The Most Slowly Declining Type Ia Supernova 2001ay
Kevin Krisciunas, Weidong Li, Thomas Matheson, D. Andrew Howell, Maximilian Stritzinger, Greg Aldering, Perry L. Berlind, M. Calkins, Peter Challis, Ryan Chornock, Alexander Conley, Alexei V. Filippenko, Mohan Ganeshalingam, Lisa Germany, Sergio Gonzalez, Samuel D. Gooding, Eric Hsiao, Daniel Kasen, Robert P. Kirshner, G. H. "Howie" Marion, Cesar Muena, Peter E. Nugent, M. Phelps, Mark M. Phillips, Yulei Qiu, Robert Quimby, K. Rines, Jeffrey M. Silverman, Nicholas B. Suntzeff, Rollin C. Thomas, Lifan Wang
aa r X i v : . [ a s t r o - ph . C O ] J un The Most Slowly Declining Type Ia Supernova 2001ay
Kevin Krisciunas, Weidong Li, Thomas Matheson, D. Andrew Howell, , MaximilianStritzinger, , Greg Aldering, Perry L. Berlind, M. Calkins, Peter Challis, RyanChornock, Alexander Conley,
Alexei V. Filippenko, Mohan Ganeshalingam, LisaGermany, Sergio Gonz´alez, Samuel D. Gooding, Eric Hsiao, Daniel Kasen, RobertP. Kirshner, G. H. “Howie” Marion, Cesar Muena, Peter E. Nugent, M. Phelps, Mark M. Phillips, Yulei Qiu, Robert Quimby, K. Rines, Jeffrey M. Silverman, Nicholas B. Suntzeff, Rollin C. Thomas, and Lifan Wang ABSTRACT
We present optical and near-infrared photometry, as well as ground-basedoptical spectra and
Hubble Space Telescope ultraviolet spectra, of the Type Iasupernova (SN) 2001ay. At maximum light the Si II and Mg II lines indicatedexpansion velocities of 14,000 km s − , while Si III and S II showed velocities of Department of Physics & Astronomy, George P. and Cynthia Woods Mitchell Institute for Fundamen-tal Physics & Astronomy, Texas A&M University, 4242 TAMU, College Station, TX 77843-4242; [email protected], suntzeff@physics.tamu.edu, [email protected], [email protected]. Department of Astronomy, University of California, Berkeley, CA 94720-3411; [email protected], [email protected], [email protected], [email protected]. National Optical Astronomy Observatories, 950 N. Cherry Avenue, Tucson, AZ 85719-4933; [email protected]. Las Cumbres Observatory Global Telescope Network, 6740 Cortona Drive, Suite 102, Goleta, CA 93117;[email protected]. Department of Physics, University of California, Santa Barbara, CA 93106-9530. The Oskar Klein Centre, Department of Astronomy, Stockholm University, AlbaNova, 10691 Stockholm,Sweden; [email protected]. Dark Cosmology Centre, Niels Bohr Institute, Universityof Copenhagen, Juliane Maries Vej 30, 2100Copenhagen Ø, Denmark; [email protected]. Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720; [email protected],[email protected], [email protected], [email protected]. Fred L. Whipple Observatory, P. O. Box 97, Amado, AZ 85645; [email protected]. Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138; [email protected], [email protected], [email protected], [email protected]. Department of Astronomy, University of Colorado, Boulder, CO 80309; [email protected]. Swinburne University of Technology, 1 John Street, Hawthorn, VIC, 3122, Australia; [email protected]. Las Campanas Observatory, Casilla 601, La Serena, Chile; [email protected]. Department of Physics, University of California, Berkeley, CA 94720; [email protected]. National Astronomical Observatories of China, Chinese Academy of Sciences, Beijing 100012, China;[email protected]. Department of Astronomy, California Institute of Technology, Pasadena, CA 91125;[email protected]. − . There is also evidence for some unburned carbon at 12,000 km s − .SN 2001ay exhibited a decline-rate parameter ∆ m ( B ) = 0.68 ± B -band photometry at t & +25 d past maximum make it the most slowlydeclining Type Ia SN yet discovered. Three of four super-Chandrasekhar-masscandidates have decline rates almost as slow as this. After correction for Galacticand host-galaxy extinction, SN 2001ay had M B = − .
19 and M V = − . not overluminous in optical bands. In near-infrared bands it was overluminous only at the 2 σ level at most. For a rise timeof 18 d (explosion to bolometric maximum) the implied Ni yield was (0.58 ± α M ⊙ , with α = L max /E Ni probably in the range 1.0 to 1.2. The Ni yieldis comparable to that of many Type Ia supernovae. The “normal” Ni yield andthe typical peak optical brightness suggest that the very broad optical light curveis explained by the trapping of the γ rays in the inner regions. Subject headings: supernovae: general — supernovae: individual (SN 2001ay)
1. Introduction
Phillips (1993) first established that Type Ia supernovae (SNe) are standardizable candlesat optical wavelengths: there is a correlation between their absolute magnitudes at maximumlight and the rate at which these objects fade. This fact allowed Type Ia SNe to be usedto determine that the expansion of the Universe is currently accelerating (Riess et al. 1998;Perlmutter et al. 1999). More recently, we have discovered that in near-infrared (IR) photo-metric bands Type Ia SNe are standard candles (Meikle 2000; Krisciunas, Phillips, & Suntzeff2004a; Wood-Vasey, et al. 2008; Folatelli et al. 2010; Mandel, Narayan, & Kirshner 2011).Except for a fraction of the rapidly declining Type Ia SNe whose prototype is SN 1991bg(Filippenko et al. 1992; Leibundgut et al. 1993), which peak in the near-IR after the timeof B -band maximum, the near-IR absolute magnitudes at maximum light are at mostonly slightly dependent on the decline-rate parameter ∆ m ( B ) (Krisciunas et al. 2009;Folatelli et al. 2010).We do not understand well how to model a Type Ia SN. For much of the previousdecade we thought that (most) Type Ia SNe were carbon-oxygen white dwarfs that approachthe Chandrasekhar limit (1.4 M ⊙ ) owing to mass transfer from a nearby companion (for areview, see, e.g., Livio 2000). Some Type Ia SNe might be mergers of two white dwarfs(Iben & Tutukov 1984). Over the past few years the double-degenerate scenario has gainedprominence among many researchers. Unfortunately, it is not yet possible to determinewith a high degree of probability that a given Type Ia SN is a single or double-degenerate 4 –event, except perhaps for the few that produced more than a Chandrasekhar mass of fusionproducts; these would result from double-degenerate mergers. An excellent summary ofevidence for different kinds of progenitors of Type Ia SNe is given by Howell (2011).The spectroscopic classification scheme stipulates that Type I SNe do not exhibit obvioushydrogen in their spectra, and that Type II SNe do show prominent hydrogen (Minkowski1941); see Filippenko (1997) for a review. Type Ia SNe show Si II in absorption, blueshiftedfrom its rest wavelength at 6355 ˚A. Two objects have been shown to have strong Si II in ab-sorption and H α in emission: SN 2002ic (Hamuy et al. 2003) and SN 2005gj (Aldering et al.2006; Prieto et al. 2006a). The interpretation is that these objects are Type Ia SNe interact-ing with circumstellar material (CSM) or (less likely) with the general interstellar medium(ISM).SN 2006X was found to exhibit variable Na I D absorption (Patat et al. 2007). Twoother examples showing variable Na I absorption are SN 1999cl (Blondin et al. 2009) andSN 2006dd (Stritzinger et al. 2010). Light from the SNe is ionizing the circumstellar medium.This leads to recombination and line variations. It should be pointed out that SNe 1999cland 2006X have large reddening that may involve multiple scattering as well as normal dustextinction (Wang 2005; Goobar 2008).It is possible to relate the peak bolometric luminosity of a Type Ia SN and the impliedamount of radioactive Ni produced in the explosion. Most events generate between 0.4 and0.7 M ⊙ of Ni (Stritzinger et al. 2006). The remarkable SN 2003fg (also known as SNLS-03D3bb; Howell et al. 2006) was sufficiently overluminous that it was regarded to have beencaused by the explosion of more than 1.4 M ⊙ of carbon and oxygen. Three additional super-Chandrasekhar-mass candidates have recently been found: SN 2006gz (Hicken et al. 2007),SN 2007if (Yuan et al. 2010; Scalzo et al. 2010), and SN 2009dc (Yamanaka et al. 2009;Tanaka et al. 2010; Silverman et al. 2010; Taubenberger et al. 2011). As summarized byScalzo et al. (2010, Table 4), these objects can be up to a magnitude more luminous at opticalwavelengths than typical Type Ia SNe. Furthermore, three of these four have extremelyslowly declining light curves (∆ m ( B ) ≈ . Hubble Space Telescope (HST)
STIS spectra(from 2001 May 2 and May 9) have already been published by Foley, Filippenko, & Jha 5 –(2008), in a study of possible luminosity indicators in the ultraviolet (UV) spectra of TypeIa SNe.
