SN 2008Ha and Low-Luminosity Transients

SN 2008ha and Low-Luminosity Transients

Ryan Foley Clay Fellow Center for Astrophysics

Stellar Death & Supernovae Collaborators

Kirshner, Challis, Friedman, Rest

Filippenko, Li, Chornock, Ganeshalingam, Silverman, Steele

Brown

Jha

Wood-Vasey SN 2008A 1 arcmin

– 50 –

– 56 –

Lick 40” V-band -20 B -19 NGC 634 DSS January 12, 2008 -18 SN 2002cx Had Low Velocities2005hk and Figure 1: Image of SN 2008A from the Lick Observator-17y 1 m Nickel telescope (left) compared to the pre-explosionLowDS SLuminosityimage (right). -20 – 50 – -19 V -18 – 56 – -17 Exceptional SN Ia that did not follow the Phillips relation -20 B -19 I max -18 Many odd characteristics, M 2005hk -17 including low line velocities -20 1.0 -19 VJ Large group (~16 objects) of -19 -18

) + constant similar objects SN 2002cx ! -17

SN 2002cx (+25 days) -20

Log (f See Jha’s talk for more I SN 2005hk (+24 days) -19 H 0.1 SN 2008A (+28 days) max -18 M -18 SN 2008A -17 4000 5000 6000 7000 8000 9000 -20 0.8 1 1.2 1.4 1.6 1.8 2 Rest Wavelength (Å) m (B) J -19 ! 15 -18 Adapted from Phillips et al. 2007 Figure 2: February 12, 2008 Keck spectrum ofFigSN. 14.—2008AThe abscolomparute magnietuddets oof anSNealoguesIa at maxiSmuNm 2002cxlight in theandBV IJH bands -17 SN 2005hk at similar epochs showing strikingplhomogeneitotted versus the ydec(llineefrt);ate pandarameabsoluter ∆m15(teB).opticalThe blackmtriagnangleituds arFeigSe.Ns8e.—in Cthoemparison of spectra of SN 2005hk at phases of -5, +13, +24, and +55 days with redshift range 0.01 < z < 0.1 whose distances were calculated from their host galaxy radial similar epoch spectra of SN 2002cx from LFC. The spectra are plotted on a logarithmic ﬂux of SN 2008A and SN 2005hk showing these arevelocouitietlierss in the-20cocomparedsmic microwavetobackagrosampleund frame aofssumnormaling a HubbSNele constIaant of H = 72 scale an0d shifted by an arbitrary constant. The wavelengths of the spectra were shifted to km s−1 Mpc−1. The red circle in each panel corresponds to SN 2005hk.H (right; panel adapted from Phillips et al. 2007). -19 the SN rest frame using the heliocentric velocities of the host galaxies given in NED. -18 -17 0.8 1 1.2 1.4 1.6 1.8 2 !m (B) 4 15 Fig. 14.— The absolute magnitudes of SNe Ia at maximum light in the BV IJH bands plotted versus the decline rate parameter ∆m15(B). The black triangles arFeigS.N8e.—in Cthoemparison of spectra of SN 2005hk at phases of -5, +13, +24, and +55 days with redshift range 0.01 < z < 0.1 whose distances were calculated from their host galaxy radial similar epoch spectra of SN 2002cx from LFC. The spectra are plotted on a logarithmic ﬂux velocities in the cosmic microwave background frame assuming a Hubble constant of H = 72 scale an0d shifted by an arbitrary constant. The wavelengths of the spectra were shifted to km s−1 Mpc−1. The red circle in each panel corresponds to SN 2005hk. the SN rest frame using the heliocentric velocities of the host galaxies given in NED. SN 2008A 1 arcmin

– 50 –

– 56 –

Lick 40” V-band -20 B -19 NGC 634 DSS January 12, 2008 -18 SN 2002cx Had Low Velocities2005hk and Figure 1: Image of SN 2008A from the Lick Observator-17y 1 m Nickel telescope (left) compared to the pre-explosionLowDS SLuminosityimage (right). -20 – 50 – -19 V -18 – 56 – 466 LI ET AL. -17 Exceptional SN Ia that did not follow the Phillips relation -20 B -19 I max -18 Many odd characteristics, M 2005hk -17 including low line velocities -20 1.0 -19 VJ Large group (~16 objects) of -19 -18

) + constant similar objects SN 2002cx ! -17

SN 2002cx (+25 days) -20

sponds to the broad-17R and I bands, so the peculiar spectral the Ca ii H and K lines, the measured Vexp of SN 2002cx is evolution in this region may be related to the unique photo- Ϫ6000 km sϪ1, only about half that of the other SNe Ia. The ∼ metric behavior of the SN0.8in the R and I 1bands. In particular1.2 , 1.4Vexp evolution1.6is also ver1.8y ﬂat, implying2 that the velocity gra- the additional emission near 7000 A˚ may contribute to the broad!m dient(B)in the ejecta of SN 2002cx is small during the period of peak around maximum and the slow late-time decline in the 15spectral observations (t p Ϫ4 to 56 days). R4band, and the lack of strong Ca ii IR triplet absorption may No apparent Si ii l6355 or S ii ll5612, 5654 lines are Fig.contribute14.— toTthheeplateauabsoaroundlute mmaximumagnituanddestheosflowSNlate-e Ia aobsert mvedaxinimourumspectralighoft SNin 2002cx,the BbutV ItheJHexpansionbandve-s time decline in the I band. More discussion of the R-band and locity measured from these lines for other SNe Ia are shown plotted versus the decline rate parameter ∆m15(B). The black triangles are SNe in the I-band light curves can be found in § 4.3. in Figure 12 for completeness. Note thatFSNig.1999ac8.—hasCothemparison of spectra of SN 2005hk at phases of -5, +13, +24, and +55 days with redshift range 0.01 < z < 0.1 whose distances were calculated from their host galaxy radial lowest Vexp measured from S ii ll5612, 5654similineslar amongepoctheh spectra of SN 2002cx from LFC. The spectra are plotted on a logarithmic ﬂux 3.6. Expansion Velocities velocities in the cosmic microwave background framseverale assuSNeminIagshowna HuinbbFigurele co12.nstTheantvaluesof H0of=Vexp72of −1 −1 scale and shifted by an arbitrary constant. The wavelengths of the spectra were shifted to km s TheMexpansionpc . Tvelocitieshe red(Vcexpir),casle inferredin eacfromh paobsernelvedcorresSNpo1999acnds tomeasuredSN 20from05htkhe. Fe ii and Fe iii lines are also minima of absorption lines in the spectra, may provide some lower than in SNe 1991T and 2000cx, althoughthe ShigherN rethanst finrame using the heliocentric velocities of the host galaxies given in NED.

clue to the nature of SN Ia explosions (Branch, Drucker, & SN 2002cx. The similarity of the Vexp evolution between Jeffery 1988; Khokhlov, Mu¨ller, & Ho¨ﬂich 1993). Figure 12 SN 2002cx and SN 1999ac is consistent with their similar shows the expansion velocities derived from several lines. The photometric behavior in the early B-band and V-band light

2003 PASP, 115:453–473 190 JHA ET AL.

TABLE 1 Spectroscopic Observations of SN 2002cx

Wavelength Range Resolution Exposure P.A. Parallactic Seeing UT Date (8) (8) (s) (deg) (deg) Air Mass (arcsec)

2003 Jan 07.65...... 3300–9430 6 2 ; 900 153.0 132.2 1.06 1.0 2003 Feb 27.66...... 3160–9420 6 800 66.0 67.3 1.28 0.7 2003 Feb 28.63...... 3160–5770 6 2 ; 2200 240.0 242.6 1.18 0.8 2003 Feb 28.63...... 6260–7540 2 2 ; 2200 240.0 242.6 1.18 0.8

has higher resolution than the others, so we have analyzed it February 27 and 28 observations correspond to 227 and 277 days separately. past B maximum light in the SN rest frame (Li et al. 2003b). We additionally show the latest SN 2002cx spectrum from the pre- 3. ANALYSIS AND RESULTS vious observing season, at an epoch of +56 days, presented by Li et al. (2003b) and analyzed in more detail by Branch et al. The late-time spectra of SN 2002cx, corrected to rest-frame 1 (2004a). For comparison we also display spectra of normal SNe wavelengths (with cz 7184 km sÀ for the host galaxy, CGCG Ia at similar epochs, including SN 1998aq (+52 days) and SN SN 2002cx-like Objects 044Form035; Falco et al. ¼1999),a are shown in Figure 1.5 Narrow, un- À 1998bu (+236 days) from our spectral database and SN 1990N resolved emission lines of H and [O ii] k3727 from a super- (+280 days) from Go´mez & Lo´pez (1998) via the online SUSPECT posed H ii region in the host galaxy are shown in light gray; database.6 We furthermore show our archive spectrum of the sub- Distinct Class they have been excised in subsequent ﬁgures. The January 7 and luminous SN 1999by (+182 days), which had an intrinsic peak brightness similar to SN 2002cx but with a much faster declining

