A Cool and Inflated Progenitor Candidate for the Type Ib Supernova 2019yvr at 2.6 Years Before Explosion
Charles D. Kilpatrick, Maria R. Drout, Katie Auchettl, Georgios Dimitriadis, Ryan J. Foley, David O. Jones, Lindsay DeMarchi, K. Decker French, Christa Gall, Jens Hjorth, Wynn V. Jacobson-Galan, Raffaella Margutti, Anthony L. Piro, Enrico Ramirez-Ruiz, Armin Rest, Cesar Rojas-Bravo
MMNRAS , 1–22 (2020) Preprint Tuesday 12 th January, 2021 Compiled using MNRAS L A TEX style file v3.0
A Cool and Inflated Progenitor Candidate for the Type Ib Supernova2019yvr at 2.6 Years Before Explosion
Charles D. Kilpatrick ★ , Maria R. Drout , , Katie Auchettl , , , , Georgios Dimitriadis ,Ryan J. Foley , David O. Jones , Lindsay DeMarchi , K. Decker French , Christa Gall ,Jens Hjorth , Wynn V. Jacobson-Galán , Raffaella Margutti , Anthony L. Piro ,Enrico Ramirez-Ruiz , , Armin Rest , , César Rojas-Bravo Center for Interdisciplinary Exploration and Research in Astrophysics (CIERA) and Department of Physics and Astronomy,Northwestern University, Evanston, IL 60208, USA David A. Dunlap Department of Astronomy and Astrophysics, University of Toronto, 50 St. George Street, Toronto, Ontario, M5S 3H4 Canada The Observatories of the Carnegie Institution for Science, 813 Santa Barbara St., Pasadena, CA 91101, USA School of Physics, The University of Melbourne, Parkville, VIC 3010, Australia ARC Centre of Excellence for All Sky Astrophysics in 3 Dimensions (ASTRO 3D) Department of Astronomy and Astrophysics, University of California, Santa Cruz, CA 95064, USA DARK, Niels Bohr Institute, University of Copenhagen, Jagtvej 128, 2200 Copenhagen, Denmark Department of Astronomy, University of Illinois, 1002 West Green St., Urbana, IL 61801, USA Department of Physics and Astronomy, Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218, USA Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, USA
Accepted 0000, Received 0000, in original form 0000
ABSTRACT
We present
Hubble Space Telescope imaging of a pre-explosion counterpart to SN 2019yvr obtained 2.6 years before its explosionas a type Ib supernova (SN Ib). Aligning to a post-explosion Gemini-S/GSAOI image, we demonstrate that there is a single sourceconsistent with being the SN 2019yvr progenitor system, the second SN Ib progenitor candidate after iPTF13bvn. We also analyzedpre-explosion
Spitzer /IRAC imaging, but we do not detect any counterparts at the SN location. SN 2019yvr was highly reddened,and comparing its spectra and photometry to those of other, less extinguished SNe Ib we derive 𝐸 ( 𝐵 − 𝑉 ) = . + . − . mag forSN 2019yvr. Correcting photometry of the pre-explosion source for dust reddening, we determine that this source is consistentwith a log ( 𝐿 / 𝐿 (cid:12) ) = . ± . 𝑇 eff = + − K star. This relatively cool photospheric temperature implies a radius of320 + − 𝑅 (cid:12) , much larger than expectations for SN Ib progenitor stars with trace amounts of hydrogen but in agreement withpreviously identified SN IIb progenitor systems. The photometry of the system is also consistent with binary star models thatundergo common envelope evolution, leading to a primary star hydrogen envelope mass that is mostly depleted but still seeminglyin conflict with the SN Ib classification of SN 2019yvr. SN 2019yvr had signatures of strong circumstellar interaction in late-time( >
150 day) spectra and imaging, and so we consider eruptive mass loss and common envelope evolution scenarios that explainthe SN Ib spectroscopic class, pre-explosion counterpart, and dense circumstellar material. We also hypothesize that the apparentinflation could be caused by a quasi-photosphere formed in an extended, low-density envelope or circumstellar matter aroundthe primary star.
Key words: stars: evolution — supernovae: general — supernovae: individual (SN 2019yvr)
Core-collapse supernovae (SNe) are the terminal explosions of starswith initial mass > 𝑀 (cid:12) (Burrows et al. 1995). This aspect of mas-sive star evolution was empirically confirmed by the discovery of theblue supergiant progenitor of SN 1987A (Podsiadlowski 1993) andsubsequent discovery of over two dozen SN progenitors in nearbygalaxies (Smartt et al. 2015, and references therein, with more dis-covered since). The majority of these stars are red supergiant (RSG) ★ Email: [email protected] progenitors of hydrogen-rich type II SNe (SNe II), although sev-eral hydrogen-poor SN IIb progenitor stars, all of which are A–Ksupergiants, have also been explored in the literature (notably forSNe 1993J, 2008ax, 2011dh, 2013df, and 2016gkg; Aldering et al.1994; Crockett et al. 2008; Maund et al. 2011; Van Dyk et al. 2014;Kilpatrick et al. 2017).To date, there is only one confirmed example of a progenitorstar to a hydrogen-stripped SN Ib; the progenitor of iPTF13bvn inNGC 5608 was initially identified as a compact Wolf-Rayet (WR)star in pre-explosion
Hubble Space Telescope ( HST ) imaging (Caoet al. 2013) and confirmed as the progenitor by its disappearance © a r X i v : . [ a s t r o - ph . H E ] J a n Kilpatrick et al. (Eldridge & Maund 2016; Folatelli et al. 2016). There are numerousupper limits on the progenitor systems of other SNe Ib in the literature(Eldridge et al. 2013), implying they tend to have low optical lumi-nosities. Overall, this relative lack of pre-explosion detections hascontributed to an ongoing debate on the nature of stripped-envelopeSN progenitors.The transition from hydrogen-rich type II to hydrogen-poor typeIIb to hydrogen-free type Ib SNe, and finally to helium-free type IcSNe is commonly understood as a continuum in final hydrogen (orhelium) mass in the envelopes of their progenitor stars (Filippenko1997; Dessart et al. 2011, 2012; Yoon et al. 2012; Yoon 2015; Dessartet al. 2015; Maund & Ramirez-Ruiz 2016). Possible mechanisms thatcan deplete stellar envelope mass include radiative mass loss (Hegeret al. 2003; Crowther 2007; Smith 2014), eruptive mass loss (Langeret al. 1994; Maeder & Meynet 2000; Ramirez-Ruiz et al. 2005;Dessart et al. 2010), and mass transfer in binary systems (Woosleyet al. 1994; Izzard et al. 2004; Fryer et al. 2007; Yoon 2017). Starswith higher initial masses or metallicities are predicted to be morestripped at the time of core collapse due to their strong radiativewinds (Heger et al. 2003). However, extremely high mass stars thatare predicted to efficiently deplete their envelopes have more compactand thus less explodable cores, which is thought to lead to a significantfraction of failed SNe, that is, direct collapse to a black hole withno luminous transient (Burrows et al. 2007; Sukhbold et al. 2016;Murguia-Berthier et al. 2020). In addition, the large relative fractionof stripped-envelope SNe in volume-limited surveys (i.e., SNe Iband Ic; Li et al. 2011; Shivvers et al. 2017a; Graur et al. 2017a,b)suggests they come from a progenitor channel including stars withinitial masses < 𝑀 (cid:12) (Smith et al. 2011; Eldridge et al. 2013).Mass transfer in a binary system is therefore an appealing alter-native mechanism to strip massive star envelopes as the majority ofmassive stars are observed to evolve in binaries (Sana et al. 2012;Kiminki & Kobulnicky 2012), and binary interactions can lead to awide variety of outcomes based on mass ratio, orbital period, andthe characteristics of each stellar component (e.g., Wu et al. 2020).In particular, Case B (during helium core contraction; Kippenhahn& Weigert 1967) or Case BB (after core helium exhaustion for a starwith previous Case B mass transfer; Delgado & Thomas 1981) masstransfer can remove nearly all of a star’s hydrogen envelope, althoughthis process typically stops before hydrogen is completely depleted(Yoon et al. 2010; Yoon 2015, 2017). Stars with a small amountof hydrogen remaining might also swell up in the latest stages ofevolution (Divine 1965; Habets 1986; Götberg et al. 2017; Laplaceet al. 2020) and fill their Roche lobes to restart mass transfer. If themass transfer is non-conservative, that is some of the material is notaccreted by the companion star, this scenario can lead to dense cir-cumstellar material (CSM) in their local environments. Thus, whenthe primary star explodes the SN ejecta might encounter and shockthis material, producing strong thermal continuum and hydrogen andhelium line emission at optical wavelengths (i.e., SN IIn and Ibnfeatures; Vanbeveren et al. 1979; Claeys et al. 2011; Maund et al.2016; Yoon et al. 2017; Smith 2017; Götberg et al. 2019). Thus, thefinal envelope mass, radius, and composition of the star can result inSNe with diverse photometric and spectroscopic properties (James &Baron 2010) ranging from type II to type IIn to type Ic-like evolution.One prediction from this model of binary mass transfer is thatthere may be a continuum between SNe with type IIb and Ib-likebehaviour, depending on their final hydrogen mass. Dessart et al.