2. Optical and IR Photometry
SN 2001ay was discovered on 2001 April 18.4 (UT dates are used throughout this paper)by Swift & Li (2001). Its position was α = 14 h m . s , δ = +26 ◦ ′ . ′′ (J2000), some10 . ′′ west and 9 . ′′ north of the nucleus of the spiral galaxy IC 4423. Basic informationon SN 2001ay is given in Table 1. Figure 1 shows the host galaxy and SN, and identifies anumber of field stars of interest.A sizable fraction of our optical photometry of SN 2001ay was obtained with the 0.76-m Katzman Automatic Imaging Telescope (KAIT) at Lick Observatory (Li et al. 2000;Filippenko et al. 2001). KAIT images are 6 . ′ × . ′ in size, with a scale of 0 . ′′ per pixel.For the first three nights of photometry with the Lick Observatory Nickel 1-m telescope, theCCD camera had a chip with 0 . ′′ pixels in 2 × . ′ × . ′ . Thereafter, a new chip was installed, having 0 . ′′ pixels in 2 × . ′ × . ′ . The two different CCD chips employed with the Nickel telescopehad considerably different quantum efficiencies at blue wavelengths.The U BV RI photometry of some field stars near SN 2001ay is found in Table 2. Cali-bration of optical photometry was accomplished from observations carried out on five pho-tometric nights (two with the CTIO 0.9-m, two with the Nickel, and one with the CTIO1.5-m telescopes) using the standard stars of Landolt (1992). A subset of our photometry(the KAIT data) has already been published by Ganeshalingam et al. (2010, Table 6), butwithout K- and S-corrections. Our data presented here were reduced in the iraf environ-ment. Transformation of the data to the system of Landolt (1992) was accomplished usingequations equivalent to those of Wang et al. (2009a). In their Table 1, the reader will findcharacteristic color terms used for the transformations. Our own direct determinations ofthese color terms are entirely consistent with their values.Table 3 gives the J s and H -band photometry of some of the field stars. We give theaverages obtained on seven photometric nights using the Las Campanas 1-m Swope telescope.The calibration was accomplished using a stand-alone version of daophot , some fortran iraf is distributed by the National Optical Astronomy Observatory, which is operated by the Associationof Universities for Research in Astronomy, Inc., under cooperative agreement with the National ScienceFoundation (NSF). J s = 13.718 ± H = 13.251 ± J = 13.718 ± H = 13.214 ± J s = 15.153 ± H = 14.881 ± J = 15.132 ± H = 14.807 ± U -band magnitude was determined from aperture photometry; no U -bandtemplate was available. For the final night of Nickel 1-m photometry, PSF magnitudeswere derived, but without template subtraction. Aperture photometry without templatesubtraction was conducted for all other optical images. To compare the photometry of different SNe, one should transform the data to the rest-frame equivalent by subtracting the so-called K-corrections (Oke & Sandage 1968; Hamuy et al.1993; Kim, Goobar, & Perlmutter 1996; Hogg et al. 2002; Nugent, Kim, & Perlmutter 2002).For each SN one computes the time since maximum brightness, and then obtains a rest-frametimescale by dividing the differential times by 1 + z , where z is the object’s heliocentric red-shift. We have worked out the optical K-corrections using actual spectra of SN 2001ay,adopting the Bessell (1990) filter profiles as reference; these are given in Table 5 and shownin Figure 2.To account for differences in SN photometry that result from the use of differenttelescopes, CCD chips, and filters, we use the method of spectroscopic corrections (S-corrections), as outlined by Stritzinger et al. (2002) and Krisciunas et al. (2003). For op-tical bands we adopt as reference the filter specifications of Bessell (1990). At minimum one constructs the effective filter profiles using the laboratory transmission curves of thefilters multiplied by an appropriate atmospheric function that includes atmospheric absorp-tion lines; then, one multiplies that result by the quantum efficiency of the CCD chip as afunction of wavelength. Telescope and instrument optics also contribute to the effective filter Due to the long delay in assembling the data and writing this paper, some images are no longer retrievablefrom magnetic storage media. Only aperture photometry was possible in 2001, when some of the data werereduced, owing to the lack (at the time) of host-galaxy templates. appropriate to Lick Observatory (elevation 1290 m). For the CTIO 0.9-m data we have useda different atmospheric function appropriate to Cerro Tololo (elevation 2215 m). In practice,one arbitrarily shifts the effective filter profiles toward longer or shorter wavelengths so thatsynthetic photometry based on spectra of standard stars gives color terms that match thoseobtained from actual photometry of photometric standards. We used spectra of 50 stars fromthe sample of Stritzinger et al. (2005) to calculate our synthetic magnitudes using an iraf script written by one of us (N.B.S.). This script was then used to calculate the S-correctionsbased on spectra of SN 2001ay itself.In Figure 3 we show the B - and V -band S-corrections for SN 2001ay. One adds thesecorrections to the photometry. As one can see in Figure 3, if an object like SN 2001aywere observed with the CTIO 0.9-m and Nickel 1-m telescopes, there could be photometricdiscrepancies as large as ∼ .
06 mag in the B band, depending on the epoch. In the V bandthe corresponding differences are much smaller.Even after the application of the S-corrections, we find that the Nickel 1-m B -banddata based on aperture magnitudes are systematically 0.09 mag fainter than the KAITmeasurements. For the first three nights of Nickel imaging we were able to eliminate thesesystematic differences by means of template subtraction. The images subsequently obtainedwith the Nickel 1-m telescope and its newer chip did not lend themselves to image subtractionusing the KAIT templates. (Depending on the number of field stars, the angular size of anearby galaxy in the field, and the seeing, the image rescaling and remapping algorithm canfail.) As a result, we have devised a third set of corrections to the photometry based on imageswith the new chip in the Nickel camera. Using addstar within the daophot package of iraf , we were able to add an artificial star to the KAIT image-subtraction templates at thepixel location of the SN. This artificial star can be scaled to give a standard magnitude of adesired value such as B = 17.97 and V = 17.14 mag, just about the brightness of the SN on2001 May 14 (JD 2,452,043.8), in the middle of the run of Nickel 1-m aperture photometry.Artificial stars having the identical PSF magnitude as the fake star at the location of theSN are placed at blank places in the KAIT templates using addstar . Since the Nickel 1-maperture photometry was typically derived with a software aperture of radius 10 pixels (px)and a sky annulus ranging from 12 to 20 px, we can then compare the aperture magnitudesof the fake SN with aperture photometry of the fake stars in the blank places of the image.We use a software aperture radius and sky annulus of identical size in arc seconds to thatused for the Nickel 1-m photometry. In this way, we can obtain an estimate of the systematic The atmospheric functions combine the mean broad-band extinction values with atmospheric absorptionlines. B -band magnitudes from Nickel 1-m aperture photometry ofthe SN were 0.08 mag fainter than what we would have measured without the presence ofthe host galaxy. Similarly, in the V band we found that the Nickel 1-m aperture magnitudeswere fainter by 0.05 mag. For the two nights of CTIO 0.9-m aperture photometry, when theSN was more than 1 mag brighter, the SN was too faint by 0.03 mag in B and V withoutthis correction. Thus, by applying the S-corrections and these additional corrections, wecan reconcile almost all systematic differences in the B - and V -band photometry obtainedwith different telescopes, using different cameras, and using different data-reduction methods(i.e., PSF magnitudes with image-subtraction templates vs. aperture magnitudes withoutsubtractions).Further justification of this method comes from a consideration of the optical photome-try from the first Nickel 1-m chip and the photometry from one night with the CTIO 1.5-mtelescope. Data from these four nights can be derived using the KAIT host-galaxy templates.In the B band, photometry of the SN using PSF magnitudes and image subtraction was 0.07to 0.16 mag brighter than aperture photometry using an annulus for the subtraction of thesky level. In the V band, the photometric values of the SN on the first three Nickel 1-mnights were 0.06 mag brighter than values derived from aperture photometry. This resultis contrary to typical experience; normally, aperture photometry of a SN is brighter thanexpected (not fainter) if the SN is located on top of some part of the host galaxy, becausethe light in the aperture is not entirely due to the SN. That this is not the case here mustbe due to the relative distributions of host-galaxy light at the SN position and in the skyannulus.Differences between