5 light curve and a very different early-time spectrum (Garnavich Throughout this paper we employ a logarithmic wavelength axis to facilitate comparison of line widths. et al. 2004). Despite the clearly peculiar nature of SN 2002cx at early times (Li et al. 2003b; Branch et al. 2004a), the +56 day spectrum does • not show gross differences compared to the normal SN 1998aq, Low ejecta velocity apart from lower expansion velocities in the lines. However, the late-time spectra below appear very different from normal SN Ia counterparts such as SN 1998bu and SN 1990N, whose ﬂux in the • Low luminosity optical is dominated by broad, blended emission lines of [Fe ii] and [Fe iii] (e.g., Ruiz-Lapuente et al. 1995; Mazzali et al. 1998). Much of the high-frequency structure in the SN 2002cx observations is due to real features, and in general these narrow lines • Hot photosphere (similar to do not correspond to resolved versions of the features seen in SN 1998bu and SN 1990N. SN 1991T) at early times The late-time SN 2002cx spectra do show a relatively broad feature coincident with the strong [Fe iii] k4700 feature seen in the SN 1998bu and SN 1990N spectra. However, unlike normal SNe Ia whose spectra change dramatically between 2 and 8– • Lacks second NIR maximum 9 months past maximum (compare SN 1998aq with SN 1998bu and SN 1990N in Fig. 1), the late-time spectra of SN 2002cx bear a resemblance to its +56 day spectrum, as shown in Figure 2. Over this wavelength range, the +227 day spectrum merely • Late-time spectrum lacks shows more resolved lines with less blueshifted absorption. So, rather than forbidden iron emission, the broad 4700 8 feature strong forbidden lines seen at late times in SN 2002cx may be the same species as in the +56 day spectrum, which can be modeled with P Cygni proﬁles

6 See http://suspect.nhn.ou.edu/~suspect.

Fig. 1.—Optical spectra of SN 2002cx (blue) compared with normal SNe Ia at similar epochs (SN 1998aq, SN 1998bu, anJhad SN 19 9et0N; blal.ack ) 2006and the subluminous SN 1999by (violet). The spectra have been arbitrarily scaled and shifted. Narrow emission lines from an H ii region superposed along the line of sight to SN 2002cx are shown in light gray. The epochs listed correspond to SN rest-frame days past B maximum light. The SN 1990N comparison spectrum is from Go´mez & Lo´pez (1998), while we have observed SN 1998aq on 1998 June 18 and SN 1998bu on 1999 January 10 with the Lick Observatory 3 m Shane Fig. 2.—Comparison between the +56 day spectrum of SN 2002cx (black) telescope (+KAST) and SN 1999by on 1999 November 9 with Keck II (+LRIS). and the +227 day spectrum (blue). SN 2008ha Had Very Low Ejecta Velocity

10 Foley et al.

At t ≈ 6 days, vej ≈ 2000 km/s 0.3 07J 2007 Mar 18 He I 2.5 Very narrow features 91T 0.2 0.1 2.0 08ha Matches SN 2002cx when 2000 0.0 convolved with a 800 km/s 1.5 0.3 89B +79 d + Constant Gaussian ! 08ha 0.2 1.0

Relative f 0.1 Matches SN 1991T when 08ha convolved with a 2000 km/s 0.5 800 0.0

02cx 04aw +64 d Gaussian ! 0.0 0.2 3000 4000 5000 6000 7000 8000 9000 Rest Wavelength (Å) 0.1 Relative f Na D Fig. 6.— Optical spectra of SNe 1991T, 2002cx, and 2008ha. All 0.0 three SN 2008ha spectra are the +7 day Lick spectrum, but two 1.0 versions have been smoothed with 800 and 2000 km s−1 Gaussians. 05cs +304 d H" The spectra of SNe 1991T and 2002cx both have phases of +20 days 0.8 relative to B maximum. The spectra of SNe 1991T and 2002cx 0.6 have been redshifted (after being deredshifted by their recession 0.4 velocity) by velocities of 7500 and 3000 km s−1, respectively. 0.2 0.0 that all six objects are broadly consistent, having many 1.0 similar spectral features. In particular, all objects have 05E [Ca II] Ca II 0.8 08ha +62 d strong Na D and Fe II lines. The main differences are 0.6 that SN 2008ha has narrower lines than those of the other 0.4 objects, SNe 2005E, 2005cs, and 2008ha have strong II II 0.2 [Ca ] and Ca emission lines, SNe 2004aw, 2005E and 0.0 2005cs have [O I] emission, SN 2005cs has a strong Hα 4000 5000 6000 7000 8000 9000 line, and SN 2007J has prominent He I emission lines. Rest Wavelength (Å) The strong [Ca II] and Ca II lines in the latest SN 2008ha spectrum present an interesting clue to the composition of the SN ejecta. The late-time ejecta of Fig. 7.— Late-time optical spectra of SNe 1989B, 2004aw, 2005E, 2005cs, 2007J, and 2008ha (black line in all panels). The spectra of SNe II cool through line emission from H transitions. SNe 1989B and 2004aw have been redshifted (after correcting for II I −1 Similarly, SNe Ia and Ib/c cool through [Fe ] and [O ] their recession velocities) by a velocity of 3000 km s . All spec- lines, respectively. In all cases, these lines dominate the tra have many of the same features, particularly Fe II lines. The late-time spectra of the objects. Late-time spectra of main differences are low line velocities for SNe 2005cs, 2007J, and 2008ha, Hα emission for SN 2005cs, He I emission for SN 2007J, SNe 2002cx and 2005hk show no strong forbidden emis- strong Ca II and [Ca II] emission for SNe 2005E, 2005cs, 2007J, sion lines (although there are weak [Ca II] lines and lines and 2008ha, and [O I] lines for SNe 2005E and 2005cs. For clar- from other intermediate-mass elements), indicating that ity, we have interpolated over nebular emission lines from the host the SN ejecta were still quite dense at least a year af- galaxies. ter maximum light (Jha et al. 2006; Sahu et al. 2008). From our measured ejecta velocity and ejecta mass measurement in § 4, our +62 day spectrum should have an generate a synthetic spectrum, one inputs a blackbody 9 −3 electron density of ∼10 cm . Integrating the flux in temperature (TBB), a photospheric velocity (vph), and for the Ca lines, [Ca II]/Ca II ≈ 0.5. An electron temper- each involved ion, an optical depth at a reference line, an ature of ∼ 7000 K is appropriate for this density and excitation temperature (Texc), the maximum velocity of line ratio (Ferland & Persson 1989). At ∼200 days after the opacity distribution (v ), and a velocity scale (v ). − max e maximum, the electron density should be ∼108 cm 3, at This last variable assumes that the optical depth declines which point we should be able to measure the relative exponentially for velocities above vph with an e-folding abundance of O/Ca (Fransson & Chevalier 1989). How- scale of ve. The strengths of the lines for each ion are de- ever, if the ejecta are in dense clumps, the density may termined by oscillator strengths and the approximation stay high for a longer time. of a Boltzmann distribution of the lower-level popula- tions with a temperature of Texc. 3.3. SYNOW Model Fits In Figure 8, we present our +14 day spectrum of To investigate the details of our SN spectra, we use SN 2008ha with a synthetic spectrum generated from the supernova spectrum-synthesis code SYNOW (Fisher SYNOW. This fit has TBB = 5500 K and vph = et al. 1997). Although SYNOW has a simple, paramet- 600 km s−1, and eleven ions are used. Several ions, which ric approach to creating synthetic spectra, it is a power- we consider to be “secure” identifications, are important ful tool to aid line identifications which in turn provide for the spectral formation of SN 2008ha; they account for insights into the spectral formation of the objects. To either the most prominent or multiple line features, and SN 2008ha Had a Very Fast 6 Foley et al.

