 2.058 µm line is narrower and more should be obtained to test for the presence of the two well-defined than the blend around 1 µm, of which we take the red strong He I 1.083 µm and He I 2.058 µm lines. absorption trough to be due to He I 1.083 µm. Both He I lines have similar velocities based on the blueshifted maximum absorption 8. (see Table 9). The spectra are normalized to the local continuum. DOUBLE-PEAKED OXYGEN LINES & ASPHERICAL EXPLOSION GEOMETRY amount) has important implications for understanding In Figure 19 we present the Keck spectrum of which progenitor stars give rise to which kind of SN and SN 2008D taken at tVmax = 91.4 days (i.e., ∆t = 109 the exact mechanism of mass-loss. Despite the impor- days), on April 28, 2008. It shows nebular emission lines tance of NIR spectra, only one SN Ib has been monitored on top of some residual, underlying photospheric spec- spectroscopically in the NIR (SN 1999ex, Hamuy et al. trum. This is expected, as SNe appear to turn fully 2002), while another has been presented in unpublished transparent ∼ 100 − 200 days after maximum light, and form (SN 2001B, Gerardy et al. 2004). some SNe take up to 1 year to complete the transition Claims of significant helium as seen in NIR spectra of (Mazzali et al. 2004). Here, we are witnessing this tran- SNe Ic have been made for SN Ic 1994I (Filippenko et al. sition period for SN 2008D. Below we analyze the nebular 1995; Clocchiatti et al. 1996) and in the GRB-SN 1998bw emission lines of SN 2008D and show that they indicate associated with GRB980425 (Patat et al. 2001). The a torus- or ring-like structure for the oxygen-emitting re- claims rely on identifying a strong absorption feature at gion. ∼1.04 µm with He I 1.083 µm. At the same time, the ab- At sufficiently late times, the whole SN ejecta become sence of strong optical He lines in those SNe Ic is usually optically thin in the continuum; hence, spectra obtained attributed to a number of factors (e.g., level depopula- during this time period afford a deeper view into the core tion, blending, see Patat et al. 2001). of the explosion than spectra taken during the early pho- Comparison of the NIR spectra of proper SNe Ib tospheric phase, since the shapes of the emission lines 1999ex, 2001B, and 2008D with those of SNe Ic (see are in principle governed by the geometry and distri- Taubenberger et al. 2006 and references therein) shows bution of the emitting ejecta. In the absence of hy- that all SNe display a strong feature at ∼1.04 µm but drogen, oxygen is the primary coolant in the ejecta of only SNe Ib show additional absorption at 1.95−2.0 µm stripped-envelope SNe at late epochs when the gas is neu- due to blueshifted He I 2.058 µm, while all SN Ic do tral or at most singly ionized (Uomoto & Kirshner 1986; not. For the few SNe Ib in general (Gerardy et al. 2004) Fransson & Chevalier 1987), and when densities are suf- and for SN 2008D in particular, the absorption velocity ficiently low for forbidden lines to be the strongest ones 21

the strong Ca II NIR triplet. 3.5 The spectrum of SN 2008D bears strong resemblance SN 2008D (t = 91.4 d) Vmax Ca II NIR to that of SN 1997ef (Matheson et al. 2001) at compa- 3.0 [O I] λλ6300,6363 rable epochs (∼ 100 days after maximum light), which [Ca II] λλ7291,7324 Mazzali et al. (2004) were able to reproduce via spectral 2.5 O I λ7774 synthesis models as the superposition of photospheric and nebular components. While the overall distribu- 2.0 tion of absorption and emission lines is similar between + Constant

λ SN 2000D and SN 1997ef, the most prominent differ- 1.5 ence between the two SN is the presence of conspicuous double-peaked oxygen emission lines in SN 2008D. 1.0 Scaled f To study the line profiles of SN 2008D more closely, 0.5 we plot the main relatively unblended emission lines, [O I] λλ6300, 6363, O I λ7774, and [Ca II] λλ7291, 4000 5000 6000 7000 8000 9000 7324, in velocity space in Figure 20. The two oxygen Rest Wavelength (Å) lines show the same unusual double-peaked profile, while [Ca II] λλ7291, 7324 does not exhibit it. For both oxygen Fig. 19.— Spectrum of SN 2008D at 3 months after maximum lines, the redder peak is at zero velocity, and the trough −1 light (April 28, 2008). The main emission lines in the spectrum are between the two peaks is shifted by ∼ −800 km s marked, and their rest wavelengths indicated. We caution that the from zero velocity. The total separation between the line at 7300 A˚ could be a blend of [O II] λλ7319, 7330 and some −1 Fe lines, in addition to [Ca II] λλ7291, 7324, as marked. two peaks is ∼ 2000 km s for both the forbidden and the permitted oxygen lines. With the chosen zeropoint for [Ca II] λλ7291, 7324 (namely 7307.5 A,˚ the straight mean of the two calcium lines), its peak is also blueshifted 7 by ∼800 km s−1, but blending with other lines ([O II] SN 2008D λλ7319, 7330, Fe) could be affecting the line shape, and (t =91.4 d) Vmax the exact determination of zero velocity. 6 1.5 x f([Ca II] λλ7291,7324) While optical-depth effects sometimes generate double- peaked [O I] λλ6300, 6363 because of its doublet nature (most notably in SN 1987A; Spyromilio 1991; 5 Leibundgut et al. 1991; Li & McCray 1992), it is not the culprit in SN 2008D, since (a) the separation of the doublet (3000km s−1) is larger than the observed separation between the double peaks (2000 km s−1), and 4 λ 2 x f(O I 7774) (b) O I λ7774, a multiplet with very small separations