aperture photometry and PSF photometry are usually larger inredder bands. In fact, similar experiments with adding artificial stars to the KAIT templatesof SN 2001ay show that the aperture magnitudes with the Nickel 1-m telescope are too brightby 0.01 mag in R and I for the size of the aperture and annulus used. We never obtainednear-IR subtraction templates with the camera used on the Las Campanas Observatory 1-mtelescope, and that camera has since been decommissioned. So, we must adopt the availablenear-IR aperture magnitudes.The I -band photometry is particularly problematic from 15 to 30 d after the time of B -band maximum. Consider the effective filter profiles shown in Figure 4. While the KAIT I -band filter very well approximates the Bessell I filter, the filter used on the Nickel 1-mtelescope does not. As the SN achieved its reddest optical colors, a significant excess amountof light is let through by the Nickel 1-m I -band filter. This leads to positive S-corrections;to correct such photometry to Bessell filter photometry requires making the Nickel 1-m 9 –photometry fainter. Our attempts to reconcile up to 0.35 mag differences between theKAIT and Nickel 1-m photometry were not successful. Apparently, the effective filter profileof the Nickel 1-m camera is even more nonstandard than our profile based on laboratorytraces of the filter, knowledge of the quantum efficiency as a function of wavelength, and ageneric atmospheric function applicable to Lick Observatory. While we list all of our opticalphotometry in Table 4, the Nickel 1-m I -band photometry is not included in our plots orused in the analysis.The R - and I -band S-corrections are shown in Figure 5. Note that the photometry inTable 4 includes the K- and S-corrections, plus the corrections mentioned above for CTIO0.9-m and Nickel 1-m photometry reduced using aperture magnitudes. Since the K-correctedphotometry from one night with the YALO 1-m telescope and four nights with the LCO 2.5-m telescope are in good agreement with photometry obtained with other telescopes, andwe have no effective transmission curves for the LCO 2.5-m filters used, we have derivedno further corrections for this small fraction of our photometry. The interpolated BV RI
K-corrections and S-corrections are listed in Table 6. In spite of the rationale outlined above to reconcile as much of the optical photometryas we can, for a derivation of the maximum brightness and decline rate of SN 2001ay werestrict ourselves to the photometry based on PSF magnitudes with image subtraction.After subtracting the derived K-corrections and adding the S-corrections, we scale the timesince maximum light to the rest frame by dividing by 1 + z . We derive a time of B -bandmaximum of JD 2,452,022.49 (= 2001 April 23.0), with an uncertainty of perhaps ± B -band data is ∆ m ( B )= 0.68 ± m ( B ) = 0.72 ± m ( B )= 0.78 ± m ( B )= 0.72 ± B - and V -band light curves of SNe 2001ay, 2005eq, and 2009dc are illustratedin Figure 6. For the first three weeks after maximum light SNe 2001ay and 2009dc wereremarkably similar, but at t ≈ B - and V -band light curves sheds light on the central density and progenitor mass ofsingle-degenerate explosions (H¨oflich et al. 2010), a comparison of the observational features At present, pipelines for SN surveys use template spectra at different epochs since maximum light, warpthem, and, adopting appropriate effective filter profiles, calculate the K- and S-corrections at the same time.
10 –and modeling of very slow decliners may yield similar insights.In Figure 7 we show the unreddened B − V colors of SN 2001ay, the unreddenedlocus of Lira (1995) and Phillips et al. (1999), and the ∆ m ( B ) = 0.83 mag locus ofPrieto, Rest, & Suntzeff (2006b). On the basis of some of the spectra published here,Branch et al. (2006) classified SN 2001ay as a “broad-line” Type Ia SN. This is the sub-type that Wang et al. (2009b) suggest is intrinsically redder than normal Type Ia SNe, oroccurs in dustier environments, with R V ≈ .
7. However, Figure 7 shows that, if anything,SN 2001ay is bluer than other normal Type Ia SNe.In Figure 8 we show the corresponding R - and I -band light curves of SN 2001ay, alongwith loci derived from the data of SNe 2005eq and 2009dc. The behavior of SN 2001ay inthese photometric bands is clearly more like that of SN 2009dc than of SN 2005eq. Note thatthe I -band secondary maximum of SN 2001ay is essentially as bright as the first maximum.Figure 9 shows the near-IR photometry of SN 2001ay and the other slow decliners,SNe 2005eq (Folatelli et al. 2010; Contreras et al. 2010) and 2009dc (Stritzinger et al. 2011).SN 1999aw (Strolger et al. 2002) and SN 2001ay were the first objects known to exhibit suchflat H -band light curves. (Both H -band light curves are admittedly somewhat ragged, how-ever.) SN 2005eq was the first to show such early near-IR maxima ( t ≤ − H -band brightness of SN 2009dc increased steadily over the time frame that the bright-ness of SN 2001ay was constant.
3. Spectroscopy
We obtained spectra of SN 2001ay on 20 dates using 9 different telescopes; 7 of thosenights had spectra with more than one telescope. Spectra were obtained with the 1.5-mtelescope of the Fred L. Whipple Observatory (FLWO) on 11 nights. Eight spectra weretaken on four nights with the MMT 6.5-m telescope; four with the 2.16-m telescope atthe Xinglong Station of the Beijing Astronomical Observatory (BAO); three with the LasCampanas Observatory 2.5-m du Pont telescope; two each with the Lick 3-m Shane reflector(Kast spectrograph) and
HST ; and one each with the Kitt Peak 2.1-m, Kitt Peak 4-m, andKeck II telescopes. The spectroscopic observations with the FLWO 1.5-m and the MMT aresummarized in Table 8; we may refer to this as the “CfA set.” A log of other spectroscopicobservations is given in Table 9. Some, but not all, of the data were obtained at (or closeto) the parallactic angle (Filippenko 1982) in order to minimize the effects of atmosphericdispersion.In Figure 12 we show a temporal sequence of spectra. Here we have combined the May 2 11 –spectrum from
HST with the April 29 spectrum from the KPNO 2.1-m telescope. The rathernoisy BAO spectra from May 20, 25, and 30 are not illustrated, but the May 10 spectrumfrom BAO fills a gap in the temporal coverage. We show additional spectra in Figure 13.Figure 14 includes a portion of our spectrum of SN 2001ay obtained with the KeckEchellette Spectrograph and Imager (ESI; Sheinis et al. 2002) from 2001 April 22 ( t = − − − . TheSi II line (rest wavelength λ = 6355 ˚A) and the Mg II line ( λ = 4481 ˚A) are blueshiftedby 14,000 km s − at this epoch. There is some gas of these species blueshifted as much as20,000 km s − . As shown in Figure 12, by t = +17 d the Si II absorption is only blueshiftedabout 10,000 km s − . In our Keck spectrum there is a small absorption dip at an observed wavelength of λ ≈ λ = 6580 ˚A) blueshifted by 12,000 km s − . However, other explanations are possible.An absorption feature at 6520 ˚A in our Galaxy or in the Earth’s atmosphere could leadto a misidentification, given that telluric lines were not removed from the Keck spectrum.One possible such feature is telluric absorption at 6515 ˚A due to atmospheric water vapor(Matheson et al. 2000, Appendix). We have better evidence of C II absorption in SN 2001ayfrom the λ = 4745 ˚A line, which is seen in the galaxy’s rest frame at 4555 ˚A, correspondingto an identical blueshift of 12,000 km s − . We see no evidence of the corresponding C II linewith λ = 7234 ˚A.One might expect the C II to be concentrated in the outer layers, and therefore ata higher velocity of approach than sulfur or silicon (Khokhlov, M¨uller, & H¨oflich 1993;Gamezo et al. 2003). But if there is mixing or clumpiness, most species could be observedover a broad range of velocities.In Figure 15 we see the two components of Na I D from gas in the Milky Way andgas in the host galaxy of SN 2001ay. The presence of these lines implies some nonzeroamount of reddening and extinction. We find equivalent widths of 0.090 ± ± B − V color excess, we find E ( B − V ) Gal = 0 . ± .
006 mag and E ( B − V ) host = 0 . ± . ± .
05 mag for E ( B − V ),which is a more realistic estimate of the uncertainty of the reddening toward SN 2001ay. For the rest wavelengths of these and other species, see Table 1 of Wang et al. (2006). In their table,however, the rest wavelength of Mg II should read 448.1 nm, not 447.1 nm.
12 – The Galactic component may be compared to E ( B − V ) = 0 .
019 mag obtained bySchlegel, Finkbeiner, & Davis (1998).In Figure 16 we see the (blueshifted) velocities of Si II λ λ ±
420 and ±
330 km s − . The λ λ v = 226 ±
32 km s − d − , while the λ v = 171 ±
35 km s − d − . The Si II velocity gradient is considerably higher than70 km s − d − , the criterion of Benetti et al. (2005) to include SN 2001ay with other “highvelocity gradient” Type Ia SNe.
4. Discussion4.1. Reddening
As mentioned above, the Galactic reddening along the line of sight to SN 2001ay is E ( B − V ) = 0 .
026 mag, and the host-galaxy component is E ( B − V ) = 0 .
072 mag. Thus, forSN 2001ay, E ( B − V ) total ≈ .