TABLE 7 Light Curve Light-Curve Fit Parameters Absolute Magnitude Oﬀset SN Name SN Type Stretch B V R I J H K SN 2002cx 02cx-like 0.77 3.7 3.2 3.3 3.2 · · · · · · · · · SN 2005hk 02cx-like 0.73 3.9 3.7 3.7 3.7 3.4 3.6 3.3 SN 2005hg Ib 0.70 3.4 3.3 3.4 3.3 4.0 4.0 3.5

!14.0 SN 2008ha ∆m15(B) = 2.17 +0.5 d 12 !14.5 ) 1 ! Å 2

! !15.0

cm UVM2 1

! R Fastest SN ever (by far) !15.5 B UVW2 V I

) (erg s UVW1 ! U !16.0 J 14 K ! 5 log(f H !16.5 No NIR second maximum K !17.0 0.2 0.4 0.6 0.8 1.0 2.0 H ! 4 Observed Wavelength (µm) 16 Fig. 3.— SED of SN 2008ha (solid crosses) constructed from UVOIR broad-band filters at t = 0.5 days (for the UV and optical, SN 2002cx matches when the NIR points are from t = 3.8 d). Each filter is labeled. The UV W 2 and UV M2 filters yield only upper limits. We also plot J ! 3 near-maximum SEDs of SN 2005cs (dashed crosses) and SN 2005hk “stretched” by 0.7. (dotted crosses) shifted to match the V -band flux of SN 2008ha. SN 2008ha has less UV flux than SN 2005hk and significantly less UV flux than SN 2005cs. 18 Apparent Brightness (mag) I ! 1.5 R ! 1 telescope, the Blue Channel spectrograph (Schmidt et al. 1989) on the 6.5 m MMT telescope, and the Low Reso- Rise time is ~10 days lution Imaging Spectrometer (LRIS; Oke et al. 1995) on the 10 m Keck I telescope. Unf + 0.5 Standard CCD processing and spectrum extraction 20 were accomplished with IRAF. The data were extracted using the optimal algorithm of Horne (1986). Low-order V polynomial fits to calibration-lamp spectra were used to establish the wavelength scale, and small adjustments derived from night-sky lines in the object frames were UVW1 U B applied. For the MagE spectra, the sky was subtracted from the images using the method described by Kelson 0 20 40 60 80 100 (2003). We employed our own IDL routines to flux cal- ibrate the data and remove telluric lines using the well- Rest Days Relative to B Max exposed continua of the spectrophotometric standards (Wade & Horne 1988; Foley et al. 2003). Fig. 2.— [UV W 1]UBVRIJHK (circles and squares) and unfiltered (crosses, with the label “Unf”) light curves of SN 2008ha. In this Section, we will present spectra from the liter- Our unfiltered magnitudes closely approximate the R band. The ature, including a spectrum from Perets et al. (in prep., uncertainties for most data points are smaller than the plotted 2009), and previously unpublished spectra of SNe 2005E, symbols. Also plotted are comparison light curves of SNe 2002cx 2005cs, and 2007J. Our spectra of SNe 2005E and 2005cs (dashed lines), 2005hg (dotted lines), and 2005hk (solid lines) after applying a stretch of 0.77, 0.70, and 0.73 (respectively) and offset were obtained with LRIS mounted on Keck I on 2005 to match the peak in each band. Our PANIC and NIRI JHK points Mar. 11.3 and 2006 Apr. 27.5, respectively. Our spectra are plotted as squares to distinguish them from the lower signal-to- of SN 2007J were obtained with LRIS on 2007 Jan. 21.4 noise ratio PAIRITEL points. The Swift [UV W 1]UBV points are also plotted as squares. We see that the light curves of SN 2008ha and 2007 Mar. 18.3. are well matched by those of the comparison light curves stretched by a factor of ∼0.7. 3.1. Spectroscopic Properties of SN 2008ha In Figure 4, we present several spectra of SN 2008ha, from the classification spectrum (Foley 2008) to a spectrum obtained in 2009 January (corresponding to phases SN 2008A 1 arcmin

– 50 –

– 56 –

Lick 40” V-band -20 B -19 NGC 634 DSS January 12, 2008 -18 SN 2008ha Had an Extremely2005hk Figure 1: Image of SN 2008A from the Lick Observator-17y 1 m Nickel telescope (left) compared to 56 the pre-explosionLowDS SLuminosityimage (right). (and-20 Ni Mass) – 50 – -19 V -18 – 56 – -17 MB = -13.74 mag -20 MV = -14.21 mag B -19 I max -18 Arnett’s Law gives M 2005hk -17 56 -3 M Ni = 3 x 10 Msun -20 1.0 VJ -19 Ejecta mass is ~0.2 Msun -18

) + constant (not 1.4 Msun) SN 2002cx ! -17

SN 2002cx (+25 days) -20

Log (f I SN 2005hk (+24 days) -19 H 0.1 SN 2008A (+28 days) max -18 M -18 SN 2008A -17 4000 5000 6000 7000 8000 9000 -20 0.8 1 1.2 1.4 1.6 1.8 2 Rest Wavelength (Å) m (B) J -19 ! 15 -18 Figure 2: February 12, 2008 Keck spectrum ofFigSN. 14.—2008AThe abscolomparute magnietuddets oof anSNealoguesIa at maxiSmuNm 2002cxlight in theandBV IJH bands -17 SN 2005hk at similar epochs showing strikingplhomogeneitotted versus the ydec(llineefrt);ate pandarameabsoluter ∆m15(teB).opticalThe blackmtriagnangleituds arFeigSe.Ns8e.—in Cthoemparison of spectra of SN 2005hk at phases of -5, +13, +24, and +55 days with redshift range 0.01 < z < 0.1 whose distances were calculated from their host galaxy radial similar epoch spectra of SN 2002cx from LFC. The spectra are plotted on a logarithmic ﬂux of SN 2008A and SN 2005hk showing these arevelocouitietlierss in the-20cocomparedsmic microwavetobackagrosampleund frame aofssumnormaling a HubbSNele constIaant of H = 72 scale an0d shifted by an arbitrary constant. The wavelengths of the spectra were shifted to km s−1 Mpc−1. The red circle in each panel corresponds to SN 2005hk.H (right; panel adapted from Phillips et al. 2007). -19 the SN rest frame using the heliocentric velocities of the host galaxies given in NED. -18 -17 0.8 1 1.2 1.4 1.6 1.8 2 !m (B) 4 15 Fig. 14.— The absolute magnitudes of SNe Ia at maximum light in the BV IJH bands plotted versus the decline rate parameter ∆m15(B). The black triangles arFeigS.N8e.—in Cthoemparison of spectra of SN 2005hk at phases of -5, +13, +24, and +55 days with redshift range 0.01 < z < 0.1 whose distances were calculated from their host galaxy radial similar epoch spectra of SN 2002cx from LFC. The spectra are plotted on a logarithmic ﬂux velocities in the cosmic microwave background frame assuming a Hubble constant of H = 72 scale an0d shifted by an arbitrary constant. The wavelengths of the spectra were shifted to km s−1 Mpc−1. The red circle in each panel corresponds to SN 2005hk. the SN rest frame using the heliocentric velocities of the host galaxies given in NED. SN 2008A 1 arcmin

– 50 –

– 56 –

) + constant (not 1.4 Msun) SN 2002cx ! -17

SN 2002cx (+25 days) -20

Log (f I SN 2005hk (+24 days) -19 H 0.1 SN 2008A (+28 days) max -18 M -18 SN 2008A -17 4000 5000 6000 7000 8000 9000 -20 0.8 1 1.2 1.4 1.6 1.8 2 Rest Wavelength (Å) m (B) J -19 ! 15 -18 Figure 2: February 12, 2008 Keck spectrum ofFigSN. 14.—2008AThe abscolomparute maSNgnietu d2008hadets oof anSNealoguesIa at maxiSmuNm 2002cxlight in theandBV IJH bands -17 SN 2005hk at similar epochs showing strikingplhomogeneitotted versus the ydec(llineefrt);ate pandarameabsoluter ∆m15(teB).opticalThe blackmtriagnangleituds arFeigSe.Ns8e.—in Cthoemparison of spectra of SN 2005hk at phases of -5, +13, +24, and +55 days with redshift range 0.01 < z < 0.1 whose distances were calculated from their host galaxy radial similar epoch spectra of SN 2002cx from LFC. The spectra are plotted on a logarithmic flux of SN 2008A and SN 2005hk showing these arevelocouitietlierss in the-20cocomparedsmic microwavetobackagrosampleund frame aofssumnormaling a HubbSNele constIaant of H = 72 scale an0d shifted by an arbitrary constant. The wavelengths of the spectra were shifted to km s−1 Mpc−1. The red circle in each panel corresponds to SN 2005hk.H (right; panel adapted from Phillips et al. 2007). -19 the SN rest frame using the heliocentric velocities of the host galaxies given in NED. -18 -17 0.8 1 1.2 1.4 1.6 1.8 2 !m (B) 4 15 Fig. 14.— The absolute magnitudes of SNe Ia at maximum light in the BV IJH bands plotted versus the decline rate parameter ∆m15(B). The black triangles arFeigS.N8e.—in Cthoemparison of spectra of SN 2005hk at phases of -5, +13, +24, and +55 days with redshift range 0.01 < z < 0.1 whose distances were calculated from their host galaxy radial similar epoch spectra of SN 2002cx from LFC. The spectra are plotted on a logarithmic flux velocities in the cosmic microwave background frame assuming a Hubble constant of H = 72 scale an0d shifted by an arbitrary constant. The wavelengths of the spectra were shifted to km s−1 Mpc−1. The red circle in each panel corresponds to SN 2005hk. the SN rest frame using the heliocentric velocities of the host galaxies given in NED. It is Difficult to Determine the Progenitor/Explosion Mechanism