+ Constant − λ (at most 3 A,˚ i.e., 130 km s 1, smaller than our spectral resolution), shows the same double-peaked profile as 3 [O I] λλ6300, 6363 in SN 2008D. Scaled f Such double-peaked oxygen emission lines have recently been observed in a number of stripped- 2 f([O I] λλ6300,6363) envelope SNe (Mazzali et al. 2005; Maeda et al. 2007; Modjaz et al. 2008c; Maeda et al. 2008, S. Taubenberger et al. 2008, in prep.), and have been interpreted in 1 the framework of global asphericities, where the emitting oxygen-rich ejecta are situated in a ring- or torus- like structure viewed along the equatorial plane. Both −10000 −5000 0 5000 10000 −1 Modjaz et al. (2008c) and Maeda et al. (2008) suggest Velocity [km s ] that such global asphericities are common during core collapse. Maeda et al. (2008) specifically suggest that Fig. 20.— The main late-time emission lines of SN 2008D in their observed fraction of double-peaked profiles (40 ± velocity space in the rest frame of NGC 2770. The zeropoint for [O I] λλ6300, 6363 here is 6300 A˚ while that of [Ca II] λλ7291, 10%) is consistent with the hypothesis that all stripped 7324 is taken at 7307.5 A˚ (straight mean of the two calcium lines). core-collapse events are mildly aspherical, given their We caution that [Ca II] might be a blend. Zero velocity is indi- models in Maeda et al. (2006). cated by the dashed line. The two oxygen lines show the same conspicuous double-peaked profile, indicating that the two horns While other SNe IIb and Ib have been shown to ex- cannot be due to the doublet nature of [O I] λλ6300, 6363. In both hibit some small-scale structure in oxygen (Spyromilio cases, the redder peak is at zero velocity, and the trough between 1994; Matheson et al. 2000a; Sollerman et al. 1998; the two peaks is shifted by ∼ −800 km s−1 from zero velocity. II Elmhamdi et al. 2004), indicating some clumping in the With the chosen zeropoint for [Ca ] λλ7291, 7324, its peak is density distribution of oxygen probably due to instabil- also blueshifted by ∼800 km s−1, though blending could be affecting the line shape and the exact location of zero velocity. See text ities, the clear symmetric double peaks in SN 2008D for more details. are different and rather indicate a global scale for the I anisotropy. The mechanism giving rise to such a global in the spectrum. In SN 2008D we detect [O ] λλ6300, anisotropy could lie in the intrinsic explosion physics 6363 and O I λ7774 along with [Ca II] λλ7291, 7324 and 22 of core collapse, since each of the various theoretical 9. X-RAY INTERPRETATION core-collapse models, be it neutrino driven (Scheck et al. Now that we have established that the optical super- 2006), acoustic (Burrows et al. 2006), or magneto driven nova properties of SN 2008D lie within the observed (Dessart et al. 2008), finds that symmetry breaking is es- range of those of SNe Ib, we have have to address the sential for a successful explosion. However, since the ob- question of origin of the X-ray emission associated with served spectra are a one-dimensional projection of an in- it. As mentioned, the Swift X-ray spectrum can be either herently three-dimensional density distribution, any pro- fit by an absorbed power law or an unabsorbed BB spec- posed solution or model for the SN geometry is not nec- trum, with the power-law model having slightly higher essarily unique. Alternatively, in case the progenitor of statistical significance. Furthermore, it is possible that SN 2008D is part of a binary, binary interaction or a the X-ray spectrum of XRT 080109 is a combination of merger might be modulating the geometry of the explo- both, similarly to that of the peculiar XRF 060218, which sion, though this speculation will be very hard to verify. could be clearly decomposed into equal parts of power Nevertheless, other lines of evidence, such as polarization law and BB components since it had a very high S/N. studies of SNe II, Ib, and Ic (e.g., Leonard & Filippenko 2005), neutron-star kick velocities (Wang et al. 2006), To attempt to break the degeneracy between the var- and young SN remnant morphologies (Fesen et al. 2006) ious emission models, we turn to additional considerations. The power-law fit implies an additional H- strongly suggest that asphericities are generic to the core- +2.1 21 −2 collapse process. Future studies and larger sample sizes equivalent column density NH = 5.2−1.8 × 10 cm will be able to probe if the degree of asphericity is cor- which, given standard conversions between NH and red- related with any other SN property (such as luminosity) dening (Predehl & Schmitt 1995), would imply AV = or with any environmental property (e.g., metallicity). 2.9±1.0 mag. This range is consistent with the AV ≈ 1.8 mag estimated based on the optical data of SN 2008D— We note the slight, but apparent blueshift from zero −1 i.e., the fact that the colors (including B−H) and spectra velocity (of the order ∼ 800 km s , Figure 20) in the of SN 2008D match very well those of other well-known oxygen trough, if the trough between the two peaks is SNe Ib after dereddening. regarded as the symmetry center. The same blueshift The BB fit does not require absorption in addition to might be also present in the calcium peak, depending that expected from our Galaxy (N < 1.5×1021 cm−2 at on the choice of zero velocity, which is more uncertain H 90% confidence). At face value, this seems inconsistent than for the oxygen lines due to possible blending with with optical observations of SN 2008D that indicate ab- other lines. This blueshift, if real, could be due to a sorption along the line of sight, unless the dust-to-metals number of reasons: (a) residual opacity before reach- ratio is higher in the host galaxy of SN 2008D. Never- ing the full nebular phase (since our spectrum was taken ≈ at t ≪ 200 days, when other SN spectra taken at theless, the BB fit is still consistent with AV 1 mag Vmax at 90% confidence. Thus, the N considerations for dis- similar epochs show similar bulk blueshifts; S. Tauben- H tinguishing between the two models are not necessarily berger et al., 2008, in prep.), (b) peculiar distribution compelling. of 56Co which is powering the emission at such later epochs (as suggested for SN 1993J, Spyromilio 1994), (c) As the next consideration, we turn to the size of the derived BB radius. The pure BB fit implies that the large-scale asymmetry (as suggested for SN Ib 1996N, X 9 Sollerman et al. 1998), or (d) offset of the SN rest-frame emitting radius must be very small, RBB ≈ 10 cm (see −1 also Li (2008)); this is ∼ 1−2% of the solar radius and velocity from that of NGC 2770 by 800 km s . How- ∼ 100 times smaller than our estimated progenitor stel- ever, given that the rotation curve of the host galaxy, lar radius ( R = (8 ± 7) × 1010 cm, see § 6.3). The NGC 2770, extends to only ∼400 km s−1(Haynes et al. ⋆ apparently small RBB could mean that the emitting sur- 1997), option (d) is less likely. face had a small filling factor (i.e., the radius was much Future multi-epoch late-time spectra (especially those larger), but the X-rays may have been from thin fila- taken tVmax & 200 days) are needed to differentiate be- ments. A similar case has been observed for SN 1987A, tween the various scenarios as each one predicts a dif- where the X-rays originate from the forward and reverse ferent behavior of the blueshift over time. If it is due shocks. Alternatively, the emitting material might oc- to residual opacity, the most probable case, we predict cupy much less than a sphere, in case the BB emission that the blueshift will go to zero for spectra tVmax & 200 is beamed. While a pure thermal X-ray spectrum is pre- days, while if dust formation is the culprit, the apparent dicted by breakout models (e.g., Matzner & McKee 1999; blueshift should increase over time. If the blueshift re- Calzavara & Matzner 2004), special conditions or mech- mains at a constant value over time (as seen in SN 1996N anisms would have to be invoked in this case to explain Sollerman et al. 1998), then bulk blueshifts or unipolar the unphysically small radius of the BB emitting area. blob ejections might need to be invoked as suggested for On the other hand, the combined fit (§ 2.2) of BB SN 2005bf (Maeda et al. 2007). and power law to the X-ray spectrum of XRT 080109 In summary, we infer from the presence of double peaks (with a ratio of ∼1, as in XRF 060218/SN 2006aj) in the two different oxygen lines in SN 2008D that global X 11 predicts RBB ≈ 10 cm, which is much more phys- asphericities were probably involved during the explo- ically plausible (and consistent with our derived value sive event of SN 2008D, regardless of the specific mech- of R⋆), and a NH that is consistent with the optical anism that actually leads to a successful explosion. We observations. While the data on XRT 080109 are con- note that this aspect of global anisotropy will influence sistent with such a combined fit, it is not clear what the interpretation and modeling of SN 2008D, such that the dominant underlying mechanisms are that give rise spherically symmetric models are probably not sufficient 56 to the two-component model that the data appear to to derive the physical SN parameters (e.g., Mej, Ni). 23 favor. Even for XRF 060218/SN 2006aj, which had inherent distribution will need to carefully consider de- a much higher S/N spectrum, the exact origin of the tector effects and rest-frame properties. X-ray emission is still debated: the suggested models 1) Energy output: The prompt high-energy emission 45 range from pure shock breakout for the thermal com- released in XRT 080109 (EX,iso ≈ 6 × 10 erg) is 2−3 ponent (Campana et al. 2006), to emission following the orders of magnitude lower than that in the peculiar late dissipation of the fireball bulk kinetic energy in a XRF 060218 (Campana et al. 2006; Butler et al. 2006), spherical system (Ghisellini et al. 2007), emission from and 3−4 orders of magnitude lower than in XRFs in gen- a nascent magnetar (Soderberg et al. 2006), and to fi- eral. nally explaining both thermal and power-law compo- 2) Duration: XRFs observed by HETE-2 and Swift are nents with a model in which a shock propagates out- seen to have similar temporal profiles and durations com- ward from a compact progenitor star into a dense wind pared to long gamma-ray bursts, exhibiting the common (Waxman et al. 2007). FRED profile. XRT 080109 can be adequately fit by such With these caveats in mind, we proceed to comparing a FRED profile (§ 9.1), but the transient’s duration is an the X-ray properties of XRF 060218 and of XRT 080109. order of magnitude longer than the median T90 for other We find that these two X-ray events actually appear XRFs and X-ray-rich GRBs (XRRs), which is roughly quite similar: (1) their X-ray spectra are comparable, 30 s (Sakamoto et al. 2005, 2008). Both XRT 080109 while that of XRT 080109 is at much lower S/N; and and XRF 060218/SN 2006aj have the longest durations (2) the X-ray light curves are similar if one allows for (T90 = 470 ± 30 sec and 2100 ±100 sec, respectively, stretching (up to factor of 0.1). Analogously, if we use Fig. 21). We note the caveat that satellites are typi- the prescription for computing the shock breakout ra- cally inhibited from triggering on long timescales due to dius by Waxman et al. (2007) (their equation [4]), with background modulations and XRT 080109 could not have values from our combined power law plus BB fit for been detected with any other survey X-ray telescope be- XRT 080109/SN 2008D (Table 1), we obtain a shock fore Swift, and the published T90 for the sample of XRFs 11 breakout radius of Rshock =5×10 cm, for kT =0.1 keV in Sakamoto et al. (2005, 2008) are in the observer frame. 45 X and Ethermal =6.7 × 10 erg. Also, since Rshock > RBB, we similarly conclude and speculate that the expansion of the shock might have happened in an aspherical man- ner, which meshes with our observations that the SN core ejecta themselves were not spherical (§ 8). However, XRT 080109 is significantly less X-ray luminous than XRF 060218 (by a factor of ∼ 700 in peak 3 luminosity and by a factor of ∼ 10 in EX ). We note that as a different scenario, Soderberg et al. (2008) adopt the pure power-law fit to the X-ray spectrum of XRT 080109 . They furthermore adopt a shock breakout origin for the X-ray power-law spectrum where the nonthermal component is caused by upscattering of thermal shock breakout photons (i.e., Comptonization), as described by Wang et al. (2007) and Waxman et al. (2007), and derive similar, but slightly larger values for & × 11 Rshock,S08 7 10 cm. 5 9.1. Comparison with X-ray Flashes 4 From a high-energy perspective, it is of interest to know where the prompt event fits in with respect to other 3 known classes. The closest association might be with X- ray flashes (Heise et al. 2001) categorized by relatively

short durations, nonthermal emission, and peak energies Number 2 XRT 080109/ Epeak < 25 keV, and believed to be softer analogues of SN 2008D cosmological GRBs (Sakamoto et al. 2008). Given the 1 importance of a physical context, already there has been some debate about whether XRT 080109 belongs in such 0 a class (Xu et al. 2008; Soderberg et al. 2008; Li 2008). −2 −1 0 1 2 3 4 Here we compare four properties of XRT 080109 with log f those observed in a sample of classical XRFs: duration O/X