098 mag. Normal Galactic dust is characterized by an averagevalue of R V = A V /E ( B − V ) = 3 . R V = 3 . R V = 2 . ± . A B = 0 . ± . A V =0 . ± . A R = 0 . ± . A I = 0 . ± . A J = 0 . ± .
02, and A H = 0 . ± . Garnavich et al. (2004) give
BV I decline-rate relations for what was then the knownrange of ∆ m ( B ) for Type Ia SNe. The slowest decliner used by Prieto, Rest, & Suntzeff However, Poznanski et al. (2011) recently found that the uncertainty of the reddening derived from theNa I D lines may be substantially larger than this.
13 –(2006b) for their light-curve fitting template algorithm was SN 1999aa, with ∆ m ( B )= 0.83mag. For such a decline-rate parameter, the implied absolute magnitudes at maximum are M B = − . M V = − .
39, and M I = − .
85 on an H = 72 km s − Mpc − distancescale. Three of the four super-Chandrasekhar-mass Type Ia SN candidates discussed byScalzo et al. (2010) are significantly more luminous than this.In Figure 10 we show the absolute V -band magnitudes at maximum light of SNe 2001ay,2003fg, 2005eq, 2006gz, 2007if, and 2009dc, along with the V -band decline-rate relation ofGarnavich et al. (2004). Figure 11 gives the near-IR absolute magnitudes of Type Ia SNe attheir respective maxima, including SNe 2001ay and 2005eq. Three of the four super-Chandracandidates are overluminous, while all other very slow decliners have “normal” maximumbrightness.Based on our best optical photometry (using PSF magnitudes and image-subtractiontemplates), we find apparent magnitudes for SN 2001ay of B max = 16.71, V max = 16.64, R max = 16.65, and I max = 16.79. These values include the K-corrections and S-corrections.Given the extinction values listed above and the distance modulus of the host galaxy given inTable 1, at maximum light we find M B = − . ± . M V = − . M R = − .
10, and M I = − .
90 mag (with uncertainties of ± .
10 mag); thus, SN 2001ay was not overluminousat optical wavelengths.Typically, Type Ia SNe peak in the near-IR ∼ B -band maximum(Krisciunas et al. 2004b). SN 2005eq peaked at least 7 d prior to B maximum. We surmisethat SN 2001ay may have been ∼ .
12 mag brighter at maximum light than our first J s measurement, or J s (max) ≈ H band SN 2001ay exhibited a very flat lightcurve; we adopt H max = 16 . ± .
08 mag. The corresponding extinction-corrected absolutemagnitudes at IR maximum are M J s = − . ± .
10 and M H = − . ± .
10. Thesevalues may be compared to the mean values of all but the late-peaking fast decliners fromKrisciunas et al. (2009), namely h M J i = − .
61 and h M H i = − .
31 mag. The standarddeviations of the distributions of the near-IR absolute magnitudes are about ± .
15 mag.In the J s and H bands, SN 2001ay was overluminous by ∼ σ and 2 σ , respectively; hence,SN 2001ay was not statistically significantly brighter in the near-IR than other Type Ia SNe. In Figure 17 we show the bolometric light curve of SN 2001ay, based on our broad-bandphotometry. Though Type Ia SNe near maximum light have spectral energy distributionsthat peak at optical wavelengths, they also emit UV and IR light. One might scale the 14 –integrated flux by a factor of ∼ . D ≈
129 Mpc;see Table 1) and a scale factor of 1.15, we obtain a peak derived bolometric luminosity (4 πD times the bolometric flux) of 1 . × erg s − . Figure 17 also shows the bolometric lightcurves of SNe 2007if and 2009dc.At maximum light the luminosity produced by radioactive Ni is given by L max = α E Ni , (1)where E Ni is the energy input from the decay of Ni, evaluated at the time of bolometricmaximum. Arnett’s Rule implies that α = 1 (Arnett 1982), meaning that the gamma-raydeposition matches the bolometric flux at maximum light. However, the value of α canactually range from 0.8 to 1.3, depending on the explosion model. For a delayed detona-tion model α = 1.2 is appropriate (Khokhlov, M¨uller, & H¨oflich 1993, Fig. 36); see alsoH¨oflich & Khokhlov (1996). Howell et al. (2006) adopt α = 1.2.In Figure 18 we show the radioactive decay energy deposition function fit to the last fewpoints of the bolometric light curve. In this figure we also show the cases of complete trappingof the γ rays and complete γ -ray escape. We adopt a rise time of 18 d from explosion to bolo-metric maximum, comparable to observational results of Garg et al. (2007) for other TypeIa SNe. Hayden et al. (2010) find an average rise time in the B band of 17.38 d, with a rangeof 13 to 23 d, while the B -band rise time determined by Ganeshalingam, Li, & Filippenko(2011) is about 18 d. Figure 18 uses α = 1.0. The energy deposited by 1 M ⊙ of Ni into the explosion of a Type Ia SN is given byStritzinger & Leibundgut (2005) as E Ni (1 M ⊙ ) = (6 . × ) e − t R / . + (1 . × ) e − t R / . , (2)where t R is the rise time in days from the moment of explosion to the time of bolometricmaximum. The e -folding times of Ni and Co are 8.8 and 111.3 d, respectively. For t R =18 d and α = 1.0, L max = 2.07 × erg s − M − ⊙ ; for α = 1.2, L max = 2.48 × erg s − M − ⊙ . The writers of this section could not come to a consensus on the best value of α to adopt, so in whatfollows we give the results for α = 1.0 and 1.2. Baron et al. (2011) choose α = 0.9 for their SN 2001ay model.
15 –There was a nondetection of SN 2001ay on April 5.4, which was 17.1 rest-frame daysbefore the time of B -band maximum. The upper limit of the brightness on that date was R ≈ . ∼ R -band maximum. Although we havean image of the host galaxy at t = −
17 d, we would need to reach a much fainter magnitudelimit to place a useful constraint on the time of the explosion. The lack of premaximumphotometry does not allow us to determine the rise time of SN 2001ay.To obtain the Ni yield in solar masses, we simply divide the peak luminosity (1.20 × erg s − ) by the coefficient given above for the adopted rise time (e.g., 2 . × × α for t R = 18 d); the result is 0.58/ α M ⊙ of Ni. SN 2001ay certainly did not produce morethan a Chandrasekhar mass of Ni.A 5% uncertainty in the value of H leads to a 10% uncertainty in the luminositycalculated from the optical photometry. Given the uncertainties in the parameter α inEquation 1 and uncertainty in the adopted bolometric rise time, the minimum uncertaintyin the Ni yield is about ± .
15 M ⊙ divided by α .We would like to estimate the mass in the ejecta of SN 2001ay. Following Jeffrey (1999)and Stritzinger et al. (2006), the “fiducial time” ( t ) is the time scale for the ejecta to becomeoptically thin: t = (cid:18) M ej κq π (cid:19) v e , (3)where M ej is the ejecta mass, κ is the γ -ray mean opacity, q is a dimensionless scale factor,and v e is the e -folding velocity of an exponential model’s density profile. Stritzinger et al.(2006) adopt κ = 0.025 cm g − and v e = 3000 km s − . They also choose q = 1 /
3, meaningthat Ni was distributed throughout the ejecta; q = 1 for high concentrations of Ni at thecenter of the ejecta. Using the last three points in our bolometric light curve (from t = 49 to111 d), adopting α = 1.2, and using a Ni yield of 0.48 M ⊙ we derive a “fiducial time” of 66.2 ± α = 1.0 and a Ni yield of 0.58 M ⊙ gives t = 57.2 ± t = 66.2 d and the values of q , v e , and κ adopted by Stritzinger et al.(2006), the implied ejecta mass is 4.4 M ⊙ , which is clearly wrong. For t = 57.2 d and the Swift & Li (2001) incorrectly give the date of this nondetection as 2001 April 4.4.