10 Foley et al.

SN 2008ha had low luminosity, 0.3 07J 2007 Mar 18 He I 2.5 low kinetic energy91T , low ejecta 0.2 0.1 2.0 08ha mass, a fast light2000 curve, no H/He, 0.0

1.5 0.3 89B +79 d + Constant spectra similar! to SNe Ia and Ic 08ha 0.2 1.0

Relative f 0.1 08ha 0.5 800 0.0

SN 2002cx-like02cx objects span a 04aw +64 d ! 0.0 56 0.2 range of luminosity/3000 4000 5000Ni 6000mass,7000 8000 9000 Rest Wavelength (Å) 0.1 Relative f Na D have mostlyFig. 6late-type.— Optical spectra ohostsf SNe 1991T, 2002cx, and 2008ha. All 0.0 three SN 2008ha spectra are the +7 day Lick spectrum, but two 1.0 versions have been smoothed with 800 and 2000 km s−1 Gaussians. 05cs +304 d H" The spectra of SNe 1991T and 2002cx both have phases of +20 days 0.8 relative to B maximum. The spectra of SNe 1991T and 2002cx 0.6 have been redshifted (after being deredshifted by their recession 0.4 Foley et al.velo cinvestigatedity) by velocities of 7500 an d43 000 km s−1, respectively. 0.2 0.0 possibilities:that a lfallback,l six objects are electrbroadly coonnsist ent, having many 1.0 similar spectral features. In particular, all objects have 05E [Ca II] Ca II 0.8 08ha +62 d capture, .Ia,strong andNa D adeflagrationnd Fe II lines. The main differences are 0.6 that SN 2008ha has narrower lines than those of the other 0.4 objects, SNe 2005E, 2005cs, and 2008ha have strong II II 0.2 [Ca ] and Ca emission lines, SNe 2004aw, 2005E and 0.0 2005cs have [O I] emission, SN 2005cs has a strong Hα 4000 5000 6000 7000 8000 9000 Valenti et lial.ne, a ndprSNeferr2007Jedhas prao mmassiveinent He I emis sion lines. Rest Wavelength (Å) The strong [Ca II] and Ca II lines in the latest star originSN (fallback/EC)2008ha spectrum present an interesting clue to the composition of the SN ejecta. The late-time ejecta of Fig. 7.— Late-time optical spectra of SNe 1989B, 2004aw, 2005E, 2005cs, 2007J, and 2008ha (black line in all panels). The spectra of SNe II cool through line emission from H transitions. SNe 1989B and 2004aw have been redshifted (after correcting for II I −1 Similarly, SNe Ia and Ib/c cool through [Fe ] and [O ] their recession velocities) by a velocity of 3000 km s . All spec- lines, respectively. In all cases, these lines dominate the tra have many of the same features, particularly Fe II lines. The late-time spectra of the objects. Late-time spectra of main differences are low line velocities for SNe 2005cs, 2007J, and 2008ha, Hα emission for SN 2005cs, He I emission for SN 2007J, SNe 2002cx and 2005hk show no strong forbidden emis- strong Ca II and [Ca II] emission for SNe 2005E, 2005cs, 2007J, sion lines (although there are weak [Ca II] lines and lines and 2008ha, and [O I] lines for SNe 2005E and 2005cs. For clar- from other intermediate-mass elements), indicating that ity, we have interpolated over nebular emission lines from the host the SN ejecta were still quite dense at least a year af- galaxies. ter maximum light (Jha et al. 2006; Sahu et al. 2008). From our measured ejecta velocity and ejecta mass measurement in § 4, our +62 day spectrum should have an generate a synthetic spectrum, one inputs a blackbody 9 −3 electron density of ∼10 cm . Integrating the flux in temperature (TBB), a photospheric velocity (vph), and for the Ca lines, [Ca II]/Ca II ≈ 0.5. An electron temper- each involved ion, an optical depth at a reference line, an ature of ∼ 7000 K is appropriate for this density and excitation temperature (Texc), the maximum velocity of line ratio (Ferland & Persson 1989). At ∼200 days after the opacity distribution (v ), and a velocity scale (v ). − max e maximum, the electron density should be ∼108 cm 3, at This last variable assumes that the optical depth declines which point we should be able to measure the relative exponentially for velocities above vph with an e-folding abundance of O/Ca (Fransson & Chevalier 1989). How- scale of ve. The strengths of the lines for each ion are de- ever, if the ejecta are in dense clumps, the density may termined by oscillator strengths and the approximation stay high for a longer time. of a Boltzmann distribution of the lower-level popula- tions with a temperature of Texc. 3.3. SYNOW Model Fits In Figure 8, we present our +14 day spectrum of To investigate the details of our SN spectra, we use SN 2008ha with a synthetic spectrum generated from the supernova spectrum-synthesis code SYNOW (Fisher SYNOW. This fit has TBB = 5500 K and vph = et al. 1997). Although SYNOW has a simple, paramet- 600 km s−1, and eleven ions are used. Several ions, which ric approach to creating synthetic spectra, it is a power- we consider to be “secure” identifications, are important ful tool to aid line identifications which in turn provide for the spectral formation of SN 2008ha; they account for insights into the spectral formation of the objects. To either the most prominent or multiple line features, and It is Difficult to Determine the Progenitor/Explosion Mechanism

14 Foley et al. 10 Foley et al. tively, mounted on the Lick 3 m Shane telescope. The SN 2008haspectra o fhadSN 200 4lowgw wer luminositye obtained with L,R IS mounted 0.3 0.6 07J 2007 Mar 18 He I on the Kec2.5 k I telescope. Classiﬁcations based on these 91T 0.2 low kineticspectra energywere origina,ll ylowrepor tejectaed by Polla s et al. (1992), 02cx!like II Ia 0.1 Foley &2.0F08hailippenko (2005) and Filippenko & Foley mass, a fast light2000 curve, no H/He, 0.00.5

(2005), a1.5nd Foley et al. (2006) for SNe 1991bj, 2004gw, 0.3 89B +7991bg d !like Ib + Constant spectraand similar200! 5hn, res topect iSNevely. SN Ia199 1andbj was rIcecently identi- 08ha 0.2 ﬁed as a 1.0potential member of this class (Stanishev et al. Relative f 0.10.4 91T!like Ic 2007). SN 208ha004gw was originally classiﬁed as a probable 0.5 800 0.0