(2012) find that progenitor stars with as little as 10 − 𝑀 (cid:12) hydrogenenvelope mass would produce a SN whose spectra exhibit broad H 𝛼 line emission up to 10 days after maximum light (although otherstudies find the envelope mass can be as large as 0.02–0.03 𝑀 (cid:12) with no H 𝛼 signature; Hachinger et al. 2012). Stars on either edge ofthis dividing line are expected to vary not only in the spectroscopicevolution of their resulting SN but also their appearance in pre-explosion imaging. Above this threshold, spectroscopic evolutionshould be similar to archetypal SNe IIb such as SN 1993J (Filippenkoet al. 1993; Richmond et al. 1994; Woosley et al. 1994), and theprogenitor star can inflate to radii >
400 R (cid:12) (Yoon 2017; Laplace et al.2020). Indeed, the progenitor of SN 1993J was a K-type supergiant(Nomoto et al. 1993; Aldering et al. 1994; Fox et al. 2014). Incontrast, stars with final hydrogen-envelope masses low enough thatthey would be classified as a type Ib SN prior to maximum light areonly expected to inflate to radii of at most ∼
100 R (cid:12) (Yoon et al. 2012;Yoon 2015, 2017; Kleiser et al. 2018; Laplace et al. 2020) and inmany cases remain significantly smaller. This should result in hotterprogenitor stars for a given luminosity.Intriguingly, some SNe Ib exhibit signatures of circumstellar in-teraction with hydrogen-rich gas weeks to months after explosion,which suggests their progenitor stars (or binary companions) recentlyreleased this material. The best-studied example to date is SN 2014C(Milisavljevic et al. 2015; Margutti et al. 2017; Tinyanont et al. 2016,2019), which was discovered in NGC 7331 at ≈
15 Mpc, but severalother stripped-envelope SNe with similar evolution have been pre-sented in the literature (e.g., SNe 2001em, 2003gk, 2004dk, 2018ijp,2019tsf, 2019oys; Chugai & Chevalier 2006; Bietenholz et al. 2014;Chandra 2018; Mauerhan et al. 2018; Pooley et al. 2019; Sollermanet al. 2020; Tartaglia et al. 2020) as well as the initially hydrogen-free superluminous SN iPTF13ehe (Yan et al. 2017). Although non-conservative mass transfer or common envelope ejections have beenproposed as the source of this material (Sun et al. 2020), it is still un-clear what evolutionary pathways lead to these apparently hydrogen-stripped stars or what exact mechanism causes an ejection timedonly years before explosion (up to 1 𝑀 (cid:12) of hydrogen-rich CSM forSN 2014C in Margutti et al. 2017).Understanding how common stripped-envelope SNe with circum-stellar interactions are might aid in ruling out less likely mechanisms,but constraining the exact rate is difficult as few SNe are close andbright enough to follow to late times and stripped-envelope SNe tendto be further extinguished in their host galaxies (Stritzinger et al.2018). Some SN Ib exhibit clear signatures of circumstellar inter-action with helium-rich material at early times (so-called SNe Ibn,with narrow emission lines of helium indicative of interaction be-tween SN ejecta and slow-moving, circumstellar helium; Pastorelloet al. 2008; Shivvers et al. 2017b), potentially from massive, helium-rich WR stars undergoing extreme mass loss immediately beforeexplosion (Smith et al. 2017). However, events from this class arerare and there with significant photometric and spectroscopic diver-sity (Hosseinzadeh et al. 2017). Margutti et al. (2017) analyzed 183SNe Ib and Ic with late-time radio observations and found that 10%exhibit evidence for rebrightening consistent with SN 2014C-likeevolution, implying this phenomenon may be relatively common.However, volume-limited samples with light curves beyond 100 daysof discovery (when most of these interactions occur; Sollerman et al.2020) are small (e.g., in Li et al. 2011; Shivvers et al. 2017a), and sothere may be an observational bias preventing precise constraints onthe intrinsic rate of these interactions in SNe Ib/c.In this paper we discuss a progenitor candidate for the SN Ib2019yvr discovered in NGC 4666 on UTC 2019 December 2712:30:14 (MJD 58844.521) by the Asteroid Terrestrial impact Last MNRAS , 1–22 (2020) rogenitor of SN 2019yvr Alert System (ATLAS; Smith et al. 2019) . We present early-timelight curves and spectra of SN 2019yvr demonstrating that it re-sembles several other SNe Ib and is spectroscopically most similarto iPTF13bvn, albeit with much more line-of-sight extinction thanmost known SNe Ib. Although we do not present any observationsbeyond 35 days from discovery, we note that SN 2019yvr exhibitedsignatures of circumstellar interaction at >
150 days from discovery,with evidence for relatively narrow H 𝛼 , X-ray, and radio emission atthese times (Auchettl et al. in prep.). From this information, we inferthat SN 2019yvr is similar to SN 2014C, with early-time type Ib-likeevolution but transitioning around 150 days to a light curve poweredby shock interaction with CSM at all wavelengths.NGC 4666 has deep Hubble Space Telescope /Wide Field Camera3 (
HST /WFC3) imaging in F438W, F555W, F625W, and F814Wbands (roughly
𝐵𝑉𝑅𝐼 , respectively) that covers the site of SN 2019yvr2.6 yr before its explosion (Shappee et al. 2016; Foley et al. 2016;Graur et al. 2018). Compared with limits on the progenitor starsof other SNe Ib in the literature (Eldridge et al. 2013) as well asthe detection of the progenitor star of iPTF13bvn (Cao et al. 2013),these data are among the deepest pre-explosion imaging for anySN Ib. We compare follow-up adaptive optics-fed imaging to the pre-explosion
HST images and identify a single progenitor candidate withF555W = . ± .
03 mag and F555W − F814W = . ± .
04 mag(AB mag; Table 2).Using constraints on the interstellar host extinction to SN 2019yvrinferred from photometry and spectra of the SN itself, we characterizethe progenitor candidate’s intrinsic spectral shape and find it is con-sistent with a star with log ( 𝐿 / 𝐿 (cid:12) ) = . ± . 𝑇 eff = + − K.This is much cooler than most SN Ib progenitor stars are generallythought to be (including the progenitor star of iPTF13bvn; Cao et al.2013; Bersten et al. 2014; Folatelli et al. 2016), implying that ifthe counterpart is dominated by the SN 2019yvr progenitor star itmust be significantly inflated compared with expectations for SN Ibprogenitor systems. Analyzing Binary Population and Spectral Syn-thesis (BPASS; Eldridge et al. 2017) stellar evolution models, wefind that the
HST photometry could be consistent with a 19 𝑀 (cid:12) pro-genitor star with a relatively low-mass ( ≈ 𝑀 (cid:12) ), close companionstar that undergoes common envelope evolution and sheds most ofits hydrogen envelope. However, all of these models predict a sig-nificant residual hydrogen envelope mass and are thus in conflictwith the observed type Ib spectral class of SN 2019yvr. Therefore,we hypothesize that the progenitor star may have shed its remain-ing hydrogen envelope through pre-SN eruptive mass ejection in thelast 2.6 yr before explosion. Otherwise, the apparently inflated radiusmay be caused by a much lower mass of hydrogen forming a compactquasi-photosphere in the progenitor star’s circumstellar environmentsoon before explosion.Throughout this paper, we assume a distance to NGC 4666 of 𝑚 − 𝑀 = . ± . . ± . 𝑧 = . 𝐸 ( 𝐵 − 𝑉 ) = .
02 mag(Schlafly & Finkbeiner 2011). SN 2019yvr is also called ATLAS19benc. We use SN 2019yvr throughoutthis paper for consistency with follow-up reports.