16 –higher Ni yield, the implied ejecta mass is 3.3 M ⊙ . Adopting q = 1 instead could cut thesedown by a factor of three. Choosing q = 1 / v e = 1500 km s − would give 0.8 to 1.1 M ⊙ for the ejecta mass. The lower e -folding velocity is only slightly less than what one derivesfrom the three-dimensional models of R¨opke & Hillebrandt (2004). Given the uncertaintiesof the parameters necessary for Eq. 3 and the lack of high-quality photometry from ∼
50 to100 d after maximum light, we feel that a robust calculation of the ejecta mass of SN 2001ayis beyond the scope of this paper.We note that Taubenberger et al. (2011) derived an ejecta mass for SN 2009dc of2.8 M ⊙ , based on the expansion velocity and the timescale around maximum brightness,which depends on the optical opacity. They describe their result as “an utmost challenge forall scenarios that invoke thermonuclear explosions of white dwarfs.” Wagers, Wang, & Asztalos (2010) have used wavelet analysis to search for spectroscopiccorrelations of Type Ia SNe. Their results are based on the analysis of a few dozen relativelynearby objects ( z . . m ( B ).In Figure 19 we illustrate a comparison of the spectra of SNe 2001ay, 2005eq, and2009dc. The most obvious difference between SN 2001ay and these other two objects isthe much larger blueshift of Si II seen in SN 2001ay prior to B maximum. Also, the C IIabsorption in SN 2009dc is much stronger. SN 2005eq was apparently the hottest of thethree, given the stronger presence of doubly ionized iron and the weakness of singly ionizedspecies. This is more like the classical slow decliner SN 1991T (Filippenko 1997), which is,in fact, how Contreras et al. (2010, Table 1) classify SN 2005eq.In Table 10 we give a summary of some observational characteristics of SNe 2001ay,2005eq, and 2009dc. Given the divergence of the light curves of SN 2005eq seen in Figure 6compared to the other two objects, the larger decline-rate parameter of SN 2005eq (∆ m ( B )= 0.78 mag) from Contreras et al. (2010) is more likely correct than the value of ∆ m ( B )= 0.72 mag from Folatelli et al. (2010), though we have used the latter to plot SN 2005eqin Figure 10. SNe 2001ay and 2009dc have almost the same optical decline rate. SN 2001ayhad normal optical peak brightness and high Si II velocity. SN 2009dc was overluminous atoptical maximum and had much lower Si II velocity. Their H -band light curves are unlikethose of any other Type Ia SNe observed thus far, except for SN 1999aw. 17 –To investigate the details of our SN spectra, we use the SN spectrum-synthesis codeSYNOW (Fisher et al. 1997). Although SYNOW has a simple, parametric approach tocreating synthetic spectra, it is a powerful tool to aid line identifications which in turn provideinsights into the spectral formation of the objects. To generate a synthetic spectrum oneinputs a blackbody temperature ( T BB ), a photospheric velocity ( v ph ), and for each involvedion, an optical depth at a reference line, an excitation temperature ( T exc ), and the maximumvelocity of the opacity distribution ( v max ). Moreover, it assumes that the optical depthdeclines exponentially for velocities above v ph with an e -folding scale of v e . The strengthsof the lines for each ion are determined by oscillator strengths, and the approximation of aBoltzmann distribution of the lower level populations is set by the temperature T exc .In Figure 20, we present our − T BB = 20,500 K and v ph = 12 ,
000 km s − . Themajority of the observed features are well fit by the synthetic spectrum. The ions usedin the fit, as labeled in the figure, are commonly observed in the near-maximum spectraof Type Ia SNe (Branch et al. 2005) with the exception of C II. Although the inclusion ofC II in the fit produces an absorption feature at ∼ λ ∼ λ t ≈ +6 d and 23 d. Figure 21shows the result for the t ≈ +6 d spectrum. The model spectrum has T BB = 14,000 K and v ph = 11,000 km s − , and has all the regular ions observed in a Type Ia SN. C II is no longerneeded, but a relatively strong Na I line is now observed at ∼ t ≈
23 d spectrum (not shown) requires T BB = 8500 K and v ph = 9000 km s − .Finally, in Figure 22 we show a model fit to our Keck spectrum using the SYNAPPScode of Thomas et al. (2011). The physical assumptions SYNAPPS uses match those ofSYNOW (Fisher et al. 1997), so findings are restricted to identification of features and notquantitative abundances. But where SYNOW is completely interactive, SYNAPPS is auto-mated. This relieves the user from iterative adjustment of a large number of parameters (over50 variables) to gain fit agreement and assures more exhaustive searching of the parameterspace. SYNAPPS can be thought of as the hybridization of a SYNOW-like calculation with aparallel optimization framework, where spectral fit quality serves as the objective function tooptimize. A good fit constrains explosion models through interpretive spectral feature iden-tification, with the main result being the detection or exclusion of specific chemical elements.The velocity distribution of detected species within the ejecta can also be constrained.SYNAPPS indicates that SN 2001ay had a photospheric velocity of 10,800 km s − at 18 – t = −
5. Conclusions
We have presented the available spectra, as well as optical and near-IR photometry, ofSN 2001ay.While SN 2001ay is the most slowly declining Type Ia SN ever discovered, it was notoverluminous in optical bandpasses. In near-IR bands it was overluminous only at the 2 σ level at most. Unlike other very slow decliners such as SNe 2003fg, 2007if, and 2009dc,which were significantly overluminous, we do not have any evidence that SN 2001ay was asuper-Chandrasekhar-mass explosion.SN 2001ay showed evidence for C II, but it was much weaker than in SN 2009dc. Atearly times Mg II and Si II were observed in SN 2001ay at high velocity (14,000 km s − and higher). By contrast, spectra of SN 2009dc did not show large velocities for Mg II andSi II. On the basis of a small number of super-Chandrasekhar-mass candidates, it seems thatthese objects exhibit rather low velocities, possibly the result of retardation due to a shell ofmaterial arising from the disruption of the less massive white dwarf in a double-degeneratesystem.SN 2001ay produced (0.58 ± α M ⊙ of Ni, considerably less than a Chan-drasekhar mass. The value of α probably lies in the range 1.0 to 1.2. Naively, one mightconclude that SN 2001ay was a single-degenerate explosion, but this is hardly a firm con-clusion. The very broad light curve might be explained by the trapping of the γ rays in theexplosion. An explanation of the extremely slow decline will be discussed in a subsequentpaper (Baron et al. 2011).The work presented here is based in part on observations made with the NASA/ESA Hubble Space Telescope , obtained at the Space Telescope Science Institute, which is op-erated by the Association of Universities for Research in Astronomy, Inc., under NASAcontract NAS5-26555; the Cerro Tololo Inter-American Observatory and the Kitt Peak Na-tional Observatory of the National Optical Astronomy Observatory, which is operated bythe Association of Universities for Research in Astronomy, Inc. (AURA) under cooperativeagreement with the NSF; the MMT Observatory, a joint facility of the Smithsonian Institu-tion and the University of Arizona; the Fred L. Whipple Observatory; the Lick Observatory 19 –of the University of California; the Las Campanas Observatory; the Beijing AstronomicalObservatory; and the W. M. Keck Observatory, which was generously funded by the W. M.Keck Foundation and is operated as a scientific partnership among the California Instituteof Technology, the University of California, and NASA. We thank the staffs at these obser-vatories for their efficient assistance, Don Groom for taking some of the Nickel 1-m images,and Rachel Gibbons, Maryam Modjaz, Isobel Hook, and Saul Perlmutter for other observa-tional assistance. We are grateful to Peter H¨oflich, Alexei Khokhlov, and Eddie Baron forcomments on § REFERENCES