(as parameterized by T90), energy output, the shape of Fig. 21.— Top: Comparison of the observed T90 of XRT 080109 the X-ray spectrum, and X-ray-to-optical flux ratio. with those of XRFs and XRRs observed by Swift (shaded areas) While in each case XRT 080109 is not necessarily and by HETE-2 (clear areas) (Sakamoto et al. 2005, 2008). Bot- tom: Comparison of the observed f (i.e, ratio between opti- an obvious outlier in each observed distribution, to- O/X × 2 − 3 gether all four considerations paint a strong picture that cal and X-ray fluxes) of XRT 080109 (fO/X = 4 (10 10 )) with those of XRFs and XRRs observed by Swift and BeppoSAX XRT 080109, along with XRF 060218/SN 2006aj, is (D’Alessio et al. 2006). The plotted values include also upper lim- different phenomenologically from the rest of classical its on fO/X for the cosmological XRFs and XRRs, as denoted as XRFs. However, future studies trying to uncover the left-hand arrows. 24

3) X-ray spectrum: Both XRT 080109 and XRF 060218/SN 2006aj, appears qualitatively dif- XRF 060218/SN 2006aj have a significant X-ray compo- ferent from the observed distribution of classical XRFs, nent that (a) is thermal and (b) appears to soften over when considering all four discussed properties together time, with very high S/N for XRF 060218/SN 2006aj (X-ray luminosity, duration, spectral evolution, and (Campana et al. 2006; Butler 2007). The late X- optical-to-X-ray ratio). ray emission of XRF 060218 fades at a rate compa- While at broad brush, the nominal progenitors of rable to that often observed for GRB X-ray after- XRFs and SNe Ib are similar, it is difficult to connect glows. However, the spectrum is markedly softer (e.g., XRT 080109 as part of the same continuum of XRF prop- Soderberg et al. 2006) and also softens strongly in time erties, since one possible scaling relation is already vio- (Butler 2007), whereas GRB X-ray afterglows exhibit lated: while one might expect duration to inversely cor- little spectral evolution at the same epoch: GRB and relate with X-ray luminosity based on XRT 080109 and XRF X-ray afterglows have photon indices of Γ ≈ classical XRFs alone, XRF 060218/SN 2006aj does not 2, while that of XRF 060218/SN 2006aj was Γ = scale accordingly (it is more luminous than XRT 080109, 3.41 ± +0.13 (Butler & Kocevski 2007b). There is ev- but also has a longer duration). In any case, future de- idence that the XRT 080109 afterglow spectrum (from tections and studies that include detection thresholds, +1.7 the Chandra data) was also softer with Γ = 3.6−1.4 are expected to show unambiguously whether these two (Pooley & Soderberg 2008) than the early spectrum. events are part of a continuous distribution of XRFs, or 4) We also compare the radiation output in members of a different physical class of events. XRT 080109 for different wavelengths to that 9.2. of XRFs and their afterglows (Fig. 21). For Rate Analysis: Detecting SN Shock Breakout XRT 080109/SN 2008D, the ratio between the optical during X-ray Missions 3 4 bolometric and X-ray luminosity is Lopt/LX ≈ 10 − 10 Since SN 2008D appears to be a spectroscopically at 70 ksec. This corresponds to a ratio of fluxes “normal” SN Ib, one could expect similar X-ray events 2 3 fO/X = 4 × (10 − 10 ), while the observed ratio is from some or all of the SNe in this class. These might fO/X <70 for classical XRF afterglows (D’Alessio et al. have been detected in X-ray transient searches, and 2006) and GRB afterglows (De Pasquale et al. 2003) at may place constraints on the rate of such events. The comparable times (using the same definition and units Ariel 5 and HEAO experiments (Pye & McHardy 1983; of fO/X as in D’Alessio et al. 2006). In other words, the Ambruster & Wood 1986; Connors et al. 1986) have con- optical to X-ray ratio for XRT 080109/SN 2008D ducted searches for fast X-ray transients, and have found is ∼ 103 larger than for XRF afterglows. For tens of sources that indicate global rates of order thou- XRF 060218/SN 2006aj, the ratio cannot be pre- sands per year. However, the detection thresholds of cisely determined since it was not observed in the these experiments were about 20 − 100 times higher desired bandpass at such early time. than the fluence of the X-ray transient associated with The two main features commonly observed in XRFs SN 2008D, which means that they would have detected such flares only to a distance of a few Mpc. Since the rate are not measured for XRT 080109, namely detection of −1 gamma-ray emission and peak of the energy distribution of core-collapse SNe is of the order of 0.01 SN year in (E ). such a small volume (Cappellaro et al. 1999, J. Leaman peak et al., in prep.), we conclude that none of the sources dis- While no gamma rays were detected from XRT 080109, covered by those searches is such a shock breakout X-ray we cannot rule out that the nondetection of gamma-ray transient, which does not allow us to place constraints emission could be due to insufficient detector sensitiv- on the nature of the source, or its prevalence. ity. The nondetection of a gamma-ray counterpart by the Swift/BAT instrument during pointed observations Looking into the future, the proposed sensitive X- places a 3σ upper limit of 8 × 10−8 erg cm−2 in the ray all-sky monitoring mission ARGOS-X (R. Remillard, 2008, private communication), with a fluence threshold 15−150 keV band (Page et al. 2008). For XRT 080109, −11 −2 −1 we find from extrapolating the power-law fit (based on of ∼ 5×10 erg cm s (for 1.5−25 keV), should detect such X-ray shock breakout events at a predicted rate the observed Swift X-ray spectrum, see § 2.2) to the BAT 3 −1 +5.8 −8 of ∼ 10 yr with its 60% sky cover-coverage (assuming range a fluence of S(15 − 150 keV)ex = 3.9−2.4 × 10 −2 the local SN Ib/c rate). This value is broadly consistent erg cm . Thus, BAT would not have seen any signifi- with an independent estimate of such X-ray transients cant gamma-ray emission that might have accompanied due to shock breakout by Soderberg et al. (2008). the prompt X-ray phase if the power-law behavior continues into the BAT range. We remain cautious about the extent to which the observed X-ray emission of XRT 080109/SN 2008D The spectrum of XRT 080109 measured across the is typical for all SNe Ib or SNe Ic, since, besides Swift/XRT range (0.3−10 keV) does not indicate a peak XRT 080109/SN 2008D, we have only one other such (see § 2.2). While Li (2008) conclude from a similar X- observed case (the broad-lined SN Ic 2006aj associated ray spectral fit for XRT 080109 that its E has to lie peak with the peculiar XRF 060218). somewhere between the Swift/XRT and the optical/UV If we use the geometry of the explosion as a guide and ranges, we note that the existence of Epeak in this case is a conjecture which relies on the assumption that the assume that only SNe with double-peaked oxygen profiles prompt radiation output in XRT 080109 is due to syn- (see § 8) give rise to such X-ray transients, then the rate of such X-ray shock breakout events slightly reduced to chrotron radiation whose characteristic frequency is be- 2 3 −1 low the X-ray and above the optical range. ∼ 10 − 10 yr . In summary, XRT 080109, along with In any case, future sensitive all-sky X-ray surveys coupled with early-time optical follow-up observations will 25 certainly answer the question whether XRT 080109 is a radius if EK /Mej is well determined. Our optical and maverick or a regular amongst SNe with observed early- NIR data allow us to construct a reliable measurement time X-ray emission. Clearly, if XRT 080109 turns out of the bolometric output for this stripped-envelope SN, to be a peculiar event among SNe Ib, then our inferred and, in conjunction with the value for EK /Mej as mea- rate would need to be revised. sured by Soderberg et al. (2008), to estimate the stellar radius of its progenitor, R⋆. We measure R⋆ 0.1–2 R⊙, 10. CONCLUSIONS which is much smaller than typical radii of WN stars, and marginally consistent with those of early-type WN We have presented extensive early-time photometric stars. (UBV Rr′Ii′JHK ) and spectroscopic (optical and NIR) s We find that the Swift X-ray spectrum can be fit data on SN 2008D as well as X-ray data on the associated equally well by an absorbed power law or an unab- XRT 080109. Our data span a time range of 2 hours to sorbed BB and propose a superposition of both, which 109 days after the detection of the X-ray transient and has not been suggested before in the literature. If we establish that SN 2008D is a spectroscopically normal concentrate on the thermal X-ray component and follow SN Ib at maximum light, with a relatively long rise time Waxman et al. (2007), we derive a shock breakout radius (18 days) and a modest optical peak luminosity of M = V that is ∼ 5 − 10 times larger than the progenitor radius. −17.0 ± 0.3 mag and M = −16.3 ± 0.4 mag, after cor- B XRT 080109 is different from the apparent distribution of recting for E(B − V ) = 0.6 ± 0.1 mag. Using our Host classical X-ray flashes as evidenced by its long duration, full set of optical to NIR data, we are able to carefully low X-ray luminosity, softening of the X-ray spectrum, estimate the (large) extinction SN 2008D suffered. and very high optical to X-ray luminosity ratio, though The light curves of the SN show the characteristic future studies on the inherent distribution will need to shape of shock breakout (∆t ≈ 0.1 − 5 days, most pro- 56 fold in detector effects and consider rest-frame proper- nounced in the blue bands) and Ni decay (∆t> 5 days, ties. In any case, future detections are expected to show see also Soderberg et al. 2008). Careful comparison with unambiguously whether these two events are part of a other stripped-envelope SNe caught shortly (∼ 1 day) af- continuous distribution of XRFs, or members of a dif- ter shock breakout indicate that there is diversity in the ferent physical class of events. Assuming XRT 080109 progenitor properties that set the shock breakout behav- has a typical X-ray luminosity for SN Ib shock breakout ior. From our small sample, we find that the shock break- (something about which we remain cautious), we predict out component in SN-GRBs appears to subside faster 100−1000 future detections of such X-ray events per year than in regular stripped-envelope SNe. We also present by sensitive all-sky X-ray telescopes. the earliest optical (BV r′i′) and NIR (JHK ) detec- s We further strongly encourage multi-epoch late-time tions of shock breakout (at 0.84 and 0.71 days), pro- observations (especially for more than 200 days after viding powerful constraints for models of the breakout maximum light) of SN 2008D as it becomes fully trans- emission mechanism. Our densely time-sampled optical parent, to monitor the behavior of the double-peaked spectral sequence, along with our NIR spectra, uncovers oxygen lines and their associated possible blueshift, in the emergence epoch of the helium lines. The early-time order to closely constrain the geometry of the explosion spectra reveal high expansion velocities (30,0000−15,000 −1 as revealed by the core ejecta. Moreover, such obser- km s ) up to 6−7 days after outburst, and one specific vations would be important to monitor for any late-time spectrum (at ∆t =1.84 days) shows a double-absorption interactions between SN ejecta and the presumably prior- feature, that is no longer visible in spectra taking 1 day shed hydrogen envelope to further study the exact mode later. In a spectrum taken 3 months after maximum of mass loss and search for episodes of sudden mass loss. light, we detect double-peaked profiles in the oxygen lines On a philosophical note, XRT 080109/SN 2008D beau- that indicate global asphericities in the SN ejecta, and tifully illustrates the importance of serendipitous discov- argue that they cannot be due to optical-depth effects. ery in science. Swift/XRT caught the X-ray signature Furthermore, we recommend for future searches for he- of shock breakout because it was monitoring the SN Ib lium in SNe Ic that J- and K-band spectra are obtained I I 2007uy in the same nearby galaxy, and two independent to test for the presence of both He 1.083 µm and He groups alerted the rest of the community of their dis- 2.058 µm. covery; our ground-based telescopes provided the crucial We note that Soderberg et al. (2008) presented a simi- earliest optical and NIR detections because, similarly, lar and independent analysis on XRT 080109/SN 2008D they were monitoring SN 2007uy as part of our SN ob- using the same Swift data and independent optical data. serving campaign. A unique aspect of our work is the very early-time (< 1 day after outburst) optical and NIR data on ACKNOWLEDGMENTS SN 2008D, and the late-time spectra in which SN 2008D turns transparent and reveals double-peaked oxygen MM thanks Roger Chevalier, Phil Chang, Chris lines. The early-time data allow us to test the predic- Matzner, Peter Nugent, Alceste Bonanos, Andrew Mac- tions of the cooling stellar envelope blackbody model of Fadyen, and Tsvi Piran for stimulating discussions and Soderberg et al. (2008) for SN 2008D. Those light curves acknowledges the Miller Institute for Basic Research in are based on calculations by Waxman et al. (2007) for a Science for support. We thank S. S. Piatek, and T. Pryor specific progenitor stellar radius, ejecta mass, and ki- for helping to obtain some of the spectra presented here. netic energy, and they are consistent with our early-time Some of the observations reported here were obtained at data, given the parameters in Soderberg et al. (2008). the MMT Observatory, a joint facility of the Smithsonian Furthermore, we provide a simple equation (based on Institution and the University of Arizona, and at the W. Waxman et al. 2007) to compute the stellar progenitor M. Keck Observatory, which was made possible by the 26 generous financial support of the W. M. Keck Founda- TABASGO Foundation and NSF grant AST–0607485 to tion. We wish to extend special thanks to those of Hawai- AVF. Supernova research at Harvard University is sup- ian ancestry on whose sacred mountain we are privileged ported by NSF grant AST 06–06772 to RPK. DP and to be guests. The Apache Point Observatory 3.5-m tele- NRB were partially supported through the DOE Sci- scope, which is owned and operated by the Astrophysical DAC grant DE–FC02–06ER41453. JSB was partially Research Consortium, was used to secure the spectra ob- supported by the Hellman Faculty Fund. JXP is par- tained by GSS reported herein. This work is based on ob- tially supported by NASA/Swift grants NNG06GJ07G servations obtained at the Gemini Observatory, which is and NNX07AE94G and an NSF CAREER grant (AST– operated by the Association of Universities for Research 0548180). Work by WHDV and SSO was performed in Astronomy, Inc., under a cooperative agreement with under the auspices of the U.S. Department of Energy the NSF on behalf of the Gemini partnership: the Na- by Lawrence Livermore National Laboratory under Con- tional Science Foundation (US), the Particle Physics and tract DE–AC52–07NA27344. GSS wishes to thank M. Astronomy Research Council (UK), the National Re- Drosback, B. Keeney, M. Moe, and K. O’Malia for relin- search Council (Canada), CONICYT (Chile), the Aus- quishing CU time, S. Hawley for awarding APO Direc- tralian Research Council (Australia), CNPq (Brazil), and tor’s Discretionary Time, R. McMillan for “a little more CONICET (Argentina). PAIRITEL is operated by the time” to complete the data acquisition on one night, and Smithsonian Astrophysical Observatory (SAO) and was the APO Observing Specialists for their assistance. GSS made possible by a grant from the Harvard University is partially supported by NASA grants NNX06AG43G Milton Fund, a camera loan from the University of Vir- and NNX06AG44G, and BPH and GDI are partially sup- ginia, and continued support of the SAO and UC Berke- ported by NASA grant NAG5–7697. CHB acknowledges ley. The PAIRITEL project are further supported by support from the Harvard Origins of Life Initiative. HM NASA/Swift Guest Investigator grant NNG06GH50G. and PG thank Alan Tokunaga for arranging the IRTF KAIT was made possible by generous donations from observations on short notice. Sun Microsystems, Inc., the Hewlett Packard Company, This research has made use of NASA’s Astrophysics AutoScope Corporation, Lick Observatory, the NSF, the Data System Bibliographic Services (ADS) and the University of California, and the Sylvia and Jim Katz- NASA/IPAC Extragalactic Database (NED) which is op- man Foundation; we also thank the TABASGO Founda- erated by the Jet Propulsion Laboratory, California In- tion for continued support. HM is a visiting astronomer stitute of Technology, under contract with the National at the Infrared Telescope Facility, which is operated by Aeronautics and Space Administration. We thank the the University of Hawaii under Cooperative Agreement Swift team and the observers who provided their data no. NCC 5–538 with the National Aeronautics and Space and analysis through the GCN. We collectively thank Administration, Science Mission Directorate, Planetary the staffs of the MMT, APO, Keck, Lick, and Gemini Astronomy Program. Observatories for their assistance. SN research at UC Berkeley is supported in part by the