SN 2002cx-likeSN Ia with02cx“a n objectsumber of spect spanral pecul iaar ities” (Foley 04aw +64 d ! & Filipp0.0enko 2005), but56Gal-Yam (2005) suggested that 0.2 range of luminosity/3000 4000 5000Ni 6000mass,7000 8000 9000 0.3 it may be a SN Ic. FinallyRest, F Wavelengthilippenk (Å)o & Foley (2005) 0.1 Relative f reclassified the object as a SN Ia. Additional observing Fraction Na D have mostlyFig. 6late-type.— Optical spectra ohostsf SNe 1991T, 2002cx, and 2008ha. All 0.0 detaitlhsrecae nSNb2e00fo8hunda speinctrtahoarseethreef+er7encesday Liacks swpelecltrausmi,nbuFtot-wo 1.0 versions have been smoothed with 800 and 2000 km s−1 Gaussians. 0.205cs +304 d H" ley etThael.sp(2ec0t0ra9o)f. SINneF19ig91urTean1d12,0w02ecxcobmothpaharveetphehasseps ectof +r2a0 odafys 0.8 relative to B maximum. The spectra of SNe 1991T and 2002cx 0.6 thesehoavbejectbeesn troedssphiectftedra(aoftfeSNr be2in0g02decxre.dsThiheftedspbyectthreairsrhoecewssnion 0.4 Foley et al.velo cinvestigatedity) by velocities of 7500 an d43 000 km s−1, respectively. here demonstrate clearly that each of these SNe belongs 0.2 0.00.1 possibilities:to thethaSNt a l2fallback,l0s0ix2cxobcljectass.ar(eI ncielectrbrodenadltyalcolyon,nsaisstt enritk,inghavningummberany 1.0 05E [Ca II] Ca II of peculsimialarrSNspectclarsaslesfeahaturdesthei. Irnfirpasrttimculemarb,eralsl diobsjcoectvers haedve 0.8 II 08ha +62 d capture,in 1.Ia,9s9t1ro.)ng andINn aaddiD atdeflagrationndionFteo theslinese .14Tohebjectmasin, tdiherffereencesis oneare 0.6 that SN 2008ha has narrower lines than those of the other 0.40.0 potenotbijaectl ms,emSNbere :2SN005E2,00270J0.5cs, and 2008ha have strong II II 0.2 [Ca ] and Ca emission lines, SNe 2004aw, 2005E and 0.0 E S0 Sa!Sab Sb Sbc Sc Scd!Sm Irr SN2020050cs7Jha, wvhie c[Oh Iha] emd missainoyn,sSNimil2a0r0it5icses hatosSNa st2r0o0ng2cxHα 4000 5000 6000 7000 8000 9000 Valenti et lial.ne, a ndprSNeferr2007Jedhas prao mmassiveinent He I emis sion lines. at early times (Filippenko et al. 2007a), eventually de- Rest Wavelength (Å) The strong [Ca II] and Ca II lines in the latest Galaxy Type star originvelopSNed (fallback/EC)H20e08emhaissspioectn rluminespr(Fesilenippt enkan ionteteresalt.ing200cl7ueb),tun-o the composition of the SN ejecta. The late-time ejecta of Fig. 7.— Late-time optical spectra of SNe 1989B, 2004aw, 2005E, like any other SN in this class. Although this object Fi20g0.5c1s2, .2—007FJr,aacntdio2n0o0f8haos(tblgaaclkaxliinees iinnaallgpiavnenelsm).oTrphheoslpoegcictraal obfin SNe II cool through line emission from H transitionsI . foSr Nthee19(8su9Bb)calnadss2e0s0o4af wSNhav2e00b2ecexn-lrikedeshoibftjedcts(a(ftbelracckorhreicstionggrafmor), has many similarities to SN 2002cx, the presenceII of He I −1 Similarly, SNe Ia and Ib/c cool through [Fe ] and [O ] SNth1ei9r9r1ebcge-slsiikoen ovbeljoecittsie(sp)ubryplae vheilsotcoigtyraomf )3,0S0N0 k1m99s1T-.likAellosbpjeecc-ts linesliinesn i,tsressppectirvaelcay. sItns adoll ubtcaseso,nthesitseincllinesusdoionminan tehitshe tra have many of the same features, particularly Fe II lines. The late-time spectra of the objects. Late-time spectra of (bmluaeinhdisitffoegrreanmce)s, aSreNleowIIli(ngereveenlochitisietsogforarmSN),eS2N00e5cIsb, 2(y00el7lJo,wanhdis- 2008ha, Hα emission for SN 2005cs, He I emission for SN 2007J, class.SNSNe 2020020cx8haandal2s0o05hahks sshoomwenosigsnitroficangnftordibiddenfferencesemis- togram), SNIeI Ic (orangIIe histogram), and normal SNe Ia (black II strong Ca and [Ca ] emission for SNe 2005E, 2005cs, 2007J, fromsiSNon l2ines002(cxal.thoHugowhevtherer, ewaerebweleaievke[Cthaa t]tlhesineseadindfferlines- linaen)d. 2008ha, and [O I] lines for SNe 2005E and 2005cs. For clar- from other intermediate-mass elements), indicating that ity, we have interpolated over nebular emission lines from the host encestheareSNcausejectedamwaerinleystbilyl qtheuitelodensw-enere agtyleaexsplt oasiyoean rinaf- galaxies. SN 2t0er08mhaaxriamtherum tlihaghnt (bJyhadiffeteraencesl. 200i6n; Saprohgueniettaolr.s.2008). From our measured ejecta velocity and ejecta mass measurement in § 4, our +62 day spectrum should have an generate a synthetic spectrum, one inputs a blackbody 5.2. Host Galaxies of 9SN 2−0302cx-like Objects and SNe Ic have significantly different distributions from electron density of ∼10 cm . Integrating the flux in temperature (TBB), a photospheric velocity (vph), and for the Ca lines, [Ca II]/Ca II ≈ 0.5. An electron temper- noearmchailnSNvolevedIa.ion,SNaenIobptaindcal SNdept1h9a9t1Ta -lrefikere enceobjectline,s haanve Using the ex∼tended sample of SN 2002cx-like objects, ature of 7000 K is appropriate for this density and mexarcigitnaatilolyn tdiemfferperenatturdiest(rTiebutxc),iotnshe fmroamximnoumrmvaelloSNcitye oIfa. we inlvineestriagtaitoe(tFheerlahondst&gaPlaerxsisesono1f9t8hi9)s.clAatss∼. 2I0n0Tdaablysea9ft,er the opacity distribution (vmax), and a velocity scale (ve). 8 −3 In a similar study, Kelly et al. (2008) found that the lo- the detmaaxiilmedumho, sttheinfeloectrmroantiodensn foirty1s4homulemd bbeer∼s1a0ndcm1 po,-at This last variable assumes that the optical depth declines which point we should be able to measure the relative caextiponsonenoftSNiallye Ifao,r Ivb,eloacindtiesII abllofvoellovpwhedwittheh agnalea-fxoyldilignght, abundance of O/Ca (Fransson & Chevalier 1989). How- tential member is presented. We exclude SN 2007J from wshicaleleSNof eveI.c Twheerestmrengoretlhsikelofyttheo blienesfoundfor eainchtheionbrairgehde-test the hoevsert-g, iaflathexy ejaectnalaysairse, ibutn densitseinclclumuspsion, tdoheesdensnoittysimg-ay termined by oscillator strengths and the approximation stay high for a longer time. regofioansBoofltazmgaannlaxydi, sitndiributcatioingn omf othere mlowasersi-lveve prel opgoenipultao-rs nificantly change the results. tions with a temperature of Texc. 3.3. for SNe Ic than SNe Ib or II. Except for SN 1991bg-like None of the SN 200S2YcxN-lOikWe Mobojdectel Fs ithas ve elliptical In Figure 8, we present our +14 day spectrum of To investigate the details of our SN spectra, we use obSNject2s0,08nohaSNwithclaasssyhantshetaicsigspniectficarumntlygenerdifferatenedt frhoomst- hostst;hehoswupevererno, vina sopurectexrumtended-synthessamisplcoe,de1/SY14N(O3/W14()Foisb-her SYNOW. This fit has TBB = 5500 K and vph = galaxy mo−rphology distribution from SN 2002cx-like ob- jectsetarael.ho19st9ed7). inAlS0tho(ugS0h tSYhroNugOWh Saha)sgaalsaixmiesple,. Apacoram-et- 600 km s 1, and eleven ions are used. Several ions, which ric approach to creating synthetic spectra, it is a power- jectwes.coAnsltiderhougtohbtehes“seecurrese”ultisdenmtaiyficachationgnse, awriethimapsolritgahnttly parisfouln tsoaoml pltoeaiwdalsinegeneridenattiedficatusioinsngwthihechSNineturfonundprovinide difffoerr tenhetsspaectmplrael (foKArmIaTtioins biofaSNsed2t0o08lumha;inotheyusagccoalaunxiest f)oror insights into the spectral formation of the objects. To either the most prominent or multiple line features, and galaxies monitored by KAIT (which includes the host with increased numbers, we cannot say that SN 2002cx- of SN 2008ha). Comparing the host-galaxy morphology like objects have a different host-galaxy morphology dis- of various (sub)classes of SNe may indicate similarities tribution from any SN class other than SN 1991bg-like in progenitors. For consistency, we use the host-galaxy events. classification scheme presented by Leaman et al. (2009), We find that 12% of the normal SN Ia hosts in the derived mainly from NED, and we divide the galaxies KAIT sample are elliptical galaxies. If SN 2002cx-like into eight morphology bins. In Figure 12, the fraction of events have the same host-morphology distribution as host galaxies in these eight morphological bins for several normal SNe Ia, we would expect to have found 1.7 ± 1.3 SN classes is presented. SNe (1.8 ± 1.3 if we include SN 2007J) of the SN 2002cx The well-known results that SN 1991bg-like objects class in elliptical galaxies, consistent with our finding (Filippenko et al. 1992a; Leibundgut et al. 1993) tend zero members in elliptical galaxies. We find that 31% to be found in early-type galaxies (relative to normal of SNe Ia come from E through S0/a galaxies. If we SNe Ia; Howell 2001), while SNe Ib, Ic, II, and 1991T-like assume that SN 2002cx-like objects and normal SNe Ia SNe have distributions skewed to later-type host galaxies have the same host-galaxy morphology distribution, then relative to the distribution of normal SNe Ia, are immedi- we would expect 4.4 ± 2.1 SNe (4.7 ± 2.2 if we include ately apparent in Figure 12. SN 2002cx-like events have, SN 2007J) of the SN 2002cx class in early-type galaxies. on average, later-type host galaxies than normal SNe Ia. This prediction is consistent with our measurement of 2 Examining Figure 12, all distributions other than that members in early-type galaxies. of SN 1991bg-like SN host galaxies are somewhat similar. Performing a Kolmogorov-Smirnov (K-S) test on 5.3. The Host Galaxy of SN 2008ha the data, we find that SN 1991bg-like objects, SNe II, Can Constrain Progenitor with Additional Observations