Aldering, G., et al. 2006, ApJ, 650, 510Arnett, W. D. 1982, ApJ, 253, 785Baron, E., et al. 2011, in preparationBeers, T. C., Kriessler, J. R., Bird, C. M., & Huchra, J. P. 1995, AJ, 109, 874Benetti, S., et al. 2005, ApJ, 623, 1011Bessell, M. S. 1990, PASP, 102, 1181Blinnikov, S. I., R¨opke, F. K., Sorokina, E. I., Gieseler, M., Reinecke, M., Travaglio, C.,Hillebrandt, W., & Stritzinger, M. 2006, A&A, 453, 229Blondin, S., Prieto, J. L., Patat, F., Challis, P., Hicken, M., Kirshner, R. P., Matheson, T.,& Modjaz, M. 2009, ApJ, 693, 207 20 –Branch, D., Baron, E., Hall, N., Melakayil, M., & Parrent, J. 2005, PASP, 117, 545Branch, D., et al. 2006, PASP, 118, 560Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245Contreras, C., et al. 2010, AJ, 139, 519Filippenko, A. V. 1982, PASP, 94, 715Filippenko, A. V. 1997, ARA&A, 35, 309Filippenko A. V., Li W., Treffers R. R., & Modjaz M. 2001, in Small-Telescope Astronomyon Global Scales, ed. W. P. Chen, C. Lemme, & B. Paczy´nski (San Francisco: ASP),121Filippenko, A. V., et al. 1992, AJ, 104, 1543Fisher, A., Branch, D., Nugent, P., & Baron, E. 1997, ApJ, 481, L89Folatelli, G., et al. 2010, AJ, 139, 120Foley, R., Filippenko, A. V., & Jha, S. W. 2008, ApJ, 686, 117Freedman, W., et al. 2001, ApJ, 553, 47Gamezo, V. N., Khokhlov, A. M., Oran, E. S., Chtchelkanova, A. Y., & Rosenberg, R. O.2003, Science, 299, 77Ganeshalingam, M., et al. 2010, ApJS, 190, 418Ganeshalingam, M., Li, W., & Filippenko, A. V. 2011, MNRAS, in pressGarg, A., et al. 2007, AJ, 133, 403Garnavich, P., et al. 2004, ApJ, 613, 1120Goobar, A. 2008, ApJ, 686, 103Hamuy, M., Phillips, M. M., Wells, L. A., & Maza, J. 1993, PASP, 105, 787Hamuy, M., et al. 2003, Nature, 424, 651Hayden, B. T., et al. 2010, ApJ, 712, 350Hicken, M., et al. 2007, ApJ, 669, L17 21 –H¨oflich, P., & Khokhlov, A. 1996, ApJ, 457, 500H¨oflich, P., et al. 2010, ApJ, 710, 444Hogg, D. W., Baldry, I. K., Blanton, M. R., & Eisenstein, D. J. 2002(arXiv:astro-ph/0210394)Howell, D. A., et al. 2006, Nature, 443, 308Howell, D. A. 2011, Nature Communics., in press (arXiv:1011.0441)Iben, I., Jr., & Tutukov, A. V. 1984, ApJS, 54, 335Jeffrey, D. J. 1999 (arXiv:astro-ph/9907015)Khokhlov, A., M¨uller, E., & H¨oflich, P. 1993, A&A, 270, 223Kim, A., Goobar, A., & Perlmutter, S. 1996, PASP, 108, 190Krisciunas, K., Phillips, M. M., & Suntzeff, N. B. 2004a, ApJ, 602, L81Krisciunas, K., Prieto, J. L., Garnavich, P. M., Riley, J.-L. G., Rest, A., Stubbs, C., &McMillan, R. 2006, AJ, 131, 1639Krisciunas, K., et al. 2003, AJ, 125, 166Krisciunas, K., et al. 2004b, AJ, 128, 3034Krisciunas, K., et al. 2007, AJ, 133, 58Krisciunas, K., et al. 2009, AJ, 138, 1584Landolt, A. U. 1992, AJ, 104, 340Leibundgut, B., et al. 1993, AJ, 105, 301Li, W., et al. 2000, in Cosmic Explosions, ed. S. S. Holt & W. W. Zhang (New York: AIP),103Lira, P. 1995, Master’s Thesis, Univ. ChileLivio, M. 2000, in The Greatest Explosions since the Big Bang: Supernovae and Gamma-RayBursts, ed. M. Livio, N. Panagia, & K. Sahu (Baltimore: STScI), 334Maeda, K., et al. 2010, Nature, 466, 82 22 –Mandel, K. S., Narayan, G., & Kirshner, R. P. 2011, ApJ, in press (arXiv:1011.5910)Matheson, T., Filippenko, A. V., Ho, L. C., Barth, A. J., & Leonard, D. C. 2000, AJ, 120,1499Meikle, W. P. S. 2000, MNRAS, 314, 782Minkowski, R. 1941, PASP, 53, 224Munari, U., & Zwitter, T. 1997, A&A, 318, 269Nugent, P., Kim, A., & Perlmutter, S. 2002, PASP, 114, 803Oke, J. B., & Sandage, A., 1968, ApJ, 154, 21Patat, F., et al. 2007, Science, 317, 924Perlmutter, S., et al. 1999, ApJ, 517, 565Persson, S. E., Murphy, D. C., Krzeminski, W., Roth, M., & Rieke, M. J. 1998, AJ, 116,2475Phillips, M. M. 1993, ApJ, 413, L105Phillips, M. M., Lira, P., Suntzeff, N. B., Schommer, R. A., Hamuy, M., & Maza, J. 1999,AJ, 118, 1766Phillips, M. M., et al. 2003, in From Twilight to Highlight, ed. W. Hillebrandt, B. Leibundgut(Berlin: Springer-Verlag), 193Poznanski, D., Ganeshalingam, M., Silverman, J. M., & Filippenko, A. V. 2011, MNRAS,in pressPrieto, J. L., et al. 2006a (arXiv:0706.4088)Prieto, J. L., Rest, A., & Suntzeff, N. B. 2006b, ApJ, 647, 501Riess, A. G., et al. 1998, AJ, 116, 1009Riess, A. G., et al. 2009, ApJ, 699, 539R¨opke, F. K., & Hillebrandt, W. 2004, A&A, 420, L1Scalzo, R. A., et al. 2010, ApJ, 713, 1073Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525 23 –Sheinis, A. I., Bolte, M., Epps, H. W., Kibrick, R. I., Miller, J. S., Radovan, M. V., Bigelow,B. C., & Sutin, B. M., 2002, PASP, 114, 851Silverman, J. M., Ganeshalingam, M., Li, W. D., Filippenko, A. V., Miller, A. A., & Poz-nanski, D. 2010, MNRAS, 410, 585Skrutskie, M. F., et al. 2006, AJ, 131, 1163Stritzinger, M., et al. 2002, AJ, 124, 2100Stritzinger, M., & Leibundgut, B. 2005, A&A, 431, 423Stritzinger, M., Suntzeff, N. B., Hamuy, M., Challis, P., Demarco, R., Germany, L., &Soderberg, A. M. 2005, PASP, 117 810Stritzinger, M., Leibundgut, B., Walch, S., & Contardo, G. 2006, A&A, 450, 241Stritzinger, M., et al. 2010, AJ, 140, 2036Stritzinger, M., et al. 2011, in pressStrolger, L.-G., et al. 2002, AJ, 124, 2905Swift, B., & Li, W. D. 2001, IAU Circ. 7611Tanaka, M., et al. 2010, ApJ, 714, 1209Taubenberger, S., et al. 2011, MNRAS, in press (arXiv:1011.5665)Thomas, R., et al. 2011, PASP, in pressWagers, A., Wang, L., & Asztalos, S. 2010, ApJ, 711, 711Wang, L. 2005, ApJ, 635, L33Wang, L., Baade, D., H¨oflich, P., Wheeler, J. C., Kawabata, K., Khokhlov, A., Nomoto, K.,& Patat, F. 2006, ApJ, 653, 490Wang, X., et al. 2009a, ApJ, 697, 380Wang, X., et al. 2009b, ApJ, 699, L139Wood-Vasey, W. M., et al. 2008, ApJ, 689, 377Yamanaka, M., et al. 2009, ApJ, 707, L118 24 –Yuan, F., et al. 2010, ApJ, 715, 1338
This preprint was prepared with the AAS L A TEX macros v5.2.
25 –Table 1. Properties of SN 2001ay and its Host GalaxyParameter ValueHost galaxy IC 4423Host-galaxy type a SbcHeliocentric radial velocity b − CMB-frame radial velocity c − Distance modulus d . ± . E ( B − V ) Gal ± E ( B − V ) host ± α (J2000) 14 h m s . δ (J2000) +26 ◦ ′ . ′′ . ′′ . ′′ B -band maximum 2,452,022.49 ± . B -band maximum 2001 April 23.0∆ m ( B ) 0.68 ± M B (max) − . ± . M V (max) − . ± . M R (max) − . ± . M I (max) − . ± . M J (max) − . ± . M H (max) − . ± . a From HyperLEDA. b Beers et al. (1995), via NED. c From NED. d Using H = 72 km s − Mpc − (Freedman et al. 2001). 26 –Table 2. Optical Field-Star Sequence near SN 2001ay a ID b U (mag) B (mag) V (mag) R (mag) I (mag)1 17.112 (0.043) 16.399 (0.019) 15.480 (0.008) 14.859 (0.018) 14.386 (0.015)2 18.224 (0.047) 17.262 (0.020) 16.193 (0.007) 15.467 (0.019) 14.921 (0.016)3 16.705 (0.022) 16.846 (0.019) 16.367 (0.008) 15.981 (0.022) 15.622 (0.021)4 17.933 (0.035) 18.271 (0.030) 17.864 (0.005) 17.503 (0.025) 17.131 (0.025)5 19.246 (0.174) 18.729 (0.027) 17.884 (0.015) 17.302 (0.024) 16.796 (0.033)6 16.813 (0.031) 16.853 (0.017) 16.320 (0.007) 15.917 (0.017) 15.574 (0.014)7 20.437 (0.129) 19.585 (0.015) 18.313 (0.010) 17.487 (0.017) 16.730 (0.010)8 20.137 (0.230) 19.987 (0.196) 18.674 (0.070) 18.110 (0.100) 17.467 (0.111)9 21.270 (0.289) 19.802 (0.108) 18.790 (0.059) 18.009 (0.101) 17.467 (0.074)10 . . . . . . . . . 20.719 (0.155) 18.930 (0.061) a Magnitude uncertainties (1 σ ) are given in parentheses. b The IDs are the same as in Fig. 1. Star 10, located 18 . ′′ SE of Star 7, is not visiblein the V -band finder.Table 3. Infrared Field Star Sequence near SN 2001ay a ID b J s (mag) H (mag)1 13.718 (0.014) 13.251 (0.011)6 15.153 (0.021) 14.881 (0.008)7 15.869 (0.007) 15.281 (0.007)8 16.566 (0.030) 15.932 (0.031)10 17.838 (0.024) 17.359 (0.029) a Magnitude uncertainties (1 σ ) aregiven in parentheses. b The IDs are the same as in Fig. 1and Table 2. 27 –Table 4. Fully Corrected Optical Photometry of SN 2001ay aJD b U (mag) B (mag) V (mag) R (mag) I (mag) Telescope c Subs d a The
BV RI data in this table include the K-corrections, S-corrections, and corrections applied to photometry derivedusing aperture magnitudes without subtraction templates. Magnitude uncertainties (1 σ ) are given in parentheses. b Julian Date minus
28 – c d Image-subtraction templates used? Y = yes; N = No.