We analyzed
HST /WFC3 imaging of NGC 4666 obtained from theMikulski Archive for Space Telescopes . These data were observedover five epochs from 2017 April 21 to August 7 (Cycle 24, GO-14611, PI Graur; see Table 2), corresponding to 980 to 872 days(2.68 to 2.39 years) before discovery of SN 2019yvr. Using our anal-ysis code hst123 , we downloaded every HST image covering theexplosion site of SN 2019yvr. These comprised WFC3/UVIS flc frames calibrated with the latest reference files, including correc-tions for bias, dark current, flat-fielding, bad pixels, and geomet-ric distortion. We optimally aligned each image using
TweakReg with 1000–2000 sources per frame and resulting in frame-to-framealignment with 0.1–0.2 pix (0.005–0.010 (cid:48)(cid:48) ) root-mean-square dis-persion. We then drizzled all images in each band and epoch with astrodrizzle . With the drizzled F555W frame as a reference, weobtained photometry in the flc frames of every source on the samechip as the SN 2019yvr explosion site using dolphot (Dolphin2016). Our dolphot parameters followed the recommended set-tings for WFC3/UVIS as described in hst123 . We show a colourimage constructed from the F814W, F555W, and F438W framesobtained on 2017 June 13 in Figure 1.In addition, multiple epochs of Spitzer /Infrared Array Camera(IRAC) imaging of NGC 4666 were obtained from 2005 January 4to 2014 September 25, or roughly 15.0 to 5.3 yr before discoveryof SN 2019yvr. There was a single epoch of Channel 4 (7.9 𝜇 m)imaging that observed NGC 4666 (AOR 21999872; PI Rieke),but no Spitzer /IRAC observations cover NGC 4666 in Channel 3(5.7 𝜇 m). We downloaded the basic calibrated data ( cbcd ) framesand stacked them using our custom Spitzer /IRAC pipeline based onthe photpipe imaging and reduction pipeline (Rest et al. 2005; Kil-patrick et al. 2018a). The IRAC frames were stacked and regriddedto a pixel scale of 0.6 arcsec pixel − using SWarp (Bertin 2010).We performed photometry on the stacked frames using
DoPhot (Schechter et al. 1993) and calibrated our data with
Spitzer /IRACinstrumental response (for the cold and warm missions where appro-priate; Hora et al. 2012) in the stacked frames. Based on the PSF widthand average sky background, the average depth of the
Spitzer /IRACimages is approximately (3 𝜎 ; AB mag) 24.3 mag, 24.6 mag, and23.0 mag at 3.6, 4.5, and 7.9 𝜇 m, respectively. We observed SN 2019yvr in 𝐻 -band on 2020 March 8, or 72 daysafter discovery, with the Gemini-South telescope from Cerro Pachón,Chile and the Gemini South Adaptive Optics Imager (GSAOI; Mc-Gregor et al. 2004). We used the Gemini Multi-conjugate AdaptiveOptics System (GeMS; Rigaut et al. 2014) with the Gemini Southlaser guide star system to perform adaptive optics corrections overthe GSAOI field of view (85 (cid:48)(cid:48) × (cid:48)(cid:48) ) and using SN 2019yvr itselfto perform tip-tilt corrections. We alternated observations betweena field covering SN 2019yvr and a relatively empty patch of sky 4 (cid:48) to the south in an on-off-on-off pattern, totaling 1005 seconds of https://archive.stsci.edu/hst/ https://github.com/charliekilpatrick/hst123 http://americano.dolphinsim.com/dolphot/dolphotWFC3.pdf MNRAS , 1–22 (2020)
Kilpatrick et al. on-source exposure time over 39 frames. Using the GSAOI reduc-tion tools in
IRAF , we flattened the images with a flat-field frameconstructed from observations of a uniformly-illuminated screen inthe same filter and instrumental setup with unilluminated frames ofthe same exposure time to account for bias and dark current. We thensubtracted the sky frames from our on-source frames.GSAOI has a well-understood geometric distortion pattern (Ne-ichel et al. 2014). We used this distortion pattern to resample eachon-source frame to a corrected grid, aligned the individual exposures,and constructed a mosaic from each amplifier in the on-source frameswith the GSAOI tool disco-stu . Finally, we stacked the individ-ual frames with SWarp using an inverse-variance weighted medianalgorithm and scaling each image to the flux of isolated point sourcesobserved in every on-source exposure. The final stacked frame isshown in the upper-left inset of Figure 1 centered on SN 2019yvr.
We observed SN 2019yvr with the Swope 1.0-m telescope andDirect/4K ×
4K imager at Las Campanas Observatory, Chile from2020 January 1 to 28 in 𝑢𝐵𝑉𝑔𝑟𝑖 . Following reduction proceduresdescribed in Kilpatrick et al. (2018a), we performed all image pro-cessing and photometry on the Swope data using photpipe (Restet al. 2005). The final
𝐵𝑉𝑔𝑟𝑖 photometry of SN 2019yvr were cali-brated using PS1 standard sources (Flewelling et al. 2020) observedin the same field as SN 2019yvr and transformed into the Swopenatural system following the Supercal method (Scolnic et al. 2015).In 𝑢 -band, we calibrated our images using SkyMapper standards(Onken et al. 2020) in the same frame as SN 2019yvr.We also observed SN 2019yvr with the Las Cumbres Observatory(LCO) Global Telescope Network 1-m telescopes from 2019 Decem-ber 29 to 2020 February 3 with the Sinistro imagers and in 𝑔 (cid:48) 𝑟 (cid:48) 𝑖 (cid:48) . Weobtained the processed images (from the BANZAI pipeline; McCullyet al. 2018) from the LCO archive and processed them in photpipe ,registering each image to a corrected grid with
SWarp (Bertin 2010)and performing photometry on the individual frames with
DoPhot (Schechter et al. 1993). We then calibrated the 𝑔 (cid:48) 𝑟 (cid:48) 𝑖 (cid:48) photometryusing 𝑔𝑟𝑖 PS1 standards.All Swope and LCO photometry are listed in Table 6 and shownin Figure 2. We estimated the time of maximum light in 𝑉 -band byfitting a low-order polynomial to the overall light curve and derivea time of 𝑉 -band maximum light at MJD 58853.64 (2020 January5.64). Detailed modelling of the light curves and inferred explosionparameters will be presented by Auchettl et al. (in prep.). We triggered spectroscopic observations of SN 2019yvr on theFaulkes-North 2-m telescope at Haleakal¯a, Hawaii with the FLOYDSspectrograph (Program NOAO2020A-008, PI Kilpatrick). The spec-trum was observed on 2020 January 2 roughly 5 days after the ini-tial discovery report from ATLAS and 3 days before SN 2019yvrreached 𝑉 -band maximum. The observation was a 1500-s exposureat an average airmass of 1.35 and under near-photometric observing IRAF is distributed by the National Optical Astronomy Observatory, whichis operated by the Association of Universities for Research in Astronomy(AURA) under a cooperative agreement with the National Science Founda-tion. conditions. We reduced the spectrum following standard proceduresin IRAF , including corrections for telluric absorption and correctingthe wavelength solution for atmospheric diffraction using the skylines. The final reduced spectrum is shown in Figure 3.We also observed SN 2019yvr on the Keck-I 10-m telescope onMaunakea, Hawaii with the Low-Resolution Imaging Spectrograph(LRIS; Program 2019B-U169, PI Foley) on 27 Jan 2020, approxi-mately 22 days after 𝑉 -band maximum as seen from our light curve.The observation was a 180-s exposure obtained during morning twi-light at an average airmass of 1.16 and under near-photometric con-ditions. We reduced these data using a custom pyraf -based LRISpipeline (Siebert et al. 2020) , which accounts for bias-subtraction,flat-fielding, amplifier crosstalk, background and sky subtraction, tel-luric corrections using a standard observed on the same night andat a similar airmass, and order combination. The final combinedspectrum is shown in Figure 3.Our spectra reveal characteristic SN Ib features with strong, broadabsorption lines of He i 𝜆𝜆 − . We also note prominent linesof Na i D absorption at the redshift of NGC 4666 ( 𝑧 = . Stripped-envelope SNe Ib are known to occur in regions of highextinction in their host galaxies (Drout et al. 2011; Galbany et al.2016a,b; Stritzinger et al. 2018). However, if there is significant ex-tinction due to dust in the circumstellar environment of SN 2019yvr,it may be variable between the time the
HST images and imaging andspectra of SN 2019yvr were obtained. Moreover, we have no a prioriconstraint on the dust composition or gas-to-dust ratio in the localinterstellar environment of SN 2019yvr, which is a major factor inunderstanding the magnitude of extinction at all optical wavelengths.Based on the relatively low Milky Way reddening of 𝐸 ( 𝐵 − 𝑉 ) = .