REFERENCES Abbott, D. C. & Conti, P. S. 1987, ARA&A, 25, 113 Cohen, M., Wheaton, W. A., & Megeath, S. T. 2003, AJ, 126, 1090 Alard, C. 2000, A&AS, 144, 363 Connors, A., Serlemitsos, P. J., & Swank, J. H. 1986, ApJ, 303, Alard, C. & Lupton, R. H. 1998, ApJ, 503, 325 769 Ambruster, C. W. & Wood, K. S. 1986, ApJ, 311, 258 Crowther, P. A. 2007, ARA&A, 45, 177 Anders, E. & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, Cushing, M. C., Vacca, W. D., & Rayner, J. T. 2004, PASP, 116, 197 362 Baron, E., Branch, D., Hauschildt, P. H., Filippenko, A. V., & D’Alessio, V., Piro, L., & Rossi, E. M. 2006, A&A, 460, 653 Kirshner, R. P. 1999, ApJ, 527, 739 De Pasquale, M., et al. 2003, ApJ, 592, 1018 Baron, E., et al. 1995, ApJ, 441, 170 Dessart, L., Burrows, A., Livne, E., & Ott, C. D. 2008, ApJ, 673, Berger, E. & Soderberg, A. 2008, GCN Circ. 7159 L43 Blondin, S., Matheson, T., Modjaz, M., & Berlind, P. 2008a, Dessart, L. & Hillier, J. 2007, To appear in Conf. Proc. of Central Bureau Electronic Telegrams, 1205, 1 ”The Multicoloured Landscape of Compact Objects and their —. 2008b, Central Bureau Electronic Telegrams, 1205, 1 Explosive Progenitors”, July 2006 (astro-ph/0610136) Blondin, S. et al. 2006, AJ, 131, 1648 Dickey, J. M. & Lockman, F. J. 1990, ARA&A, 28, 215 —. 2008c, IAU Circ., 8908, 2 Elmhamdi, A., Danziger, I. J., Branch, D., Leibundgut, B., Baron, Bloom, J. S., et al. 2006, in ASP Conf. Ser. 351: Astronomical Data E., & Kirshner, R. P. 2006, A&A, 450, 305 Analysis Software and Systems XV, ed. C. Gabriel, C. Arviset, Elmhamdi, A., Danziger, I. J., Cappellaro, E., Della Valle, M., D. Ponz, & S. Enrique, 751–+ Gouiﬀes, C., Phillips, M. M., & Turatto, M. 2004, A&A, 426, Branch, D., et al. 2002, ApJ, 566, 1005 963 Burrows, A., Livne, E., Dessart, L., Ott, C. D., & Murphy, J. 2006, Ensman, L. M. & Woosley, S. E. 1988, ApJ, 333, 754 ApJ, 640, 878 Faber, S. M., et al. 2003, in Presented at the Society of Photo- Burrows, D. N., et al. 2005, Space Science Reviews, 120, 165 Optical Instrumentation Engineers (SPIE) Conference, Vol. Butler, N. R. 2007, ApJ, 656, 1001 4841, Instrument Design and Performance for Optical/Infrared Butler, N. R. & Kocevski, D. 2007a, ApJ, 663, 407 Ground-based Telescopes. Edited by Iye, Masanori; Moorwood, —. 2007b, ApJ, 668, 400 Alan F. M. Proceedings of the SPIE, Volume 4841, pp. 1657-1669 Butler, N. R., et al. 2006, ApJ, 652, 1390 (2003)., ed. M. Iye & A. F. M. Moorwood, 1657–1669 Calzavara, A. J. & Matzner, C. D. 2004, MNRAS, 351, 694 Fabricant, D., Cheimets, P., Caldwell, N., & Geary, J. 1998, PASP, Campana, S. et al. 2006, Nature, 442, 1008 110, 79 Cappellaro, E., Evans, R., & Turatto, M. 1999, A&A, 351, 459 Falco, E. E., et al. 1999, PASP, 111, 438 Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 Fesen, R. A., et al. 2006, ApJ, 645, 283 Cherepashchuk, A. M. 1982, Ap&SS, 86, 299 Filippenko, A. V. 1982, PASP, 94, 715 Clocchiatti, A., Wheeler, J. C., Brotherton, M. S., Cochran, A. L., —. 1997, ARA&A, 35, 309 Wills, D., Barker, E. S., & Turatto, M. 1996, ApJ, 462, 462 27