29 –Table 5. K-Corrections for SN 2001ay in Optical Bands aUT Date (2001) Phase (d) b Telescope ∆ B (mag) ∆ V (mag) ∆ R (mag) ∆ I (mag)Apr 21 − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − a These corrections (measured in magnitudes) are subtracted from the photometric data to correct thephotometry to the rest frame. These were calculated using the Bessell (1990) filter profiles. b The number of observer-frame days since the time of B maximum, 2001 April 23.
30 –Table 6. Interpolated Corrections to Optical Photometry aJD b K C B K C V K C R K C I S C B S C V S C R S C I Telescope c − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − a Values in Columns 2 through 9 are in magnitudes. The K-corrrections are subtracted from the data inTable 4. The S-corrections are added to the data in Table 4. b Julian Date minus c
31 –Table 7. Near-Infrared Photometry of SN 2001ay a JD b J s (mag) H (mag) Telescope c a Magnitude uncertainties (1 σ ) are given in parenthe-ses. b Julian Date minus c Table 8. Spectroscopic Observations in the “CfA Set”
UT Date Julian Day Tel. Range Res. Pos.A. Par.A. Airmass Flux Std. Sky Seeing Slit Exp. Observer(s)(2001) (˚A) (˚A) ( ◦ ) ( ◦ ) ( ′′ ) ( ′′ ) (s)Apr 21.48 2452020.98 FLWO 3720–7540 7.0 71.00 70.02 1.48 Feige34 1–2 3 1200 CalkinsApr 24.31 2452023.81 FLWO 3720–7540 7.0 0.00 − − − − −
33 –Table 9. Other Spectroscopic Observations
UT Date (2001) Phase (d) a Telescope Wavelength Range (˚A) Resolution (˚A)Apr 22 − HST
HST a Number of observer frame days since time of B -band maximum, 2001 April 23. Table 10. Comparison of Three ObjectsParameter SN 2001ay SN 2005eq SN 2009dc∆ m ( B ) 0.68 (0.05) 0.72–0.78 0.72 (0.03) M V (max) − − − a (km s − ) − − − − − − H -band light curve flat peaked early increasing a For the λ = 6355 ˚A line. The values in brackets correspond to the numbersof rest-frame days after the time of B maximum when the line was measuredat this velocity. For SN 2009dc the velocity is from Silverman et al. (2010). 34 –Fig. 1.— Finder chart for IC 4423, SN 2001ay, and some field stars in our Galaxy.Fig. 2.— K-corrections for BV RI photometry of SN 2001ay.Fig. 3.— S-corrections for B - and V -band photometry of SN 2001ay.Fig. 4.— Effective I -band transmission functions for the Bessell (1990) filter, the KAITfilter, and the Nickel 1-m filter.Fig. 5.— S-corrections for R - and I -band photometry of SN 2001ay.Fig. 6.— Comparison of the B - and V -band light curves of SNe 2001ay, 2005eq(Contreras et al. 2010), and 2009dc (Silverman et al. 2010).Fig. 7.— Unreddened B − V colors of SN 2001ay vs. time since B -band maximum.Fig. 8.— Similar to Figure 6, except for the R and I bands.Fig. 9.— Near-IR photometry of SNe 2009dc (Stritzinger et al. 2011), 2001ay, and 2005eq(Contreras et al. 2010).Fig. 10.— Absolute V -band magnitudes at maximum light of six very slowly declining TypeIa SNe vs. the decline-rate parameter ∆ m ( B ). The decline-rate relation of Garnavich et al.(2004) is also shown.Fig. 11.— Near-IR absolute magnitudes at maximum light of Type Ia SNe.Fig. 12.— Temporal sequence of spectra of SN 2001ay.Fig. 13.— Additional spectra of SN 2001ay obtained with the Lick Observatory 3-m Shanetelescope, HST , and the Las Campanas Observatory 2.5-m du Pont telescope.Fig. 14.— A portion of our highest signal-to-noise ratio spectrum of SN 2001ay, taken withthe Keck ESI on 2001 April 22.Fig. 15.— Profile of the Na I D lines in our Keck ESI spectrum. The Milky Way andhost-galaxy components are clearly present.Fig. 16.— Blueshifted velocity of two Si II lines in the spectra of SN 2001ay around maximumlight.Fig. 17.— Bolometric luminosities ( L ) of SNe 2001ay, 2007if (Scalzo et al. 2010), and 2009dc(Taubenberger et al. 2011), measured in erg s − . 35 –Fig. 18.— The best-fit radioactive decay energy deposition function (middle solid line) tothe UVOIR light curve (blue squares) of SN 2001ay.Fig. 19.— Comparison of the optical spectra of SNe 2001ay, 2005eq, and 2009dc.Fig. 20.— SYNOW modeling of the optical spectrum of SN 2001ay at 1 d before B maximum.Upper curve = data; lower curve = SYNOW model. The dashed lines indicate the additionof C II to the model.Fig. 21.— SYNOW modeling of the optical spectrum of SN 2001ay 6 d after B maximum.Upper curve = data; lower curve = SYNOW model.Fig. 22.— SYNAPPS modeling of the optical spectrum of SN 2001ay 1 d before B maximum.The blue dashed line (model spectrum) can be shifted arbitrarily along the vertical axis tomatch the actual spectrum (shown in black). 36 –Krisciunas et al. Figure 1. Finder chart for IC 4423, SN 2001ay, and some field stars in ourGalaxy. This 7 . ′ × . ′ image is a 200 s V -band exposure obtained on 2001 May 23 withthe CTIO 1.5-m telescope. North is up and east is to the left. The SN is located 10 . ′′ Wand 9 . ′′ N of the galaxy core. 37 – K - c o rr ec ti on ( m a g ) B V I R
Krisciunas et al.
Figure 2. K-corrections for
BV RI photometry of SN 2001ay. Theseaccount for the shifting of the SN spectrum to longer wavelengths owing to the redshift ofthe host galaxy. These values are to be subtracted from the photometry. Values derivedfrom the high signal-to-noise ratio MMT spectra are plotted as larger symbols. 38 – -0.06-0.04-0.0200.020.040.06 ∆ B ∆ V CTIO 0.9-mKAITNickel new
Krisciunas et al.
Figure 3. S-corrections for B - and V -band photometry of SN 2001ay,based on spectra from the FLWO 1.5-m and the MMT. The individual points are shownonly for the KAIT corrections. The corrections for the older Nickel 1-m chip are essentiallythe same as those shown for the newer Nickel 1-m chip. S-corrections are added to thephotometry to correct it to the filter system of Bessell (1990). 39 – o )00.20.40.60.81 E ff ec ti v e I- b a nd t r a n s m i ss i on Bessell Nickel 1-mKAIT
Krisciunas et al.
Figure 4. Effective I -band transmission functions for the Bessell (1990)filter, the KAIT filter, and the Nickel 1-m filter used with the newer, higher quantumefficiency chip. Each filter profile has been normalized to 1.00 at maximum throughput.We also include the t = +23 d spectrum from Figure 13 (multiplied by 10 ); it showssignificant flux past 9000 ˚A, which is measured by the Lick 1-m system but excluded in theKAIT measurements. 40 – -0.06-0.04-0.0200.020.040.06 ∆ R ∆ I KAITCTIO 0.9-mNickel new
Nickel old
CTIO 0.9-mNickel KAIT
Krisciunas et al.
Figure 5. S-corrections for R - and I -band photometry of SN 2001ay.Corrections for the older Nickel 1-m chip are shown as a dotted line; corrections for thenew Nickel 1-m chip are shown as a dashed line. For the I -band corrections with the Nickel1-m filter and both CCD chips used in the camera, we can only use spectra that extend tosufficiently long wavelengths (our Keck spectrum, two spectra taken with the Lick 3-m, andone with the KPNO 4-m). For the Nickel 1-m photometry to match the KAIT photometry,we would need the S-corrections to be a factor of 3 larger than the derived values.S-corrections are added to the photometry to correct it to the filter system of Bessell (1990). 41 – B − B m a x -10 0 10 20 30 40 50 60Rest-frame days since B-band maximum V − V m a x SN 2005eqSN 2009dc
SN 2001ay
Krisciunas et al.