02 mag and the fact that SN 2019yvr exhibited red colours (Fig-ure 2) and strong Na i D absorption, we infer that SN 2019yvr and itsprogenitor system are heavily extinguished by its host’s interstellarand/or its own circumstellar environment. Moreover, if we do notcorrect for any additional extinction, the 𝑉 -band light curve wouldpeak at only − . 𝐴 𝑉 > 𝑉 -band extinction 𝐴 𝑉 and reddening law https://github.com/msiebert1/UCSC_spectral_pipeline MNRAS000
02 mag and the fact that SN 2019yvr exhibited red colours (Fig-ure 2) and strong Na i D absorption, we infer that SN 2019yvr and itsprogenitor system are heavily extinguished by its host’s interstellarand/or its own circumstellar environment. Moreover, if we do notcorrect for any additional extinction, the 𝑉 -band light curve wouldpeak at only − . 𝐴 𝑉 > 𝑉 -band extinction 𝐴 𝑉 and reddening law https://github.com/msiebert1/UCSC_spectral_pipeline MNRAS000 , 1–22 (2020) rogenitor of SN 2019yvr Figure 1. ( Right ) Hubble Space Telescope imaging of the SN 2019yvr explosion site from 2.5 years before discovery consisting of F814W (red), F555W (green),and F438W (blue). All images are oriented with north up and east to the left. The color image on the right is 165 (cid:48)(cid:48) × (cid:48)(cid:48) , while the left-upper and left-middleimages are 38.8 (cid:48)(cid:48) × (cid:48)(cid:48) , and the left-lower image is 2.4 (cid:48)(cid:48) × (cid:48)(cid:48) . The blue box denotes the approximate location of SN 2019yvr. ( Upper left ): Gemini-S/GSAOI 𝐻 -band image of SN 2019yvr obtained 67 days after discovery of the transient. The image is centered on the location of SN 2019yvr. ( Middle left ): Pre-explosionF555W imaging of NGC 4666 showing the same location as the upper left. (
Lower left ): Pre-explosion F555W imaging zoomed into the blue box from themiddle left. The location of the SN 2019yvr progenitor candidate derived from our Gemini-S/GSAOI imaging is shown as red lines, which agrees with thelocation of a single point source as discussed in Section 4.1. parameter 𝑅 𝑉 that can be used to estimate the total extinction in the HST bandpasses as observed in pre-explosion data.
One quantity that is correlated with line-of-sight reddening in bothSNe (Stritzinger et al. 2018) and quasars (Poznanski et al. 2012) isthe equivalent width of Na i D. We detect Na i D in our 2019yvr LRISspectrum with equivalent width of 4.2 ± 𝐸 ( 𝐵 − 𝑉 ) > MNRAS , 1–22 (2020)
Kilpatrick et al. − − A pp a r e n t B r i gh t n e ss ( A B m a g ) irVgBu-1.3 Figure 2.
Swope (circle) and LCO (square) 𝑢𝐵𝑔𝑉 𝑟𝑖 light curves ofSN 2019yvr as described in Section 2.3. We denote the epoch of each ob-servation in rest-frame days (correcting for the redshift of NGC 4666 at 𝑧 = . 𝐵 -band maximum light detected at MJD 58854.28(Table 6). equivalent width regime where we find SN 2019yvr (consistent withfindings in Stritzinger et al. 2018).If we instead use the relation between 𝐴 𝑉 and Na i D equivalentwidth in Stritzinger et al. (2018), which was derived specificallyfrom SN Ib/c colour curves, we find SN 2019yvr has a line-of-sightextinction 𝐴 𝑉 = . ± . Spectra and light curves of SNe Ib similar to SN 2019yvr can be usedto constrain its line-of-sight extinction. As host extinction is a domi-nant systematic uncertainty in estimating intrinsic stripped-envelopeSN colours, any differences in broadband colours between SNe atsimilar epochs can be attributed to extinction. Here we compare ourSN 2019yvr spectra to those of other SNe Ib applying a Cardelli et al.(1989) extinction law with variable 𝐸 ( 𝐵 − 𝑉 ) and 𝑅 𝑉 to dereddenour SN 2019yvr until they closely match.Our template spectra are chosen from those of well-observedSNe Ib with low host reddening ( 𝐸 ( 𝐵 − 𝑉 ) < . 𝐸 ( 𝐵 − 𝑉 ) = .
05 mag; Deng et al. 2000; Benetti et al. 2011),SN 2007Y ( 𝐸 ( 𝐵 − 𝑉 ) = .
11 mag; Stritzinger et al. 2009), SN 2009jf( 𝐸 ( 𝐵 − 𝑉 ) ≈ . 𝐸 ( 𝐵 − 𝑉 ) = .
17 mag; Srivastav et al. 2014). We use spectra ob-tained from the Open Supernova Catalog (Guillochon et al. 2017).All template spectra were chosen to correspond to roughly the sameepoch relative to 𝑉 -band maximum as one of our two SN 2019yvr sne.space f λ ( s ca l e d + o ff s e t ) iPTF13bvn -2.1d A V =0.53 R V =3.12019yvr -3d A V =3.33 R V =4.912007Y -0.9d A V =0.35 R V =3.12019yvr -3d A V =3.14 R V =3.222009jf -3.3d A V =0.00 R V =3.12019yvr -3d A V =2.46 R V =3.011999dn -3.0d A V =0.15 R V =3.12019yvr -3d A V =3.63 R V =4.902009jf 25.6d A V =0.00 R V =3.12019yvr 22d A V =3.94 R V =4.97 He Figure 3.
Our spectra of SN 2019yvr (black) with comparison to other SNe Ib(red). All dates are indicated with a “d” with respect to 𝑉 -band maximumlight. The comparison spectra have been dereddened for Milky Way extinctionbased on values in Schlafly & Finkbeiner (2011) and dereddened for hostextinction based on values in Deng et al. (2000); Benetti et al. (2011) (forSN 1999dn), Stritzinger et al. (2009) (for SN 2007Y), Valenti et al. (2011)(for SN 2009jf), and Srivastav et al. (2014) (for iPTF13bvn). We removedthe recessional velocity for 𝑧 = . 𝑅 𝑉 parameter are given next to each SN 2019yvr spectrum.We highlight lines of He i at 𝜆𝜆 spectra. For the purposes of our fitting procedure, we assume thatthese extinction values are exact with no additional uncertainty. Fur-thermore, we assume all template spectra experienced Milky Way-like host reddening with 𝑅 𝑉 = . 𝑅 𝑉 is either unconstrained orpoorly constrained for all of these objects. We acknowledge that thispossibly biases our 𝐴 𝑉 and 𝑅 𝑉 estimates for SN 2019yvr based onthe spectroscopic fitting method, although 𝐸 ( 𝐵 − 𝑉 ) is small for ourtemplates and so this may not be a major systematic uncertainty. Forboth the SN 2019yvr and template spectra, we estimate the uncer-tainty in the specific flux ( 𝜎 𝜆 ) by taking 𝜎 𝜆 = ˜ 𝑓 𝜆 √︄(cid:12)(cid:12)(cid:12)(cid:12) − 𝑓 𝜆 ˜ 𝑓 𝜆 (cid:12)(cid:12)(cid:12)(cid:12) , (1) MNRAS000
Our spectra of SN 2019yvr (black) with comparison to other SNe Ib(red). All dates are indicated with a “d” with respect to 𝑉 -band maximumlight. The comparison spectra have been dereddened for Milky Way extinctionbased on values in Schlafly & Finkbeiner (2011) and dereddened for hostextinction based on values in Deng et al. (2000); Benetti et al. (2011) (forSN 1999dn), Stritzinger et al. (2009) (for SN 2007Y), Valenti et al. (2011)(for SN 2009jf), and Srivastav et al. (2014) (for iPTF13bvn). We removedthe recessional velocity for 𝑧 = . 𝑅 𝑉 parameter are given next to each SN 2019yvr spectrum.We highlight lines of He i at 𝜆𝜆 spectra. For the purposes of our fitting procedure, we assume thatthese extinction values are exact with no additional uncertainty. Fur-thermore, we assume all template spectra experienced Milky Way-like host reddening with 𝑅 𝑉 = . 𝑅 𝑉 is either unconstrained orpoorly constrained for all of these objects. We acknowledge that thispossibly biases our 𝐴 𝑉 and 𝑅 𝑉 estimates for SN 2019yvr based onthe spectroscopic fitting method, although 𝐸 ( 𝐵 − 𝑉 ) is small for ourtemplates and so this may not be a major systematic uncertainty. Forboth the SN 2019yvr and template spectra, we estimate the uncer-tainty in the specific flux ( 𝜎 𝜆 ) by taking 𝜎 𝜆 = ˜ 𝑓 𝜆 √︄(cid:12)(cid:12)(cid:12)(cid:12) − 𝑓 𝜆 ˜ 𝑓 𝜆 (cid:12)(cid:12)(cid:12)(cid:12) , (1) MNRAS000 , 1–22 (2020) rogenitor of SN 2019yvr Epoch Template (Epoch) 𝐴 𝑉 ,𝑇 𝑒𝑚𝑝. 𝐴 𝑉 , 𝑅 𝑉 , 𝜒 (days) (days) (mag) (mag) − − ± ± − − ± ± − − ± ± − − ± ± +