Filippenko, A. V. 2005, in Astronomical Society of the Pacific Malesani, D., et al. 2008a, ApJL submitted (arXiv:0805.1188) Conference Series, Vol. 332, The Fate of the Most Massive Stars, Malesani, D. et al. 2004, ApJ, 609, L5 ed. R. Humphreys & K. Stanek, 33–+ —. 2008b, GCN Circ. 7169 Filippenko, A. V., Li, W. D., Treffers, R. R., & Modjaz, M. Marion, G. H., Höflich, P., Gerardy, C. L., Vacca, W. D. Wheeler, 2001, in ASP Conf. Ser. 246: IAU Colloq. 183: Small Telescope J. C., & Robinson, E. L. 2008, AJ, submitted Astronomy on Global Scales, ed. B. Paczynski, W.-P. Chen, & Matheson, T., Filippenko, A. V., Ho, L. C., Barth, A. J., & C. Lemme, 121–+ Leonard, D. C. 2000a, AJ, 120, 1499 Filippenko, A. V., Matheson, T., & Ho, L. C. 1993, ApJ, 415, Matheson, T., Filippenko, A. V., Li, W., Leonard, D. C., & Shields, L103+ J. C. 2001, AJ, 121, 1648 Filippenko, A. V. et al. 1995, ApJ, 450, L11+ Matheson, T. et al. 2000b, AJ, 120, 1487 Fisher, A., Branch, D., Hatano, K., & Baron, E. 1999, MNRAS, Matzner, C. D. & McKee, C. F. 1999, ApJ, 510, 379 304, 67 Mazzali, P. A., Deng, J., Maeda, K., Nomoto, K., Filippenko, A. V., Folatelli, G., et al. 2006, ApJ, 641, 1039 & Matheson, T. 2004, ApJ, 614, 858 Foley, R. J., et al. 2006, ApJ, 645, 450 Mazzali, P. A., et al. 2006, Nature, 442, 1018 Foley, R. J., Smith, N., Ganeshalingam, M., Li, W., Chornock, R., Mazzali, P. A., et al. 2005, Science, 308, 1284 & Filippenko, A. V. 2007, ApJ, 657, L105 Mazzali, P. A. & Lucy, L. B. 1998, MNRAS, 295, 428 Fransson, C. & Chevalier, R. A. 1987, ApJ, 322, L15 Miknaitis, G., et al. 2007, ArXiv Astrophysics e-prints Fukugita, M., Shimasaku, K., & Ichikawa, T. 1995, PASP, 107, 945 Millard, J., et al. 1999, ApJ, 527, 746 Galama, T. J. et al. 1998, Nature, 395, 670 Miller, J. S. & Stone, R. P. S. 1993, Lick Obs. Tech. Rep. 66 (Santa Garg, A., et al. 2007, AJ, 133, 403 Cruz: Lick Obs.) Garnavich, P. M. & Ann, H. B. 1994, AJ, 108, 1002 Mirabal, N., Halpern, J. P., An, D., Thorstensen, J. R., & Gehrels, N., et al. 2004, ApJ, 611, 1005 Terndrup, D. M. 2006, ApJ, 643, L99 Gerardy, C. L., et al. 2004, in Cosmic explosions in three Modjaz, M. 2007, PhD thesis, Harvard University dimensions, ed. P. Höflich, P. Kumar, & J. C. Wheeler, 57–+ Modjaz, M., Bloom, J. S., Butler, N. R., Starr, D., Kirshner, R. P., Gezari, S., et al. 2008, ArXiv e-prints, 804 & Friedman, A. 2008a, GRB Coordinates Network, 7175, 1 Ghisellini, G., Ghirlanda, G., & Tavecchio, F. 2007, MNRAS, 382, Modjaz, M., Chornock, R., Foley, R. J., Filippenko, A. V., Li, W., L77 & Stringfellow, G. 2008b, Central Bureau Electronic Telegrams, Gräfener, G., Koesterke, L., & Hamann, W.-R. 2002, A&A, 387, 1222, 1 244 Modjaz, M., Kirshner, R. P., & Challis, P. 2008c, ApJL, submitted Hamann, W.-R., Gräfener, G., & Liermann, A. 2006, A&A, 457, (arXiv0801.0221), 801 1015 Modjaz, M., Li, W., Filippenko, A. V., King, J. Y., Leonard, D. C., Hamuy, M., et al. 2002, AJ, 124, 417 Matheson, T., Treffers, R. R., & Riess, A. G. 2001, PASP, 113, Hamuy, M., Suntzeff, N. B., Gonzalez, R., & Martin, G. 1988, AJ, 308 95, 63 Modjaz, M., et al. 2006, ApJ, 645, L21 Harkness, R. P., et al. 1987, ApJ, 317, 355 Mould, J. R., et al. 2000, ApJ, 529, 786 Haynes, M. P., Giovanelli, R., Herter, T., Vogt, N. P., Freudling, Munari, U. & Zwitter, T. 1997, A&A, 318, 269 W., Maia, M. A. G., Salzer, J. J., & Wegner, G. 1997, AJ, 113, Nakano, S., Kadota, K., Itagaki, K., & Corelli, P. 2008, IAU Circ., 1197 8908, 2 Heise, J., in’t Zand, J., Kippen, R. M., & Woods, P. M. 2001, Nomoto, K., Iwamoto, K., & Suzuki, T. 1995, Phys. Rep., 256, 173 in Gamma-ray Bursts in the Afterglow Era, ed. E. Costa, Ofek, E. et al. 2008, GCN Circ. 7172 F. Frontera, & J. Hjorth, 16–+ Oke, J. B., et al. 1995, PASP, 107, 375 Herald, J. E., Hillier, D. J., & Schulte-Ladbeck, R. E. 2001, ApJ, Page, K. L. et al. 2008, GCN Report 110.1 548, 932 Patat, F. et al. 2001, ApJ, 555, 900 Herbig, G. H. 1975, ApJ, 196, 129 Phillips, A. C. & Davis, L. E. 1995, in Astronomical Society of the Hillier, D. J. & Miller, D. L. 1998, ApJ, 496, 407 Pacific Conference Series, Vol. 77, Astronomical Data Analysis Hjorth, J. et al. 2003, Nature, 423, 847 Software and Systems IV, ed. R. A. Shaw, H. E. Payne, & J. J. E. Höflich, P. 1991, A&A, 246, 481 Hayes, 297–+ Hook, I., et al. 2003, in Proceedings of the SPIE., ed. M. Iye & Phillips, M. M. 1993, ApJ, 413, L105 A. F. M. Moorwood, Vol. 4841, 1645–1656 Pian, E. et al. 2006, Nature, 442, 1011 Horne, K. 1986, PASP, 98, 609 Podsiadlowski, P., Langer, N., Poelarends, A. J. T., Rappaport, S., Immler, E. et al. 2008a, GCN Circ. 7168 Heger, A., & Pfahl, E. 2004, ApJ, 612, 1044 —. 2008b, GCN Circ. 7185 Poole, T. S., et al. 2008, MNRAS, 383, 627 Kong, E. & Maccarone, T. J. 2008, ATEL 1355 Pooley, D. & Soderberg, A. 2008, GCN Circ. 7216 Kong, E. et al. 2008a, GCN Circ. 7171 Predehl, P. & Schmitt, J. H. M. M. 1995, A&A, 293, 889 —. 2008b, GCN Circ. 7164 Pye, J. P. & McHardy, I. M. 1983, MNRAS, 205, 875 Landolt, A. U. 1992, AJ, 104, 340 Quimby, R. M., Wheeler, J. C., Höflich, P., Akerlof, C. W., Brown, Leibundgut, B., Kirshner, R. P., Pinto, P. A., Rupen, M. P., Smith, P. J., & Rykoff, E. S. 2007, ApJ, 666, 1093 R. C., Gunn, J. E., & Schneider, D. P. 1991, ApJ, 372, 531 Rayner, J. T., Toomey, D. W., Onaka, P. M., Denault, A. J., Leonard, D. C. & Filippenko, A. V. 2001, PASP, 113, 920 Stahlberger, W. E., Vacca, W. D., Cushing, M. C., & Wang, Leonard, D. C. & Filippenko, A. V. 2005, in ASP Conf. Ser. S. 2003, PASP, 115, 362 342: 1604-2004: Supernovae as Cosmological Lighthouses, ed. Rest, A., et al. 2005, ApJ, 634, 1103 M. Turatto, S. Benetti, L. Zampieri, & W. Shea, 330–+ Richardson, D., Branch, D., & Baron, E. 2006, AJ, 131, 2233 Li, H. & McCray, R. 1992, ApJ, 387, 309 Richmond, M. W., Treffers, R. R., Filippenko, A. V., Paik, Y., Li, L.-X. 2008, MNRAS, accepted (arXiv:0803.0079) Leibundgut, B., Schulman, E., & Cox, C. V. 1994, AJ, 107, 1022 Li, W., Chornock, R., Foley, R. J., Filippenko, A. V., Modjaz,M., Roming, P. W. A., et al. 2005, Space Science Reviews, 120, 95 Poznanski, D., & Bloom, J. S. 2008, GRB Coordinates Network, Sakamoto, T., et al. 2008, ArXiv e-prints, 801 7176, 1 Sakamoto, T., et al. 2005, ApJ, 629, 311 Li, W. & Filippenko, A. V. 2008, Central Bureau Electronic Sauer, D. N., Mazzali, P. A., Deng, J., Valenti, S., Nomoto, K., & Telegrams, 1202, 3 Filippenko, A. V. 2006, MNRAS, 369, 1939 Li, W., et al. 2001, PASP, 113, 1178 Schawinski, K., et al. 2008, ArXiv e-prints, 803 Li, W., Jha, S., Filippenko, A. V., Bloom, J. S., Pooley, D., Foley, Schechter, P. L., Mateo, M., & Saha, A. 1993, PASP, 105, 1342 R. J., & Perley, D. A. 2006, PASP, 118, 37 Scheck, L., Kifonidis, K., Janka, H.-T., & Müller, E. 2006, A&A, Lodders, K. 2003, ApJ, 591, 1220 457, 963 Lucy, L. B. 1991, ApJ, 383, 308 Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525 Maeda, K., et al. 2008, Science, 319, 1220 Schmidt, B. P., et al. 1993, Nature, 364, 600 Maeda, K., Nomoto, K., Mazzali, P. A., & Deng, J. 2006, ApJ, 640, Schmidt, G. D., Weymann, R. J., & Foltz, C. B. 1989, PASP, 101, 854 713 Maeda, K., et al. 2007, ApJ, 666, 1069 Skrutskie, M. F., et al. 2006, AJ, 131, 1163 28