Figure 6. Comparison of the B - and V -band light curves of SNe 2001ay,2005eq (Contreras et al. 2010), and 2009dc (Silverman et al. 2010). The SN 2001ay dataare K-corrected and S-corrected, and the CTIO 0.9-m and Nickel 1-m aperture photometryincludes additional corrections discussed in the text. For SN 2001ay the upward pointingtriangles represent data which result from PSF photometry using image-subtractiontemplates. The downward pointing triangles are SN 2001ay data based on aperturephotometry. 42 – -10 0 10 20 30 40 50 60Rest-frame days since B-band maximum-0.200.20.40.60.811.21.4 ( B − V ) un r e dd e n e d SN 2001ay
Krisciunas et al.
Figure 7. Unreddened B − V colors of SN 2001ay vs. time since B -bandmaximum. Upward pointing triangles are data derived using image-subtraction templates.Downward pointing triangles represent data derived from aperture magnitudes withouttemplate subtractions. The dashed line is the locus from Prieto, Rest, & Suntzeff (2006b)for their most slowly declining object (with ∆ m ( B ) = 0.83 mag). The solid line is thecommonly used Lira-Phillips locus (Lira 1995; Phillips et al. 1999). 43 – R − R m a x -10 0 10 20 30 40 50 60Rest-frame days since B-band maximum00.511.52 I − I m a x SN 2001ay
Krisciunas et al.
Figure 8. Similar to Figure 6, except for the R and I bands. In the caseof SN 2005eq the data are taken in Sloan r and i filters, not in Kron-Cousins R and I . TheNickel 1-m I -band data are not shown. 44 – J s -10 0 10 20 30 40 50 60Rest-frame days since B-band maximum15161718 H SN 2009dcSN 2001aySN 2005eq
Krisciunas et al.
Figure 9. Near-IR photometry of SNe 2009dc (Stritzinger et al. 2011),2001ay, and 2005eq (Contreras et al. 2010). The solid line is an educated guess thatSN 2001ay may have been ∼ .
12 mag brighter in J s at maximum light than our earliest J s datum. 45 – ∆ m (B) (mag)-21-20-19-18-17 M V ( m a x ) Krisciunas et al.
Figure 10. Absolute V -band magnitudes at maximum light of six veryslowly declining Type Ia SNe vs. the decline-rate parameter ∆ m ( B ). The decline-raterelation of Garnavich et al. (2004) is also shown. 46 – -19.5-19-18.5-18-17.5-17 M J -19-18.5-18-17.5-17 M H ∆ m (B) (mag)-19-18.5-18-17.5-17 M K Krisciunas et al.
Figure 11. Near-IR absolute magnitudes at maximum light of Type IaSNe. While SNe 2001ay and 2005eq are more luminous than the average of other objects inthe H band, one would not consider them significantly overluminous. The H -band pointfor SN 2009dc corresponds to the same epoch as in the J -band plot, but SN 2009dcbrightened 0.26 mag in H over the following month. The diamond-shape points correspondto objects that peak in the near-IR later than B maximum. These objects are subluminousin all bands. See Krisciunas et al. (2009) for more details. 47 – o A) − . l og ( F λ ) + c on s t a n t −2 −1 (Keck)+1+3 +3 (MMT)+4 +6 (KPNO2m+HST)+7 +8 (LCO)+9 +17 (BAO)+23 +24 (LCO)+30 +30 (KPNO4m)+31 (MMT)+32 +32 (MMT)+37+56 +56 (MMT) SN 2001ay
Krisciunas et al.
Figure 12. Temporal sequence of spectra of SN 2001ay. The labelsindicate the number of observer-frame days with respect to B -band maximum. Elevenwhich have no telescope indicated were taken with the FLWO 1.5-m. The t = +6 dspectrum from the KPNO 2.1-m telescope has been combined with the t = +9 d HST spectrum. The solid vertical line marks the nominal wavelength of the λ = 6355 ˚A line ofSi II. The dashed vertical line shows Si II line blueshifted by 10,000 km s − . Telluricoxygen is identified by an Earth symbol. 48 – o ) − . l og ( F λ ) + c on s t a n t +7 (Lick)+16 (HST) +23 (Lick)+33 (LCO) SN 2001ay
Krisciunas et al.
Figure 13. Additional spectra of SN 2001ay obtained with the LickObservatory 3-m Shane telescope,
HST , and the Las Campanas Observatory 2.5-m du Ponttelescope. The captioning within the figure and the meaning of the vertical lines are thesame as in Figure 12. Telluric oxygen is identified by an Earth symbol. 49 – o A) − . l og ( F λ ) Si III S II Na ISi II C II
SN 2001ay ( − S IIMg II C II
Krisciunas et al.
Figure 14. A portion of our highest signal-to-noise ratio spectrum ofSN 2001ay, taken with the Keck ESI on 2001 April 22. The dashed lines for Si III and S IIcorrespond to blueshifts of 9000 km s − . The dashed magenta lines for C II correspond to ablueshift of 12,000 km s − . The dot-dashed red lines for Mg II and Si II correspond to ablueshift of 14,000 km s − . The dotted line for Na I is the rest wavelength. 50 – Observed Wavelength (Angstroms) N o r m a li ze d F l ux Galactic Host
Na I D absorption towards SN 2001ay
Krisciunas et al.
Figure 15. Profile of the Na I D lines in our Keck ESI spectrum. TheMilky Way and host-galaxy components are clearly present. 51 – -5 0 5 10 15 20Time since B-band maximum (days)-16000-14000-12000-10000-8000-6000 S ili c on v e l o c it y ( k m / s ec ) λ 6355λ 4130 Krisciunas et al.
Figure 16. Blueshifted velocity of two Si II lines in the spectra ofSN 2001ay around maximum light. Within 1 σ the gradient derived from the two lines isthe same, about 200 km s − d − . This qualifies SN 2001ay to be a “high velocity gradient”object (Benetti et al. 2005). 52 – -20 0 20 40 60 80 100 120Rest-frame days since B-band maximum41.84242.242.442.642.84343.243.443.6 l og [ L ( e r g / s ec )] SN 2001aySN 2007ifSN 2009dc
Krisciunas et al.
Figure 17. Bolometric luminosities ( L ) of SNe 2001ay, 2007if (Scalzo et al.2010), and 2009dc (Taubenberger et al. 2011), measured in erg s − . For SN 2001ay we havescaled our integrated bolometric luminosity by a factor of 1.15 to account for flux notincluded in our photometric bandpasses. We adopted a distance modulus of m − M = 35 .
55 mag from Table 1. 53 –Krisciunas et al.
Figure 18. The best-fit radioactive decay energy deposition function(middle solid line) to the UVOIR light curve (blue squares) of SN 2001ay. Here we adoptArnett’s Rule (Arnett 1982), with α = 1.0, which stipulates that the gamma-ray depositionmatches the bolometric flux at maximum light. The top solid line corresponds to the caseof complete trapping of γ rays and positrons (i.e., τ ≫ γ -ray escape ( τ ≪ γ rays 57.2 ± -17-16-15 l og F λ + c on s t a n t SN 2005eq (-5)SN 2001ay (-1)SN 2009dc (-2)4000 5000 6000 7000 8000 9000Rest Wavelength (Angstroms)-17-16-15 l og F λ + c on s t a n t SN 2005eq (+5)SN 2001ay (+5)SN 2009dc (+5) C a II S i II S i II S i II S i III S II S II C II F e III F e III O I C a II C II Krisciunas et al.
Figure 19. Comparison of the optical spectra of SNe 2001ay, 2005eq, and2009dc. 55 –
Ca II O IFe IIIFe IICo II Mg II C IISi IICo II Si IISi II C II O IMg IICa IISi IIFe IIIFe II S IIS II
SYNOW FitSN 2001ay (t=−1d)
Krisciunas et al.
Figure 20. SYNOW modeling of the optical spectrum of SN 2001ay at 1 dbefore B maximum. Upper curve = data; lower curve = SYNOW model. The dashed linesindicate the addition of C II to the model. 56 – Si IIFe IISi II S IIS IINa ISi II O I Ca II O IMg II
SN 2001ay (t=+6d)
Co IICa II Co IISi II Mg IIFe IIIFe II
SYNOW fit
Fe III
Krisciunas et al.
Figure 21. SYNOW modeling of the optical spectrum of SN 2001ay 6 dafter B maximum. Upper curve = data; lower curve = SYNOW model. 57 – l og ( F λ ) + c on s t a n t SN 2001ay (t = −1 d)SYNAPPS fit Krisciunas et al.