22 SN 2009jf ( + ± ± ± . ± Table 1.
Our best-fitting parameters for 𝑉 -band extinction ( 𝐴 𝑉 ) and 𝑅 𝑉 inferred for SN 2019yvr based on matching to template spectra as shownin Figure 3 and described in Section 3.2. As in Figure 3, the epoch of eachSN 2019yvr and template spectrum is given in days with respect to 𝑉 -bandmaximum light. 𝜒 is given in units of reduced 𝜒 / 𝜒 . We give parametersfor each template spectrum used and the inverse 𝜒 -weighted average for 𝐴 𝑉 and 𝑅 𝑉 . However, see caveats in Section 3.2. where ˜ 𝑓 𝜆 is the specific flux 𝑓 𝜆 passed through a smoothing functionwith a 50 Å window and rebinned to 1 Å resolution over the maxi-mum overlap range between the SN 2019yvr and template spectrum.Thus, the SN 2019yvr and template spectrum flux uncertainties arepropagated through our entire analysis. We then fit our LRIS andFLOYDS spectra of SN 2019yvr to the templates by calculating adereddened spectral template ˜ 𝑓 𝜆,𝑑 assuming the appropriate MilkyWay extinction and the interstellar host extinction given above. Wealso deredden SN 2019yvr for Milky Way extinction following thesame procedure yielding 𝑓 𝜆, . For both SN 2019yvr and thetemplate, we rescale the uncertainty 𝜎 𝜆 by the same factor as thedereddened spectrum. Finally, we derive the best-fitting host extinc-tion 𝐴 𝑉 , and reddening law parameter 𝑅 𝑉 , by calculating 𝐴 𝜆 from Cardelli et al. (1989) and minimizing the reduced 𝜒 value 𝜒 = 𝑁 ∑︁ 𝜆 ( ˜ 𝑓 𝜆,𝑑 − 𝐶 𝑓 𝜆, . 𝐴 𝜆 ) 𝑁 ( 𝜎 𝜆 + 𝜎 𝜆, ) (2)where 𝑁 is the total number of 1 Å wavelength bins and 𝐶 is a scalingconstant between the two spectra. Thus, our spectral fitting method isprimarily sensitive to the overall shape of the two spectra rather thanthe ratio between their fluxes. We show our best-fitting dereddenedSN 2019yvr spectra in Figure 3 and we list our best-fitting 𝐴 𝑉 and 𝑅 𝑉 parameters for SN 2019yvr in Table 1.As reported in Section 2.4, our SN 2019yvr spectra correspond toapproximately − +
22 days relative to 𝑉 -band maximum.For the latter spectrum, only SN 2009jf had a spectrum sufficientlyclose in 𝑉 -band epoch to perform a robust comparison betweenspectral shape. Thus, while the best-fitting cases all correspond to theearly-time spectrum, our second epoch serves to validate the resultsof this analysis. In this way, we derive a line-of-sight extinction toSN 2019yvr of 𝐴 𝑉 = . . 𝐴 𝑉 = . . 𝐴 𝑉 , implying that thereare systematic uncertainties in our method. We further investigate the interstellar host reddening using colourcurve templates from Stritzinger et al. (2018) and compare to ourSwope and LCO colour curves of SN 2019yvr (Figure 4). Using aCardelli et al. (1989) reddening law, we vary the values of 𝐴 𝑉 and 𝑅 𝑉 in order to derive colour corrections due to interstellar reddening.We apply these corrections to our 𝐵 − 𝑔 , 𝐵 − 𝑉 , 𝑔 − 𝑟 , and 𝑔 − 𝑖 colour −
10 0 10 20Rest-frame days from g max . . . . . . . . B - g ( A B m a g ) −
10 0 10 20Rest-frame days from B max . . . . . . B - V ( A B m a g ) −
10 0 10 20Rest-frame days from g max . . . . . . . g - r ( A B m a g ) −
10 0 10 20Rest-frame days from r max − . − . . . . . r - i ( A B m a g ) Figure 4.
Colour curves of SN 2019yvr corrected for Milky Way and inter-stellar host extinction (with 𝐴 𝑉 = . 𝑅 𝑉 = .
4) as discussed inSection 3.3. Circles correspond to our Swope photometry while squares arefor LCO photometry. We overplot templates for extinction-corrected SN Ibcolour curves from Stritzinger et al. (2018) as black lines with the 1 𝜎 uncer-tainties in each template as a gray shaded region. curves to find the best fit with Stritzinger et al. (2018) templatecolours as shown in Figure 4. The best-fitting values are quantifiedwith respect to the summed 𝜒 values in 𝐵 − 𝑔 , 𝐵 − 𝑉 , 𝑔 − 𝑟 , and 𝑟 − 𝑖 and across all epochs. The final reduced 𝜒 value for differentvalues of 𝐴 𝑉 and 𝑅 𝑉 is shown in Figure 5.We find the best-fitting values using our colour curve matchingare 𝐴 𝑉 = . + . − . mag and 𝑅 𝑉 = . + . − . and implying a best-fitting 𝐸 ( 𝐵 − 𝑉 ) = . + . − . mag. The value of 𝑅 𝑉 is limited at thehigh end by our boundary condition that 𝑅 𝑉 < .
0. Based on well-measured values of 𝑅 𝑉 for SN host galaxies (e.g., the wide variety ofSN Ia hosts presented in Amanullah et al. 2015), which tend to have1 . < 𝑅 𝑉 < .
6, we infer that our prior 𝑅 𝑉 < . Overall, the level of extinction inferred from our spectral analysis isconsistent with our estimate from the 𝑉 -band light curve as well asthe value inferred from the Stritzinger et al. (2018) Na i D relation.While the latter relationship diverges significantly at large extinctionvalues (also similar to Poznanski et al. 2012), we infer from theagreement between these three estimates that the line-of-sight hostextinction inferred for SN 2019yvr is close to the value inferred fromour spectral and colour curve analyses. However, the colour curveanalysis involves more independent measurements of the SN 2019yvroptical spectrum, and this analysis has been validated for severalSNe Ib by Stritzinger et al. (2018). Thus, although there is agreementbetween all of our methods, we infer that 𝐴 𝑉 = . + . − . mag and 𝑅 𝑉 = . + . − . is most representative of the line-of-sight extinctionto SN 2019yvr, and we use these values and our 𝜒 distribution onextinction in Figure 5 below.However, the critical question to the analysis below is how much MNRAS , 1–22 (2020)
Kilpatrick et al. R V . . . . . . . A V ( m a g ) E ( B − V ) = . m a g E ( B − V ) = . m a g . . . . . . . . . χ Figure 5. 𝜒 values as a function of assumed interstellar 𝑉 -band extinction 𝐴 𝑉 and reddening law parameter 𝑅 𝑉 and comparing SN 2019yvr colours tothe colour templates in Section 3.3 and Figure 4. The best-fitting extinctionparameters are 𝐴 𝑉 = . 𝑅 𝑉 = . 𝐸 ( 𝐵 − 𝑉 ) = .
51 mag. The yellow dashed lines show the 1- 𝜎 best-fitting limits of 𝐸 ( 𝐵 − 𝑉 ) . extinction did the SN 2019yvr progenitor star experience? While itis reasonable to assume that the interstellar host extinction inferredfrom SN 2019yvr would be the same as the extinction that its pro-genitor star experienced (especially on the 2.4–15.0 yr timescale ofour pre-explosion data), there could be additional sources of extinc-tion present when the pre-explosion data were obtained to which ourSN 2019yvr observations are not sensitive or vice versa. In particu-lar, circumstellar dust could have been present in the pre-explosionenvironment but vaporized soon after explosion, or else there couldbe material ejected by the progenitor star very soon before explosionthat was not present when the HST or Spitzer data were obtained.Below we consider both scenarios and the effects of circumstellarmaterial and extinction on our overall data set.