Smith, N. & Conti, P. S. 2008, ApJ, in press (2008arXiv0802.1742), Valenti, S., et al. 2008a, GRB Coordinates Network, 7171, 1 802 Valenti, S. et al. 2008b, Central Bureau Electronic Telegrams, 1205, Smith, N. & Owocki, S. P. 2006, ApJ, 645, L45 1 Soderberg, A. M., et al. 2008, Nature, accepted (arXiv:0802.1712) Vogt, S. S., et al. 1994, in Presented at the Society of Photo-Optical Soderberg, A. M., et al. 2006, Nature, 442, 1014 Instrumentation Engineers (SPIE) Conference, Vol. 2198, Proc. Sollerman, J., Leibundgut, B., & Spyromilio, J. 1998, A&A, 337, SPIE Instrumentation in Astronomy VIII, David L. Crawford; 207 Eric R. Craine; Eds., Volume 2198, p. 362, ed. D. L. Crawford Sollerman, J. et al. 2006, A&A, 454, 503 & E. R. Craine, 362–+ Spyromilio, J. 1991, MNRAS, 253, 25P Wade, R. A. & Horne, K. 1988, ApJ, 324, 411 —. 1994, MNRAS, 266, L61+ Wang, C., Lai, D., & Han, J. L. 2006, ApJ, 639, 1007 Stanek, K. Z. et al. 2003, ApJ, 591, L17 Wang, X.-Y., Li, Z., Waxman, E., & Mészáros, P. 2007, ApJ, 664, Stritzinger, M. et al. 2002, AJ, 124, 2100 1026 Suntzeff, N. B. & Bouchet, P. 1990, AJ, 99, 650 Waxman, E., Mészáros, P., & Campana, S. 2007, ApJ, 667, 351 Taubenberger, S., et al. 2006, MNRAS, 371, 1459 Wheeler, J. C., et al. 1993, ApJ, 417, L71+ Tominaga, N. et al. 2005, ApJ, 633, L97 Wood-Vasey, W. M., et al. 2007, ArXiv e-prints, 711 Tomita, H., et al. 2006, ApJ, 644, 400 Woosley, S. E. & Bloom, J. S. 2006, ARA&A, 44, 507 Turatto, M., Benetti, S., & Cappellaro, E. 2003, in From Twilight Woosley, S. E., Langer, N., & Weaver, T. A. 1993, ApJ, 411, 823 to Highlight: The Physics of Supernovae, ed. W. Hillebrandt & Xu, D., Zou, Y.-C., & Fan, Y.-Z. 2008, ArXiv e-prints, 801 B. Leibundgut, 200–+ Uomoto, A. & Kirshner, R. P. 1986, ApJ, 308, 685 Vacca, W. D., Cushing, M. C., & Rayner, J. T. 2003, PASP, 115, 389 29

TABLE 1 Model Parameters for X-ray Spectral Fits of XRT 080109

Parameter Power Law (PL) Fit Blackbody (BB) Fit Combined BB and PL Fita Quality of Fit (χ2/DOF ) 12.73/24 28.56/25 12.57/24 21 −2 +2.1 +2.1 Implied NH [10 cm ] 5.2−1.8 < 1.5 5.2−1.8 +0.3 c Time-integrated Γ, kT , and kT respectively 2.1−0.4 0.75 ± 0.07 0.10 ± 0.01 b 42 −1 +5 c LX,iso [ 10 erg s ] 12−2 4.6 ± 0.6 13 ± 2 b 45 +2 c EX,iso [10 erg] 5.8−1 2.4± 3 6.7 ± 4 a Assuming the same (factor unity) ratio of PL to BB flux as in XRF 060218 (Butler et al. 2006). b Un-absorbed and time-averaged value over the range 0.3−10.0 keV (for PL fit) and bolometric (for BB fit), respectively, assuming D = 31 ± 2 Mpc. c X 11 Listed only for the BB component. LX,iso and EX,iso are computed assuming RBB = 10 cm as required from the combined PL and BB fit.