Pre-explosion images • Most deﬁnitive

X-ray/radio detections/limits • Might expect relativistic ejecta or circumstellar interaction

Nebular spectrum? • Models have speciﬁc predictions, but nothing very satisfying yet

Earlier data? • Probe outer layers

Fig. 3.—Expanded views of the late-time spectra of SN 2002cx (blue), compared with Synow modeJhal spectra in cletuding Fal.e ii, Na i,2006and Ca ii (black) and O i (red). The shaded regions show line identifications (except those for Fe ii, which are too numerous) and two of the strongest unidentified lines. Fallback and Deflagration Are Most Promising

Fallback occurs when a Fallback Deflagration massive star undergoes core Progenitor >30 Msun WD collapse, creating a neutron star then shortly after the NS Metallicity Low ? Star collapses to a black hole Host All? forming Deflagration is a subsonic KE/56Ni/ Requires Requires thermonuclear burning front in Rise time/ failed balancing a white dwarf (SN 2008ha Ejecta mass explosion requires a failed deflagration) Early Unburned H/He? spectrum Material Nebular [O I], Ca II, [O I], [Fe II] spectrum [Ca II], H? Maximum-Light Spectrum Shows IMEs/Carbon

2 Foley et al.

5 Higher ejecta velocity (~4000 Si III 4 10 3 2 km/s), so larger KE, ejecta mass 1 S II 0 0.4 Si II Fe III 0.2 8 S II Not much Fe (still hot, showing S IINa D 0.0 Si II O I Si II C II !0.2 Fe III), as expected C II O I 0.4 0.3 6 0.2 08ha !1 d Plenty of IMEs, including sulfur, 0.1 + Constant ! which is only strong in SNe Ia 0.0 4 !

Scaled f 05hk !5 d !0.1

06gz !3 d !0.2 Clear detection of carbon, which 6000 7000 8000 9000 0.8 99aa !1 d Relative f 2 0.4 is rare in SNe Ia and never seen 0.0 04aw !2 d !0.4 at maximum light (now 09dc!) 5400 5600 5800 6000 6200 0.8 0.6 0 07gr !1 d 0.4 0.2 4500 5000 5500 6000 6500 7000 7500 8000 0.0 !0.2 “Spectrum is truth” - RPK Rest Wavelength (Å) 6400 6600 6800 7000 7200 − 1.2 Fig. 1.— Optical spectrum of SN 2008ha at t = 1 days rela- 0.8 tive to B maximum (solid black line) and the SYNOW model fit 0.4 (red dashed line). The spectrum is dominated by IMEs. Spectra 0.0 of SNe 1999aa (Garavini et al. 2004), 2004aw (Taubenberger et al. !0.4 7200 7400 7600 7800 8000 8200 2006), 2005hk (Chornock et al. 2006), 2006gz (Hicken et al. 2007), and 2007gr (Valenti et al. 2008) are also shown for comparison. To 0.8 match the ejecta velocity of SN 2008ha, the comparison spectra 0.4 have been redshifted (after deredshifting by their recession veloci- 0.0 !0.4 ties) by 9000, 12,000, 3000, 10,500, and 4500 km s−1, respectively. 8200 8400 8600 8800 9000 9200 Rest Wavelength (Å) flux. We convert the r-band instrumental magnitudes into R-band magnitudes using the R-band photometry Fig. 2.— Optical spectrum of SN 2008ha at t = 231 days relative to B maximum (solid black lines). (top panel): The extracted of local stanard stars (Paper I). No instrumental color spectrum of SN 2008ha. A linear fit to the continuum is shown as corrections were applied. a red dotted line. (second panel): The spectrum of SN 2008ha after subtracting the continuum. (third panel): The spectrum 3. SPECTRAL ANALYSIS from the middle panel is reproduced in black. The red line is the continuum-subtracted 227 day spectrum of SN 2002cx (Jha et al. The two new spectra of SN 2008ha are presented in 2006). (bottom four panels): The continuum-subtracted spectrum Figures 1 and 2. To compensate for the contamination of SN 2008ha from the second and third panels is reproduced in of an underlying H II region in the late-time spectrum, these panels on an enlarged scale and finer resolution to see de- we fit a line to the continuum. Subtracting the line from tailed features. The red and blue solid lines are the continuum- subtracted 227 day spectrum of SN 2002cx and the continuum- the spectrum, a residual spectrum, which is more useful subtracted 62 day spectrum of SN 2008ha (Paper I), respectively. for comparisons to other SN spectra, is created.

3.1. Maximum-Light Spectrum simple, parametric approach to creating synthetic spec- The maximum-light spectrum consists almost exclu- tra, it is a powerful tool to aid line identifications which sively of intermediate-mass elements (IMEs) with a low in turn provide insights into the spectral formation of ejecta velocity. The minima of the Si II λ6355 and the objects (see Paper I for details). SYNOW unam- O I λ7774 features are blueshifted by ∼ 3700 and biguously detects the presence of C, O, Na, Si, S, and 5100 km s−1, respectively (the absorption of the O I Fe, all common features in maximum-light SN Ia spectra feature is nearly flat, so there is a large uncertainty). (with the exception of carbon, which is rarely seen), in The Si II feature was much weaker a week later (the the early-time spectrum. Although strong Si II is the first spectrum from Paper I) and could not be directly hallmark of a SN Ia, there are examples of SNe Ic with measured, but the characteristic photospheric velocity of strong Si features (Taubenberger et al. 2006; Valenti et al. ∼2000 km s−1 is nearly half that of Si II in the maximum- 2008). light spectrum. The O I feature had a velocity that was The SYNOW fit is similar to those performed on t = 2.5 times larger at maximum than at a week after maxi- −1 and −5 day spectra of SNe 2002cx (Branch et al. mum brightness. 2004) and 2005hk (a SN 2002cx-like object; Chornock The identification of IMEs is confirmed by fitting the et al. 2006). Most of the differences in the fit are the spectrum with the supernova spectrum-synthesis code result of different wavelength ranges (e.g., contributions SYNOW (Fisher et al. 1997). Although SYNOW has a from Ca are not expected in the SN 2008ha spectrum). 2 Foley et al.

5 Si III 4 10 3 2 1 S II 0 0.4 Si II Fe III 0.2 8 S II S IINa D 0.0 Si II O I Si II C II 0.2 C II O I 0.4 0.3 6 0.2 Late-Time Data08ha 1 d Are Consistent 0.1 + Constant ! with Most Scenarios0.0 4 !

Scaled f 05hk 5 d 0.1

06gz 3 d 0.2 6000 7000 8000 9000 Attempted a late-time (230 days) 0.8 99aa 1 d Relative f 2 spectrum with Gemini (~3 hours 0.4 0.0 N&S), but still low S/N 04aw 2 d 0.4 5400 5600 5800 6000 6200 0.8 0.6 0 Object on top of an H II07gr region, 1 d 0.4 0.2 4500 5000 5500 6000 6500 7000 7500 8000 0.0 so could only get limits in 0.2 Rest Wavelength (Å) imaging 6400 6600 6800 7000 7200 − 1.2 Fig. 1.— Optical spectrum of SN 2008ha at t = 1 days rela- 0.8 tive to B maximum (solid black line) and the S56YNOW model fit 0.4 (red dashed liImagingne). The spec tshowsrum is dom ilownated by INiMEs. Spectra 0.0 of SNe 1999aa (Garavini et al. 2004), 2004aw (Taubenberger et al. 0.4 7200 7400 7600 7800 8000 8200 2006), 2005hk (Chornock et al. 2006), 2006gz (Hicken et al. 2007), and 2007gr (VClearalenti et a ldetection. 2008) are also s hofow ncarbon,for comparis owhichn. To 0.8 match the ejeiscta rarveloceit yinof SNeSN 200 8Iaha, andthe com neverparison s pseenectra 0.4 have been redshifted (after deredshifting by their recession veloci- 0.0 0.4 ties) by 9000, at12,0 0maximum0, 3000, 10,500, alightnd 4500 km s−1, respectively. 8200 8400 8600 8800 9000 9200 Rest Wavelength (Å) flux. We convert the r-band instrumental magnitudes into R-band magnitudes using the R-band photometry Fig. 2.— Optical spectrum of SN 2008ha at t = 231 days relative to B maximum (solid black lines). (top panel): The extracted of local stanard stars (Paper I). No instrumental color spectrum of SN 2008ha. A linear fit to the continuum is shown as corrections were applied. a red dotted line. (second panel): The spectrum of SN 2008ha after subtracting the continuum. (third panel): The spectrum 3. SPECTRAL ANALYSIS from the middle panel is reproduced in black. The red line is the continuum-subtracted 227 day spectrum of SN 2002cx (Jha et al. The two new spectra of SN 2008ha are presented in 2006). (bottom four panels): The continuum-subtracted spectrum Figures 1 and 2. To compensate for the contamination of SN 2008ha from the second and third panels is reproduced in of an underlying H II region in the late-time spectrum, these panels on an enlarged scale and finer resolution to see de- we fit a line to the continuum. Subtracting the line from tailed features. The red and blue solid lines are the continuum- subtracted 227 day spectrum of SN 2002cx and the continuum- the spectrum, a residual spectrum, which is more useful subtracted 62 day spectrum of SN 2008ha (Paper I), respectively. for comparisons to other SN spectra, is created.