All massive stars exhibit winds that pollute their environments withgas and dust (Smith 2014), and this material can lead to significantcircumstellar extinction when the wind is dense, clumpy, and rela-tively cool. Thus it is possible that the SN 2019yvr progenitor starexperienced significant circumstellar extinction from a shell of dustthat was vaporized before it could be observed in the SN. Auchettlet al. (in prep.) find evidence for a significant mass of hydrogen-richCSM from H 𝛼 , X-ray, and radio emission. Rebrightening in the lightcurve of SN 2019yvr beginning >
150 days after discovery suggeststhat this material is in a shell likely at > HST observations. There is no obvious sign of any such material, for example, in evolution of the Na i D profileor excess emission in early-time light curves and spectra.Dust geometries and properties most likely to be associated withcircumstellar extinction due to material close (2–10 × the photo-spheric radius as in Kochanek et al. 2012) to the progenitor star butunconstrained by our SN 2019yvr observations can be probed withour mid-infrared Spitzer /IRAC limits. Assuming this material waspresent on the timescale of the IRAC observations, we model anoptically thin shell of dust to our limits of 22.8, 23.1, and 21.5 magin IRAC bands 1, 2, and 4, respectively (see Section 4.1 for a discus-sion of the IRAC limits). A warm shell of gas and dust (
𝑇 >
200 K)would result in bright mid-infrared emission even in cases whereit is relatively compact ( < > 𝜇 m and a range of temperatures from 200–1500 K. Athotter temperatures, the dust would likely sublimate and thus wouldnot exhibit the same extinction properties or attendant mid-infraredemission. Similarly, a shell at large distances from its progenitor starmight be so cool that it does not emit significant flux at < 𝜇 mwhere our IRAC data probe, even if it has a large mass.The dust mass limits we derive are strongly temperature dependent,with the coolest temperatures yielding the weakest limits on mass( 𝑀 𝑑 < × − 𝑀 (cid:12) and 𝐿 𝑑 < × 𝐿 (cid:12) at 200 K) whereas hotterdust leads to relatively strong limits on dust mass ( 𝑀 𝑑 < × − 𝑀 (cid:12) and 𝐿 𝑑 < × 𝐿 (cid:12) at 1500 K). We used the 0 . 𝜇 m silicate dustgrain opacities from Fox et al. (2010, 2011) to calculate these limits.Assuming the same dust grain composition, we approximate thelimits on optical depth in 𝑉 -band as 𝜏 𝑉 = 𝜌𝜅 𝑉 𝑟 dust , where 𝑟 dust isthe implied blackbody radius of the dust shell, 𝜌 ≈ 𝑀 𝑑 /( / 𝜋𝑟 ) ,and 𝜅 𝑉 is the opacity in 𝑉 -band. Under these assumptions, the opticaldepth must be 𝜏 𝑉 < 𝐴 𝑉 = . 𝜏 𝑉 as in Kochanek et al. (2012) andKilpatrick & Foley (2018), these limits are not constraining on thetotal circumstellar extinction due to a compact dust shell. Indeed, cir-cumstellar dust absorption could be the dominant source of extinctionin the SN 2019yvr progenitor system, but we would have no contex-tual information from the pre-explosion Spitzer /IRAC photometry toconstrain the magnitude of that extinction.The strongest argument against such a compact, warm shell ofgas and dust is the lack of any hydrogen or helium emission as-sociated with circumstellar interaction in early-time spectra or anynear-infrared excess in the photometry as shown in Figure 3 andFigure 2. However, these arguments are biased by the epoch of thefirst observations. SN 2019yvr had a reported discovery on 2019 De-cember 27 by ATLAS with the last previous non-detection occurringon 2019 December 11 at > . 𝑜 -band (Smith et al. 2019).Subsequent non-detection reports by the Zwicky Transient Facilitygive a more constraining non-detection in 𝑔 -band at > . However, this still allows for 14 days whenSN 2019yvr could have interacted with CSM in its immediate envi-ronment. Although the first spectrum of SN 2019yvr did not exhibitevidence for flash ionization or narrow emission lines due to CSMinteraction, this would not be surprising if the explosion was alreadymore than several days old (e.g., flash ionization lasted for < MNRAS000
200 K)would result in bright mid-infrared emission even in cases whereit is relatively compact ( < > 𝜇 m and a range of temperatures from 200–1500 K. Athotter temperatures, the dust would likely sublimate and thus wouldnot exhibit the same extinction properties or attendant mid-infraredemission. Similarly, a shell at large distances from its progenitor starmight be so cool that it does not emit significant flux at < 𝜇 mwhere our IRAC data probe, even if it has a large mass.The dust mass limits we derive are strongly temperature dependent,with the coolest temperatures yielding the weakest limits on mass( 𝑀 𝑑 < × − 𝑀 (cid:12) and 𝐿 𝑑 < × 𝐿 (cid:12) at 200 K) whereas hotterdust leads to relatively strong limits on dust mass ( 𝑀 𝑑 < × − 𝑀 (cid:12) and 𝐿 𝑑 < × 𝐿 (cid:12) at 1500 K). We used the 0 . 𝜇 m silicate dustgrain opacities from Fox et al. (2010, 2011) to calculate these limits.Assuming the same dust grain composition, we approximate thelimits on optical depth in 𝑉 -band as 𝜏 𝑉 = 𝜌𝜅 𝑉 𝑟 dust , where 𝑟 dust isthe implied blackbody radius of the dust shell, 𝜌 ≈ 𝑀 𝑑 /( / 𝜋𝑟 ) ,and 𝜅 𝑉 is the opacity in 𝑉 -band. Under these assumptions, the opticaldepth must be 𝜏 𝑉 < 𝐴 𝑉 = . 𝜏 𝑉 as in Kochanek et al. (2012) andKilpatrick & Foley (2018), these limits are not constraining on thetotal circumstellar extinction due to a compact dust shell. Indeed, cir-cumstellar dust absorption could be the dominant source of extinctionin the SN 2019yvr progenitor system, but we would have no contex-tual information from the pre-explosion Spitzer /IRAC photometry toconstrain the magnitude of that extinction.The strongest argument against such a compact, warm shell ofgas and dust is the lack of any hydrogen or helium emission as-sociated with circumstellar interaction in early-time spectra or anynear-infrared excess in the photometry as shown in Figure 3 andFigure 2. However, these arguments are biased by the epoch of thefirst observations. SN 2019yvr had a reported discovery on 2019 De-cember 27 by ATLAS with the last previous non-detection occurringon 2019 December 11 at > . 𝑜 -band (Smith et al. 2019).Subsequent non-detection reports by the Zwicky Transient Facilitygive a more constraining non-detection in 𝑔 -band at > . However, this still allows for 14 days whenSN 2019yvr could have interacted with CSM in its immediate envi-ronment. Although the first spectrum of SN 2019yvr did not exhibitevidence for flash ionization or narrow emission lines due to CSMinteraction, this would not be surprising if the explosion was alreadymore than several days old (e.g., flash ionization lasted for < MNRAS000 , 1–22 (2020) rogenitor of SN 2019yvr been needed to provide meaningful constraints on the presence andtotal mass of such material. There is strong evidence for circumstellar interaction aroundSN 2019yvr in optical spectra, radio, and X-ray detections start-ing around 150 days after discovery (Auchettl et al. in prep.). Thedevelopment of narrow Balmer lines at these late times indicatesthis material is hydrogen rich. A delayed interaction points to ashell of material at a large projected separation from the progenitor( ≈ ≈ − ).A key consideration above is whether this CSM was present at thetime of the HST observations or if it was ejected in the subsequent2.6 yr before core collapse. In the latter case, any dust synthesized inthe CSM would not be present in the
HST data and thus 𝐴 𝑉 = . HST observations and follow up data of the SN,we cannot constrain this scenario. However, one prediction fromthis scenario would be an intermediate-luminosity transient associ-ated with an extreme mass loss episode over this time. We analyzedthis location of the sky and found no luminous counterparts in pre-explosion imaging from the ASAS-SN Sky Patrol (Shappee et al.2014; Kochanek et al. 2017) or the Catalina Surveys Data Release 2(Drake et al. 2009), but these limits only extend to <
13 mag givencontamination from the bright center of NGC 4666. Thus we cannotprovide a meaningful estimate on any such CSM, but we considerthe possibility that 𝐴 𝑉 < . HST observations and the time of explosion. Overall, weconsider 𝐴 𝑉 = . + . − . mag and 𝑅 𝑉 = . + . − . to represent the totalline-of-sight extinction to the SN 2019yvr progenitor system at thetime the pre-explosion HST imaging was obtained.
We obtained positions for 114 point sources in our GSAOI adap-tive optics image using sextractor (Bertin & Arnouts 1996)and compared these to the positions of the same sources in theF555W
HST /WFC3 image as obtained in dolphot . From thesecommon sources, we derived a coordinate transformation solutionfrom GSAOI → HST . We also derived the systematic uncertainty inthis transformation by splitting our sample of common astrometricsources in half, re-deriving the coordinate transformation, and thencomparing the offset between the remaining
HST sources and theirpositions from our GSAOI image and transformation. Repeating thisprocedure, we are able to derive an average systematic offset betweenour GSAOI and
HST sources. We assume the root-mean-square ofthese offsets dominates the error in our astrometric solution, whichwe find is 𝜎 𝛼 = .