TABLE 2 Photometry of comparison stars in the field of SN 2008D by KAIT and FLWO 1.2m

′ ′ ID U NU B NB V NV R/r NR I/i NI 1 16.771(029) 2 16.438(002) 3 15.563(006) 2 15.117(013) 3 14.640(016) 3 2 17.703(037) 2 17.732(003) 2 16.997(002) 4 16.590(005) 2 16.274(007) 6 3 13.611(005) 3 13.745(008) 3 13.234(010) 4 12.933(010) 4 12.639(015) 3 4 17.822(069) 2 17.865(009) 3 17.409(012) 4 17.061(010) 5 16.781(002) 6 5 − − 17.994(011) 2 17.316(010) 4 16.841(013) 4 16.498(009) 5 6 − − 18.132(014) 4 17.550(016) 3 17.115(005) 4 16.856(011) 6 7 12.466(005) 3 12.530(002) 3 − − − − −− 8 − − 18.977(012) 4 17.924(011) 4 17.227(006) 5 16.789(003) 3 9 15.003(007) 3 14.969(003) 3 14.394(016) 3 14.041(010) 3 13.734(013) 3

1 − − 16.415(049) 1 15.586(033) 1 15.326(045) 1 15.088(046) 1 3 − − 13.657(031) 1 13.191(022) 1 13.089(031) 1 12.999(031) 1 10 − − 16.498(049) 1 15.594(033) 1 15.298(045) 1 15.047(045) 1 11 − − 16.908(051) 1 15.764(035) 1 15.344(047) 1 15.004(048) 1 12 − − 18.068(089) 1 17.237(060) 1 17.038(079) 1 16.805(081) 1 13 − − 15.659(040) 1 14.827(027) 1 14.585(037) 1 14.391(037) 1 14 − − 14.364(033) 1 13.631(023) 1 13.446(031) 1 13.311(031) 1 15 − − 17.584(086) 1 17.164(058) 1 17.087(076) 1 17.040(078) 1 16 − − 16.054(045) 1 15.320(031) 1 15.105(041) 1 14.946(042) 1 17 − − 17.309(055) 1 15.966(037) 1 15.486(049) 1 15.059(051) 1 Note. — Note: uncertainties (standard deviation of the mean) are indicated in parentheses. Stars (see Fig. 3) in the top half of the table were observed with KAIT in UBV RI and the stars in the bottom half were observed with the FLWO 1.2m telescope in BV r′i′. 30

TABLE 3 Lick Observatory photometry of SN 2008D

∆ta B V RR(c)b I Tel.c 1.85 19.05(07) 18.33(04) 17.80(03) 17.78(05) 17.34(05) KAIT 2.78 19.01(03) 18.39(05) 17.84(03) − 17.36(02) Nickel 2.89 19.15(08) 18.43(05) 17.85(04) 17.82(05) 17.38(04) KAIT 3.78 19.09(03) 18.37(02) 17.89(02) − 17.37(02) Nickel 3.87 19.22(09) 18.43(05) 17.83(04) 17.81(05) 17.40(04) KAIT 4.86 19.14(06) 18.30(04) 17.78(04) 17.76(05) 17.19(04) KAIT 5.83 19.23(07) 18.22(04) 17.68(03) 17.64(05) 17.17(03) KAIT 6.86 18.95(11) 17.97(05) 17.47(06) 17.48(05) 16.97(05) KAIT 7.82 18.88(11) 17.93(05) 17.34(03) 17.31(05) 16.81(04) KAIT 8.79 18.90(10) 17.73(04) 17.27(02) 17.21(06) 16.72(03) KAIT 9.79 18.69(07) 17.69(04) 17.13(02) 17.07(06) 16.58(02) KAIT 10.77 18.54(08) 17.61(05) 17.07(02) 16.97(08) 16.41(02) KAIT 26.85 18.92(07) 17.54(03) 16.82(02) − 16.20(02) KAIT 27.79 19.01(05) 17.58(03) 16.86(02) − 16.22(02) KAIT 28.81 19.13(06) 17.64(03) 16.86(02) − 16.24(02) KAIT 29.78 19.19(05) 17.74(03) 16.95(02) − 16.27(03) KAIT 30.76 19.28(06) 17.82(03) 17.01(02) − 16.28(02) KAIT 32.75 19.52(08) 17.92(03) 17.11(02) − 16.38(02) KAIT 34.72 19.70(06) 18.06(03) 17.21(02) − 16.42(02) KAIT 35.76 19.74(12) 18.08(03) 17.26(02) − 16.47(02) KAIT Note. — Note: uncertainties of the measurements are indicated in parentheses. a b Time in days since t0. R(c)-band magnitudes derived from unﬁltered observations.c Telescope being used: KAIT = 0.76-m Katzman Automatic Imaging Telescope; Nickel = Lick Observatory 1-m Nickel telescope.

TABLE 4 Optical FLWO 1.2m Photometry of SN 2008D

MJD ∆tB V r′ i′ [days] [days] [mag] [mag] [mag] [mag] 54474.45 -0.12 > 20.10 > 19.80 > 20.20 > 20.20 54475.41 0.84 19.08(03) 18.52(04) 18.31(04) 18.17(03) 54476.43 1.87 18.98(03) 18.38(03) 18.11(03) 17.86(03) 54477.36 2.79 19.11(02) 18.44(03) 18.15(02) 17.86(02) 54478.39 3.83 19.24(02) 18.47(02) 18.17(02) 17.89(02) 54480.39 5.83 19.34(06) 18.22(04) 18.02(03) 17.72(04) 54481.51 6.94 19.10(04) 18.05(04) 17.75(05) 17.58(09) 54482.43 7.86 18.98(02) 17.94(03) 17.65(03) 17.34(03) 54483.27 8.71 18.89(05) 17.70(03) 17.51(02) 17.25(03) 54489.41 19.90 18.76(05) 17.63(03) 17.00(02) 16.87(02) 54495.36 20.80 18.36(04) 17.38(03) 17.00(02) 16.65(02) 54496.34 21.77 18.52(03) 17.41(02) 16.95(02) 16.59(02) 54497.39 22.83 18.63(03) 17.40(02) 17.06(03) 16.68(03) 54498.30 23.74 18.63(05) 17.50(04) 17.05(02) 16.65(02) 54499.35 24.79 18.72(02) 17.49(02) 17.07(02) 16.70(02) 54502.44 27.87 18.93(04) 17.51(02) 17.18(02) 16.77(02) 54503.34 28.78 19.15(03) 17.67(02) 17.22(02) 16.78(02) 54504.37 29.81 19.25(04) 17.71(02) 17.29(02) 16.83(02) 31

TABLE 5 UVOT photometry of SN 2008D

∆ta δtb Exp c Mag σ Filter [days] [days] [sec] [mag] 0.1359 0.2688 412.27 20.24 0.21 U 2.0047 1.0058 2677.07 18.62 0.03 U 2.9769 0.8591 1001.81 19.07 0.06 U 3.9479 0.9349 1314.54 19.36 0.07 U 4.9872 0.9346 1347.46 19.37 0.07 U 5.9550 0.9326 921.90 19.45 0.08 U 6.9283 0.9920 997.54 19.37 0.08 U 8.0761 0.9269 3492.92 19.31 0.05 U 9.3099 0.1476 1072.24 19.32 0.07 U 11.7894 1.0748 3179.82 19.14 0.04 U 15.1978 0.4090 2247.42 18.99 0.04 U 16.0643 0.6762 3023.59 19.03 0.04 U 17.0180 0.9388 3015.00 19.15 0.05 U 18.0722 0.6754 3217.95 19.18 0.05 U 18.8761 0.8101 2693.93 19.15 0.05 U 20.0333 0.8677 1957.61 19.19 0.05 U 20.9489 0.1390 1280.75 19.28 0.07 U 21.5476 0.1383 1103.65 19.37 0.07 U 22.9628 0.2705 1350.25 19.33 0.07 U 23.8248 0.1406 1694.19 19.39 0.07 U 24.9641 0.4028 1825.17 19.91 0.10 U 25.8991 0.1371 1576.07 19.85 0.10 U 29.7807 0.1353 899.42 20.53 0.19 U 30.8165 0.4675 895.05 20.49 0.18 U 31.7877 0.1350 847.76 20.63 0.20 U 32.8267 0.2102 1254.98 20.82 0.21 U 34.0904 0.2071 1005.21 21.32 0.35 U 34.8159 0.2725 1018.80 21.05 0.27 U 36.7836 0.7591 1327.09 21.34 0.30 U 38.7112 0.0866 3135.55 21.39 0.28 U 0.1370 0.2688 412.27 >20.75 − B 2.0265 1.0058 2579.13 19.05 0.04 B 2.9778 0.8592 1001.79 19.32 0.06 B 3.9490 0.9350 1314.47 19.31 0.05 B 4.9885 0.9345 1347.47 19.26 0.05 B 5.9558 0.9325 921.86 19.19 0.06 B 6.9291 0.9918 997.41 19.05 0.05 B 8.0799 0.9244 1749.45 18.91 0.04 B 9.3141 0.1452 536.69 18.83 0.06 B 11.7202 1.0728 1488.37 18.59 0.03 B 15.2016 0.4072 1119.61 18.51 0.03 B 16.0676 0.6744 1504.09 18.48 0.03 B 17.0211 0.9371 1511.09 18.48 0.03 B 18.0757 0.6736 1611.71 18.62 0.03 B 18.8786 0.8100 1346.53 18.59 0.03 B 20.0359 0.8657 979.26 18.55 0.04 B 20.9540 0.1380 1017.40 18.53 0.04 B 21.5520 0.1374 886.87 18.50 0.04 B 22.9660 0.2689 673.18 18.66 0.04 B 23.8315 0.1403 1625.55 18.66 0.03 B 24.9672 0.4004 911.83 18.75 0.04 B 25.9053 0.1355 1508.81 18.82 0.04 B 29.8169 0.0693 439.22 19.19 0.07 B 30.9445 0.2692 338.88 19.17 0.08 B 31.7911 0.1348 772.71 19.34 0.07 B 32.8305 0.2122 1094.90 19.92 0.09 B 34.0944 0.2041 863.57 19.84 0.09 B 34.8190 0.2738 917.95 19.64 0.08 B 36.7867 0.7575 1266.95 19.93 0.08 B 38.5493 0.1420 1265.38 20.01 0.09 B 0.1420 0.2685 412.25 >19.61 − V 2.0202 0.9383 2349.89 18.39 0.04 V 3.0145 0.7952 939.01 18.55 0.07 V 3.9543 0.9358 1314.63 18.42 0.06 V 4.9947 0.9339 1347.48 18.27 0.05 V 5.9595 0.9320 921.87 18.16 0.05 V 6.9329 0.9911 997.38 17.98 0.05 V 8.0820 0.9242 1321.31 17.90 0.04 V 9.3163 0.1443 473.94 17.94 0.07 V 11.7225 1.0722 1043.80 17.47 0.04 V 15.2036 0.4068 846.50 17.37 0.04 V 16.0693 0.6743 1183.73 17.38 0.03 V 17.0226 0.9370 1272.83 17.36 0.03 V 18.0775 0.6735 1312.98 17.36 0.03 V 18.8799 0.8104 1007.89 17.31 0.03 V 20.0373 0.8651 668.21 17.35 0.04 V 20.9438 0.1390 1280.81 17.38 0.03 V 21.5432 0.1383 1103.62 17.39 0.03 V 22.9677 0.2688 619.40 17.36 0.04 V 23.8181 0.1406 1694.14 17.42 0.03 V 24.9688 0.3995 789.30 17.54 0.04 V 25.8928 0.1385 1576.05 17.57 0.03 V 29.7764 0.1385 1083.93 17.75 0.04 V 30.9459 0.2684 93.90 17.91 0.14 V 31.7843 0.1348 847.69 17.90 0.05 V 32.8230 0.2070 1254.96 17.95 0.04 V 34.0864 0.2099 1005.20 18.23 0.06 V 34.8128 0.2711 1018.74 18.10 0.05 V 36.7804 0.7605 1327.52 18.27 0.05 V 38.5438 0.1401 1397.89 18.31 0.05 V a Time in days (middle point of the combined image) since b t0= 2008-01-09 13:32:49. Time bin in days (from the start of the ﬁrst exposure to the end of the last exposure in the combined image).c Total exposure time (in seconds). 32