5 Si III 4 10 3 2 1 S II 0 0.4 Si II Fe III 0.2 8 S II S IINa D 0.0 Si II O I Si II C II 0.2 C II O I 0.4 0.3 6 0.2 Late-Time Data08ha 1 d Are Consistent 0.1 + Constant ! with Most Scenarios0.0 4 !

Scaled f 05hk 5 d 0.1 4 Foley et al. 06gz 3 d 0.2 6000 7000 8000 9000 SED, this is consistent with the estimate from the early- Attempted a late-time (230 days) 0.8 ± × −3 56 99aa 1 d Relative f time light curve of (3.0 0.9) 10 M". The true Ni −3 2 0.4 mass may be larger than 10 M" if a significant per- spectrum with Gemini (~3 hours 18 0.0 centage of the bolometric luminosity is emitted at longer wavelengths. N&S), but still low S/N 04aw 2 d 0.4 5400 5600 5800 6000 6200 19 5. DISCUSSION AND CONCLUSIONS 0.8 0.6 The maximum-light spectrum provides a great deal of 0 Object on top of an H II07gr region, 1 d 0.4 0.2 insight regarding the nature of SN 2008ha. The spec- 4500 5000 5500 6000 6500 7000 7500 8000 0.020 trum has strong lines from IMEs, including silicon and so could only get limits in 0.2 Rest Wavelength (Å) sulfur. V09 cites the lack of obvious S II and weak Si II imaging 6400 6600 6800 7000 7200 in the t = 8 day spectrum of SN 2008ha as evidence − 1.221 56 that SN 2008ha was not a thermonuclear explosion. Our Fig. 1.— Optical spectrum of SN 2008ha at t = 1 days rela- 0.8 Co Decay observations show that SN 2008ha did generate a signif- tive to B maximum (solid black line) and the S56YNOW model fit 0.4 icant amount of Si and S, and was therefore likely the (red dashed liImagingne). The spec tshowsrum is dom ilownated by INiMEs. Spectra 0.0R Band Apparent Brightness (mag) result of a thermonuclear explosion. 0.422 of SNe 1999aa (Garavini et al. 2004), 2004aw (Taubenberger et al. The higher ejecta velocity at maximum light shows 7200 7400 7600 7800 8000 8200 2006), 2005hk (Chornock et al. 2006), 2006gz (Hicken et al. 2007), that SN 2008ha may have a larger ejecta mass and ki- and 2007gr (VClearalenti et a ldetection. 2008) are also s hofow ncarbon,for comparis owhichn. To 0.8 netic energy than previously measured (Paper I; V09). 23 match the ejecta velocity of SN 2008ha, the comparison spectra 0.4 A doubling of the ejecta velocity will increase the kinetic is rare in SNe Ia and never seen energy by a factor of 8 and double the ejecta mass. We have been redshifted (after deredshifting by their recession veloci- 0.0 1 0.4 therefore revise our estimates of the kinetic energy to be at maximum light − 0 50 100 150 200 250 × 49 ties) by 9000, 12,000, 3000, 10,500, and 4500 km s , respectively. 8200 8400 Rest8600 Days Relative 8800to B Max 9000 9200 1.8 10 ergs and Mej = 0.3M" with large uncertainties. From a week until two months after maximum bright- Fig. 3.— R-band ligRestht cu Wavelengthrve of SN 20 (Å)08ha. The circles and ness, the ejecta of SN 2008ha maintained a very simi- squares come from Paper I and V09, respectively. The lower limits lar and low velocity (Paper I). The much larger velocity flux. We convert the r-band instrumental magnitudes are from our new LDSS3 and GMOS photometry. The red dotted −3 56 in the maximum-light spectrum requires a large veloc- into R-band magnitudes using the R-band photometry Filinge.in2d.—icatOesptthiceadl escpaeycetxrupmectoefdSfrNom20g0e8nhearaatitngt 1=0 23M1 "daoyfs reNlai- tivien ttoheBSmNa(xwimituhmco(msoplliedteblγa-crkaylinteras)p.pi(ntogpapnadnael)s:olTarh-elikeextSrEacDt)e.d ity gradient. Benetti et al. (2004) suggested that a large of local stanard stars (Paper I). No instrumental color The solid and dashed black lines is the light curve of SN 2005hk velocity gradient for SNe Ia is indicative of the burning sp(ePcthriullmipsoeftSaNl. 22000087h; aS.ahAulienteaalr. fi2t00t8o) tshheiftceodnttionumuamtchis tshheowpneaaks front extending to large radii, suggesting more complete corrections were applied. brightness of SN 200s8ehcao.ndThpeadnaeslhed line has been scaled to match a trhede wdiodttthedofliSnNe. 20(08ha (see Pap)e: r TI hfoersdpeetcatirlsu)m, wohfilSe Nthe20s0o8lihda burning in objects with a high velocity gradient than aftleinre shuabstnraocttbinegentmheodciofinetdi.nuum. (third panel): The spectrum those with a low velocity gradient. The detection of car- 3. SPECTRAL ANALYSIS from the middle panel is reproduced in black. The red line is the bon in the maximum-light spectrum at the same velocity continuum-subtracted 227 day spectrum of SN 2002cx (Jha et al. as the IMEs suggests incomplete burning at these radii. The two new spectra of SN 2008ha are presented in 20i0ns6)t.r(umbotentomt/filfouterr paresneplso):nsTeshe. continuum-subtracted spectrum For SN 2008ha it is more likely that the steep velocity Figures 1 and 2. To compensate for the contamination Paper I found that the light curves of SNe 2005hk and gradient is the result of a steep density gradient with of 2S0N082ha008whera efrowmellt-mhe astecchedond iafntdhethliirgdhtpacurnelvseisofreSNpro2d0u0ce5dhkin little material at large velocities. of an underlying H II region in the late-time spectrum, thwesaesp“asnterlsetocnhed”an ebnylaragefdactscoarleoafn0d.7fi3n.erInreFsoigluurtioen3t,othesee Rde-- The late-time photometry shows that SN 2008ha did tailed features. The red and blue solid lines are the continuum- −3 56 we fit a line to the continuum. Subtracting the line from band light curve of SN 2008ha is compared to that of not produce more than a few times 10 M" of Ni, suSNbtrac2t0e0d52hk27 (dPahiy lslpipsectretumalo. f2S0N07;20Sa02hcxu aetndalt.he20c0o8n)tinbuoutmh- consistent with measurements based on the early-time the spectrum, a residual spectrum, which is more useful suwbtirtahcteadnd62wdiatyhosputecttrhiums osftrSetNc2hi00ng8h.a (TPheapesrtIr)et, rcehedspectliivgehlyt. light curve (Paper I; V09). The late-time spectrum shows for comparisons to other SN spectra, is created. curve is a good approximation of the early-time behavior that there are no extremely strong emission lines from since the SEDs are similar, but the objects had different the SN, however the relatively low S/N spectrum places 56 3.1. opacities, amounts of ejecta/ Ni, and kinetic energies, weak limits on such features. Maximum-Light Spectrum simcrpleae,tingpaaradimffeterrienc tapprdecloineachrattoe.crHeaowtievngers,ytnhethetlatice-tsipmec-e The strong carbon lines in the maximum-light spec- light curve, whose shape should be dominated mostly by trum indicates that there is unburned material at low The maximum-light spectrum consists almost exclu- trara,diitoiasctaivpeowdecaerfuly, tmoaoyl thao vaeidalinediffierdenentificastretticohnsfawcthiocr.h velocities, consistent with a deflagration (Gamezo et al. sively of intermediate-mass elements (IMEs) with a low inTtheurnejectproavoidef SNins2i0g0h2tcxs -linitkoe othebjectspsectappraeal rfotrombaetionptio-f 2003). The low ejecta and 56Ni mass are consistent with cally thick at very late times (Jha et al. 2006; Sahu et al. a deflagration of a white dwarf if SN 2008ha was a failed ejecta velocity. The minima of the Si II λ6355 and the200o8b)j,ectsosopa(seecityPeaffpecter sImfoary detstilladoils)m. inaSYte.NORegWarunadlesms,- deflagration that did not disrupt the progenitor star (Pa- O I λ7774 features are blueshifted by ∼ 3700 and bitgheuouscurlyvesdetshoectulsdtgheivepraesn enceindicaotfioCn,oO,f theNaex, pSiect, edS, ade-nd per I). −1 I cay rate at late times if SN 2008ha continues to behave It is clear that pre-maximum data is critical for un- 5100 km s , respectively (the absorption of the O Fe,likaellSNcom20m0o5hkn f.eatures in maximum-light SN Ia spectra derstanding the nature of SN 2008ha-like objects. Al- feature is nearly flat, so there is a large uncertainty). (witIhf tthehepexeaceptk lumioinnoosiftycaorfbaoSNn, wishipocwherisedrabryel5y6Nseen)i deca, yi,n though our observations are all consistent with a failed II then the luminosity in the tail of the light curveIIshould deflagration of a white dwarf, we can not completely rule The Si feature was much weaker a week later (the thebeeaporwlyer-tedimebysp56ectCorum(the. decaAlthoy ugprohductstroongf 56SiNi) decais tyhe. out other scenarios. If the progenitor of SN 2008ha was first spectrum from Paper I) and could not be directly haTllhemabrrkigohftnesa SNs ofIa,SNtheratelatree teximaesmplis estherofefSNoreediIcrectwiltyh a very massive star, one might expect a brightening in related to the 56Ni mass. Assuming the relationship be- X-rays or radio when the SN ejecta interacts with the measured, but the characteristic photospheric velocity of strong Si56features (Taubenberger et al. 2006; Valenti et al. −1 tween Ni mass and bolometric luminosity (Sutherland progenitor’s wind or shells from pre-explosion outbursts. ∼2000 km s is nearly half that of Si II in the maximum- 20&08W).heeler 1984), a solar SED for bolometric corrections, Future early-time X-ray and radio observations will help I and full γ-ray trapping, the late-time limits are consistent constrain the nature of similar events. Eventually we will light spectrum. The O feature had a velocity that was The SYNOW fit− is similar to those performed on t = 56 ≈ 3 −with M−Ni 10 M" (see Figure 3). This value is also detect a SN 2008ha-like object with deep pre-imaging 2.5 times larger at maximum than at a week after maxi- 1consandistent5wdaithy tshepectbrirgahtonesf SNs oef SN200220cx08ha(Baratnc∼6h0etdayasl. data and either detect or highly constrain the properties mum brightness. 20a0f4ter) aBndm2a0x0im5hkum.(aCSNonsi2der00i2ngcx-ltheikeuncerobjectta;inCtyhoorfnotheck of the progenitor star. The identification of IMEs is confirmed by fitting the et al. 2006). Most of the differences in the fit are the spectrum with the supernova spectrum-synthesis code result of different wavelength ranges (e.g., contributions SYNOW (Fisher et al. 1997). Although SYNOW has a from Ca are not expected in the SN 2008ha spectrum). Are SNe 2008ha and 2005E A new type of stellar explosion 3 Related?