16 WFC3/UVIS pixels (0 . (cid:48)(cid:48) ) and 𝜎 𝛿 = . . (cid:48)(cid:48) ).The position of SN 2019yvr in our GSAOI image corresponds to asingle point source in the WFC3/UVIS imaging to a precision of 0 . ≈ 𝜎 as the uncertainty on the position of this https://asas-sn.osu.edu/ source from our GSAOI is negligible). We detect this source in thedrizzled F555W image at 34 𝜎 significance, and there are no othersources at the > 𝜎 level within a separation of 0.27 (cid:48)(cid:48) or 30 timesthe astrometric uncertainty. Our WFC2/UVIS photometry is listed inTable 2.We also examined the position of SN 2019yvr in pre-explosion Spitzer /IRAC imaging. Using the same alignment method as above,we determined the location of SN 2019yvr in the
Spitzer /IRACstacked images using our GSAOI image of SN 2019yvr. Our align-ment uncertainty is typically 𝜎 ≈ . . (cid:48)(cid:48) ) fromGSAOI → IRAC in each channel. We found no evidence of a counter-part in any epoch or the cumulative, stacked pre-explosion frames.Therefore, we place an upper limit on the presence of a pre-explosioncounterpart in the stacked IRAC frames by injecting and recoveringartificial stars at the location of SN 2019yvr and using the nativeIRAC point response function for each channel. Our pre-explosionlimits for IRAC are reported in Table 3.
HST
Counterpart to SN 2019yvr
Stripped-envelope SNe are known to occur in the brightest, high-est extinction, and highest metallicity regions of their host galaxies(Galbany et al. 2016a,b). The iPTF13bvn progenitor system wasidentified in a relatively uncrowded region of NGC 5086 (Cao et al.2013) and subsequently confirmed as the actual progenitor by itsdisappearance (Eldridge & Maund 2016; Folatelli et al. 2016), but ingeneral SNe Ib/c are found in crowded regions of their host galaxies(when the surrounding environment can be resolved, as in Eldridgeet al. 2013). For example, the candidate progenitor system of thestripped-envelope SN Ic 2017ein was in an environment with severalother luminous sources (Kilpatrick et al. 2018b). This fact and thecounterpart’s high optical luminosity suggest it may in fact have anunresolved star cluster or a chance coincidence.The candidate progenitor system to SN 2019yvr does not appearextended in any of the WFC3/UVIS frames, with dolphot average sharpness = − . roundness =0 .
36, and classified as a brightstar, which is consistent with a circular point source at WFC3/UVISresolution. The source is not blended with any other nearby sourcesand has an average crowding =0 .
09. Therefore, we conclude thatthe candidate counterpart is consistent with being a single, isolatedpoint source in all of our images.One possible scenario is that the candidate source is dominatedby emission from multiple stars in a single system or open cluster(similar to those in Bastian et al. 2005; Gieles et al. 2006; Gieles &Portegies Zwart 2011). Although the candidate source is point-like,the PSF size of
HST /WFC3 in F555W is ≈ (cid:48)(cid:48) , or 4.7 pc at thedistance of NGC 4666. Many open clusters are smaller than this,and might be so compact as to resemble a point source. The F555W(roughly 𝑉 -band) absolute magnitude we infer for this source is − 𝐴 𝑉 = . > 𝜎 in a 10 (cid:48)(cid:48) region surrounding the candidate SN 2019yvr counterpartin any of the HST frames. Thus, at most 6 . or 2% of thisregion is subtended by area within 3 𝜎 (astrometric uncertainty) ofany source, which is a conservative upper limit on the probabilityof chance coincidence between the counterpart and SN 2019yvr. We MNRAS , 1–22 (2020) Kilpatrick et al.
WFC3/UVIS Photometry of 2019yvr Progenitor CandidateMJD Filter Exposure (s) Magnitude Uncertainty57864.06972 F438W 1140 26.2028 0.220757864.11750 F625W 1134 24.8352 0.046657864.17879 F555W 1200 25.4011 0.072057864.24510 F814W 1152 24.1778 0.049857890.23024 F555W 1143 25.1599 0.061257890.24828 F625W 1140 24.8008 0.044357917.36677 F438W 1140 26.3646 0.481057917.39478 F625W 1134 24.9420 0.053857917.43295 F555W 1200 25.5264 0.085357917.46140 F814W 1152 24.3412 0.058857944.68250 F555W 1143 25.2253 0.062157944.75022 F625W 1140 24.8760 0.047857972.22031 F438W 1140 25.9439 0.241057972.23837 F625W 1134 25.0531 0.055957972.28531 F555W 1200 25.4908 0.082957972.30377 F814W 1152 24.2492 0.0575Average PhotometryMJD Filter Exposure (s) Magnitude Uncertainty57917.88560 F438W 3420 26.1382 0.162257904.13112 F555W 5886 25.3512 0.031957917.74983 F625W 5682 24.8971 0.022157918.00342 F814W 3456 24.2533 0.0319
Table 2.
HST
WFC3/UVIS photometry of the SN 2019yvr progenitor candi-date. All magnitudes are on the AB system.
Spitzer /IRAC Pre-explosion LimitsAverage MJD Wavelength Exposure Limit( 𝜇 m) (s) (mag)57917.88560 3.6 1530.0 > > > Table 3.
IRAC limits on the presence of a pre-explosion counterpart toSN 2019yvr progenitor candidate. All magnitudes are on the AB system. find it is unlikely that SN 2019yvr coincides with this source bychance, although we acknowledge that this scenario cannot be ruledout definitively before we demonstrate that the source has disappeared(as in the case of iPTF13bvn; Eldridge & Maund 2016; Folatelli et al.2016).Given that SN 2019yvr coincides with a single, bright source,that source is point-like and isolated from nearby sources, and therelatively low likelihood of a chance coincidence, we consider thissource to be a credible progenitor candidate to SN 2019yvr. Belowwe assume that this object is dominated by emission from a singlestellar system that hosted the SN 2019yvr progenitor star.
We show the light curve of the SN 2019yvr progenitor candidate attimes relative to explosion in Figure 6. The source is relatively stablewith at most 0 .
47 mag peak-to-peak variability (corresponding to3.4 𝜎 ) in F555W over a baseline of 110 days. Thus we infer thatthe progenitor candidate did not exhibit any extreme variability withconstant flux at the < − − − − − − . . . . . A pp a r e n t M a gn it ud e ( A B m a g ) F438WF555WF625WF814W
Figure 6.
The pre-explosion light curve of the SN 2019yvr progenitor can-didate in all four
HST filters for which we have imaging. The source is notsignificantly variable, with at most 0 .
47 mag peak-to-peak variations as dis-cussed in Section 4.3. highly variable phase during these observations, as these events aretypically accompanied by large differences in luminosity or colour(as in pre-SN outbursts associated with SNe IIn, e.g., Smith et al.2009; Mauerhan et al. 2013; Kilpatrick et al. 2018a).Thus we are confident that the average photometry across all four
HST bands in which we detect the progenitor candidate is repre-sentative of its overall spectral energy distribution (SED). Takingan inverse-variance weighted average of across all epochs in eachband, we derive average photometry 𝑚 F438W = . ± .
162 mag, 𝑚 F555W = . ± .
032 mag, 𝑚 F625W = . ± .
022 mag, and 𝑚 F814W = . ± .
032 mag as shown in Table 2. Temporarily,ignoring any correction due to host extinction but accounting forMilky Way extinction, the source has 𝑚 F555W − 𝑚 F814W (roughly 𝑉 − 𝐼 ) of 1.065 ± 𝑀 F555W = − . 𝑇 eff = HST or Spitzer observations.