TABLE 6 NIR Photometry of SN 2008D

MJD ∆tJ H Ks [days] [days] [mag] [mag] [mag] 54474.35 -0.21 > 18.15 > 17.12 > 16.18 54475.27 0.71 17.00(03) 16.70(05) 17.09(12) 54476.33 1.77 16.54(07) 16.22(09) 16.22(10) 54477.34 2.78 16.45(07) 16.15(09) 16.13(15) 54478.34 3.78 16.44(04) 16.05(09) 15.99(12) 54479.38 4.82 16.38(10) 15.96(10) 15.81(10) 54480.34 5.78 16.23(10) 15.86(10) 15.63(07) 54481.32 6.76 15.99(10) 15.72(07) 15.55(02) 54482.26 7.70 15.88(04) 15.55(00) 15.46(05) 54483.32 8.76 15.74(08) 15.41(13) 15.21(06) 54484.35 9.79 15.62(05) 15.31(05) 15.11(07) 54485.31 10.75 15.51(02) 15.17(16) 15.05(13) 54486.35 11.79 15.43(01) 15.12(05) 14.92(05) 54497.19 22.63 15.00(04) 14.76(08) 14.53(03) 54499.26 24.70 15.07(21) 14.73(03) 14.44(04) 54505.20 30.64 15.11(05) 14.76(02) 14.54(11) 54508.24 33.68 15.25(02) 14.86(01) 14.68(13) 54509.22 34.66 15.24(05) 14.88(03) 14.68(05) 54510.25 35.69 15.32(04) 14.89(06) 14.68(03)

TABLE 7 Blackbody Fits and Bolometric Light Curves of SN 2008D from U → Ks Data

∆t TBB RBB log(LBB) log(LU−Ks ) [days] [K] [1014 cm] [ergs−1] [ergs−1]

+0.30 +0.12 0.840 11700±2600 2.79±0.33 42.01−0.23 41.69−0.12 +0.25 +0.14 1.80 11000±2000 3.41±0.34 42.03−0.20 41.86−0.13 +0.17 +0.13 2.80 8300±1100 4.68±0.35 41.86−0.14 41.79−0.12 +0.14 +0.12 3.80 7400±830 5.69±0.38 41.82−0.12 41.79−0.11 +0.13 +0.12 4.80 6900±710 6.51±0.41 41.83−0.11 41.82−0.11 +0.10 +0.12 5.80 6700±200 7.20±0.41 41.86−0.10 41.86−0.11 +0.08 +0.11 6.80 6600±190 7.82±0.41 41.90−0.10 41.90−0.11 +0.08 +0.11 7.80 6500±190 8.43±0.43 41.95−0.09 41.94−0.11 +0.08 +0.11 8.80 6400±180 8.99±0.46 41.99−0.08 41.94−0.11 +0.08 +0.11 9.80 6400±180 9.45±0.48 42.02−0.08 42.01−0.11 +0.08 +0.11 10.8 6400±170 9.93±0.50 42.05−0.08 42.04−0.10 +0.08 +0.11 11.8 6300±170 10.4±0.52 42.08−0.08 42.07−0.10 +0.08 +0.11 12.8 6300±170 10.8±0.54 42.10−0.08 42.10−0.10 +0.08 +0.11 13.8 6200±160 11.2±0.55 42.12−0.08 42.12−0.10 +0.08 +0.11 14.8 6100±160 11.7±0.57 42.13−0.08 42.14−0.10 +0.08 +0.11 15.8 6100±160 12.1±0.59 42.14−0.08 42.16−0.10 +0.08 +0.11 16.8 6000±150 12.5±0.61 42.15−0.08 42.17−0.10 +0.08 +0.11 17.8 5900±150 12.9±0.62 42.16−0.08 42.18−0.10 +0.07 +0.10 18.8 5800±140 13.3±0.64 42.16−0.07 42.18−0.10 +0.07 +0.10 19.8 5800±140 13.8±0.65 42.16−0.07 42.18−0.10 +0.07 +0.10 20.8 5700±130 14.2±0.67 42.16−0.07 42.18−0.10 +0.07 +0.10 21.8 5600±130 14.6±0.68 42.16−0.07 42.18−0.09 +0.07 +0.10 22.8 5500±120 15.0±0.70 42.15−0.07 42.17−0.09 +0.07 +0.10 23.8 5400±110 15.5±0.71 42.14−0.07 42.17−0.09 +0.07 +0.10 24.8 5200±110 16.0±0.73 42.13−0.06 42.16−0.09 +0.07 +0.10 25.8 5100±100 16.6±0.75 42.11−0.06 42.15−0.09 +0.07 +0.10 26.8 4900±94. 17.2±0.76 42.09−0.06 42.13−0.09 +0.06 +0.09 27.8 4800±87. 17.9±0.78 42.07−0.05 42.12−0.09 +0.06 +0.09 28.8 4600±81. 18.5±0.80 42.05−0.05 42.09−0.09 +0.06 +0.09 29.8 4600±78. 18.7±0.81 42.03−0.05 42.07−0.08

Note. — Adopted values: E(B − V )Host = 0.6 ± 0.1 mag, D = 31±2 Mpc, and t0= 2008-01-09 13:32:49 UT. The uncertainties include the uncertainty in reddening, but not in distance. 33

TABLE 8 Optical He I Absorption Line Velocities of SN 2008D

JD - 2400000 ∆t tVmax He I λ5876 He I λ6678 He I λ7065 [days] [days] [days] [km s−1] [kms−1] [kms−1] 54480.82 5.75 -12.7 -14200 -11800 -10500 54481.84 6.77 -11.7 -13300 -12700 -13000 54484.07 9.00 -9.4 -12300 -12400 -10900 54486.72 11.66 -6.8 -11500 -11100 -10500 54497.85 22.78 4.3 -10600 -10400 -10100 54498.94 23.88 5.4 -10600 -10200 -10200 54505.62 30.56 12.1 -10500 -10200 -9900 54507.68 32.61 14.2 -10400 -10000 -9900 54508.93 33.87 15.4 -10300 -9900 -9700 54512.86 37.79 19.4 -10000 -9600 -9500 54524.87 49.81 31.4 -9300 -8600 -8700 54535.73 60.67 42.2 -8600 -7900 -8100

TABLE 9 Doppler Velocities for NIR He I Lines of SN 2008D

a λrest(air) λmeasured Velocity λmeasured Velocity Uncertainty [µm] Jan. 21.51 km s−1 Jan. 28.1 [km s−1] [kms−1] 1.0830 1.0306 −14,900 1.0340 −13,900 570 1.2786 1.2254 −12,700 1.2292 −11,800 520 1.7002 1.6302 −12,600 1.6454 − 9,800 360 1.8688 1.7969 −11,800 1.8037 −10,600 300 2.0581 1.9760 −12,200 1.9824 −11,200 260 a Uncertainty estimated as half the wavelength diﬀerence to the next data point.