Both have low luminosity, strong [Ca II], low ejecta/56Ni mass

SN 2005E was far from its S0/Sa host, so probably not core collapse

Helium at early times, no sulfur, and Mej < Msun, so Perets et al. 2009 Perets et al. says it is likely a He detonation Figure 1. The environment of SN 2005E (technical details about the observations can be found in SI Section 1). (a) NGC 1032, the host galaxy of SN 2005E, as observed by the Sloan Digital Sky Survey (SDSS), prior to the SN explosion. The galaxy is an isolated, edge-on, early-type spiral galaxy, showing no signs of star-formation activity, warping, or interaction. Its luminosity is dominated by the cumulative contribution of a multitude of low-mass old stars (yellow light in this image). Panels (a)-(c) are 275!! × 175!!; a scale bar is provided, north is up, and east due left. (b) The LOSS10 discovery of SN 2005E on Jan. 13, 2005 (shown in negative). Note the remote location of the SN (marked with a red arrow) with respect to its host, 22.9 kpc (projected) from the galaxy nucleus and 11.3 kpc above the disk, whose edge-on orientation is well determined (panel (a)). (c) An image of NGC 1032 in the light of the Hα emission line, emitted by interstellar gas ionized by ultraviolet (UV) radiation, and a good tracer of recent star formation. There are no traces of recent star-formation activity (usually appearing as irregular, compact emission sources) near the SN location or anywhere else in the host. Panels (d)-(g) are 64!! × 36!!; a scale bar is provided. (d) Zoom-in on the location of SN 2005E in pre-explosion SDSS r-band images. No source is detected near the SN location, marked with a yellow circle (radius 3!!; the astrometric uncertainty in the SN location is < 0.5!!). The SDSS catalog does not list any objects near that position (e.g., putative faint dwarf satellites of NGC 1032), down to a typical limit of r = 22.5 mag. (e-f) Deeper photometry of the SN location. A red image is shown in panel (e), while a UV (u-band) image is shown in panel (f). At the distance of NGC 1032, the point source upper limits we ﬁnd, Mr < −7.5(−6.9) and Mu! < −8.1(−7.1) mag at 3(2)σ, respectively, indicate that we would have detected faint star-forming galaxies or star-forming regions at the SN location, or indeed even individual massive red supergiant or luminous blue supergiant stars. (g) Zoom-in on the location of SN 2005E in Hα light (see panel (c) for details). No trace of star-formation activity is seen near the SN location. There Are Few Known Low-Luminosity Transients

relation has been quantiﬁed with high pre- There are only a few objects with !#$ AB#"">-J cision and the theory is well understood. In -11 > MV > -14 and 2 < t < 100 !## AB#""&;2 AB#""%*+ contrast, the MMRD does not enjoy the same Included are SN 2008ha, SN !#" quantity or quality of light curves as those 2005E, luminous red novae, very !!% AFH"&I& of Ia supernovae. Fortunately, dedicated on- AB#""D-M AB#""%G

:-; SNe Ia

! going nova searches in M31 and the P60- faint SNe II, and “SN imposters” !!& AB!C%D4 Faint CC SNe AB#"">K AB#""%L- AB!CCC*7 FasTING project (see 4) has vastly increased !!$ AB#""!1, § Volume probed is tiny (µ ≈ 30-33 AB#""%A the number of well-sampled light curves. BEF<""?' 5 !!# mag); SNe Ia detected in 10 45+6.78*/9 LRNe The large gap between the brightest novae LBV times the volume !!" :/?' (say MV 10) and the faintest supernovae BEF%C!?' :

3 There Are Few Known Low-Luminosity Transients

:-; SNe Ia

3 Concluding Remarks

• SN 2008ha is an extreme member of the SN 2002cx class

• SN 2008ha had very low luminosity, ejecta mass, kinetic energy

• Maximum-light spectrum suggests thermonuclear explosion

• All observations can be explained by a failed deﬂagration

• Links to Ca-rich (SN 2005E-like) SNe (varying C/O:He ratio?)

• Within a few years, there should be many more examples

• There will also be plenty of new weird things Many Related Talks/Posters

Roepke: Outcomes from Deﬂagration McCully: HST and Ground-Based Scenarios Observations of SN 2008A: A Peculiar 2002cx-like Type Ia Supernova Howell: Puzzles Presented by Superluminous and Subluminous Events Moriya: Faint Supernovae with Fallback Jha: Power of Nebular Spectroscopy Pastorello: Subluminous Core-Collapse and SN 2002cx, 2005hk Supernovae Chamulak: The Nucleosynthetic Timmes: On Type Ia Supernovae From Signature of Surface Detonation in Type The Collisions of Two White Dwarfs Ia Supernovae Timmes: Trends in 44 Ti and 56 Ni from Gonzalez: The Rise-Time of Core-Collapse Supernovae Underluminous SNe Ia Valenti: A Low Energy Core-Collapse Supernova without a Hydrogen Envelope