Assuming that the counterpart is dominated by the SED of a singlestar, we estimate the luminosity and temperature of that star by fittingvarious SED models to the
HST photometry. Broadly, we use black-body and stellar SEDs obtained from Pickles & Depagne (2010).We use a full forward modeling and Monte Carlo Markov Chain(MCMC) approach to simulate the in-band apparent magnitudes as-suming the distance above and drawing extinction ( 𝐴 𝑉 ) and redden-ing ( 𝑅 𝑉 ) parameters following the 𝜒 probability distribution fromour light curve analysis in Section 3 and as shown in Figure 5. For MNRAS , 1–22 (2020) rogenitor of SN 2019yvr a blackbody with a given effective temperature 𝑇 eff and luminosity 𝐿 as well as extinction values drawn from the 𝜒 distribution dis-cussed above, we simulate an intrinsic, absolute magnitude 𝑀 𝑖 ineach band 𝑖 and convert to an apparent magnitude 𝑚 𝑖 with in-bandMilky Way extinction 𝐴 MW ,𝑖 , the implied host extinction 𝐴 𝐻 ,𝑖 , andour preferred distance modulus 𝜇 = . 𝑇 eff , 𝐿 ) and extinction parameters ( 𝐴 𝑉 , 𝑅 𝑉 ) drawn from 𝜒 .We then estimate the log likelihood for our MCMC ( 𝜒 ) using theobserved magnitude 𝑚 𝑜,𝑖 and uncertainty 𝜎 𝑜,𝑖 from Table 2 as 𝜒 = ∑︁ 𝑖 (cid:18) 𝑚 𝑖 − 𝑚 𝑜,𝑖 𝜎 𝑖 (cid:19) + 𝜒 ( 𝐴 𝑉 , 𝑅 𝑉 ) . (3)In this way we incorporate the differences between the observedand forward-modeled magnitudes as well as between the values of 𝐴 𝑉 and 𝑅 𝑉 for each trial and the best-fitting values from our colourcurve template fitting. Assuming a blackbody SED, we estimatethe best-fitting parameters 𝑇 eff = + − K and log ( 𝐿 / 𝐿 (cid:12) ) = . + . − . . Although we include the distance modulus uncertainty inour luminosity (and radius) uncertainty estimates, we did not includethis value in our fitting method as it does not affect the overall shapeof the SED. The implied photospheric radius for the best-fittingblackbodies are 𝑅 = ± 𝑅 (cid:12) . As we incorporate 𝐴 𝑉 and 𝑅 𝑉 from the light curve analysis into our models, we also constrainthese parameters with best-fitting values 𝐴 𝑉 = . + . − . mag and 𝑅 𝑉 = . + . − . assuming the intrinsic blackbody spectrum.We also compared our photometry to single-star SEDs from Pick-les & Depagne (2010). We use stars of all spectral classes, fittingonly to a scaled version of the stellar SED as a function of effectivetemperature. Using the same MCMC method, our walkers drew atemperature randomly and then chose the stellar SED with the clos-est effective temperature. The best-fitting SEDs are consistent withstars in the F4 to F0 range (intrinsic 𝑇 eff = + − K; see Figure 7and Figure 8) with an implied luminosity log ( 𝐿 / 𝐿 (cid:12) ) = . ± . 𝑅 = + − 𝑅 (cid:12) . Thus the best-fitting valuesare broadly consistent between blackbody and Pickles & Depagne(2010) model SEDs. There is some systematic uncertainty in the ex-act temperature of the latter models given the sampling of the Pickles& Depagne (2010) spectra, which are increasingly sparse for hotterstars. However, this effect is small at temperatures 5000–10,000 Kwhere there are 40 spectra of varying spectral classes.Our treatment of 𝐴 𝑉 and 𝑅 𝑉 is identical to the forward modelingapproach for the blackbodies above, and we derive best-fitting valuesof 𝐴 𝑉 = . + . − . mag and 𝑅 𝑉 = . + . − . . In both the blackbodyand stellar SED models, our luminosity estimates are on the high-luminosity end for observed core-collapse SN progenitor stars (e.g.,in Smartt et al. 2015) but consistent with most of the SN IIb progenitorstars (see Figure 9).As a further check on the effect of extinction on our derivedparameters, we show the relationship between the host extinction 𝐴 𝑉 and reddening law parameter 𝑅 𝑉 and the implied temperatureand luminosity for the Pickles & Depagne (2010) models in Fig-ure 8. Luminosity is highly correlated with variations in 𝐴 𝑉 , with 𝐴 𝑉 < . f λ [ − e r g s − c m − ˚ A − ] Figure 7.
The best-fitting SEDs to the average pre-explosion
HST photometryof the SN 2019yvr progenitor candidate. Our best-fitting blackbody has 𝑇 eff = 𝑇 eff = ( 𝐿 / 𝐿 (cid:12) )≈ tive temperature > 𝜎 level. This is in starkcontrast with the progenitor of iPTF13bvn with 𝑇 eff ≈ ,
000 K(Cao et al. 2013; Bersten et al. 2014; Eldridge & Maund 2016; Fo-latelli et al. 2016) and He stars generally, which tend to have effectivetemperatures > 𝐴 𝑉 and 𝑅 𝑉 enable a relatively tightfit temperature and luminosity as demonstrated in Figure 8. Theminimum 𝜒 / degreesoffreedom = .
5, which suggests that a single,extinguished star is well matched to our data and we cannot effectivelyconstrain scenarios with more free parameters, such as the inclusionof another star to the overall SED. However, while this analysis mightaccurately reflect the SN 2019yvr progenitor star’s evolutionary stateat 2.6 yr before explosion, it does not place any specific constraintson the pathway that led to this configuration. We further explore theimplications of a SN Ib progenitor star with these properties and theimplications for a single-star origin in Section 5.1.
We also compare the SED of the SN 2019yvr progenitor candidateto binary stellar evolution tracks from BPASS (Eldridge et al. 2017),comprising 12,663 binary star models at a single metallicity. BPASSprovides physical parameters from the binary star system throughoutits evolutionary sequence as well as in-band absolute magnitudesfor the individual components and binary system as a whole. Ouranalysis involved a direct comparison between the F438W, F555W,F625W, and F814W magnitudes of the SN 2019yvr counterpart andthe total binary emission estimated via BPASS synthetic magnitudesin F435W, F555W, SDSS 𝑟 , and F814W, respectively. As BPASS MNRAS , 1–22 (2020) Kilpatrick et al. l og ( T e ff / K ) . . . . . A V ( m a g ) log ( L / L (cid:12) ) . . . . R V log ( T eff / K ) . . . . . A V (mag) . . . . R V Figure 8.
Corner plot showing the correlation between our fit parameterslog ( 𝐿 / 𝐿 (cid:12) ) and log ( 𝑇 / K ) of a single star following a Pickles & Depagne(2010) as well as host extinction 𝐴 𝑉 and host reddening 𝑅 𝑉 as described inSection 4.4. The contours show the 1, 2, and 3 𝜎 best-fitting values derivedfrom all of our samples. Although the luminosity and 𝑉 -band host extinctionare highly correlated, the resulting luminosity and temperature are tightlyconstrained. magnitudes are provided in Vega mag, we transformed our AB magphotometry to Vega mag using the relative Vega mag − AB magzero points for all four WFC3/UVIS filters (0.15, 0.03, − − 𝑍 = . 𝑀 init = . 𝑀 (cid:12) ), mass ratios ( 𝑞 = . . ( 𝑃 / ) = 𝐴 𝑉 and 𝑅 𝑉 as free parameters, but with the walkersdrawing from the same 𝜒 distribution for these parameters as in theblackbody and stellar SED fits above. For our BPASS fits, the best-fitting models correspond to initial mass 𝑀 init = ± 𝑀 (cid:12) , initialmass ratio 𝑞 = . ± .
05 and initial period log ( 𝑃 / ) = . ± . 𝑀 init = 𝑞 = .
1, and log ( 𝑃 / ) = .
6) in Figure 9 andthe luminosity and temperature derived from our Pickles & Depagne(2010) stellar SED fits. We performed our fits by comparing the observed photometry ofthe SN 2019yvr progenitor candidate to the apparent magnitudes in-ferred for the combined flux of both stars in the BPASS models ,and so we are sensitive to scenarios where the flux from either theprimary or companion star dominates the total emission. In all of thebest-fitting models and all four bands we consider, the counterpart isdominated by emission from a ≈ 𝑀 (cid:12) primary star and the com-panion contributes very little to the overall flux. We found no otherbinary scenarios where the total flux was consistent with our pho-tometry at the time one of the stars terminated, including scenarioswhere the secondary star produced the SN explosion instead (i.e., ina neutron star or black hole binary).In the best-fitting model, the terminal state of the SN progenitoris log ( 𝐿 / 𝐿 (cid:12) ) = . 𝑇 eff = 𝑀 final = . 𝑀 (cid:12) , implying a consistent luminosity but a slightlywarmer temperature than we derive from the Pickles & Depagne(2010) models. Similar to above, there are no models at the < 𝜎 level where the exploding star has a terminal temperature > . 𝑀 (cid:12) ) is mostly unchanged with only 0 . 𝑀 (cid:12) of materialaccreted by the time the primary reaches core collapse, implyingthat most of the mass transfer in this model was non-conservative.In addition, the BPASS models predict that 0 . 𝑀 (cid:12) (0.6% massfraction) of hydrogen remains in the primary in its terminal state andno model with < . 𝑀 (cid:12) is consistent with our HST photometryat the 3- 𝜎 level. The best-fitting extinction values for this BPASSmodel was 𝐴 𝑉 = . ± . 𝑅 𝑉 = . + . − . .The primary effect of the BPASS evolutionary models compared tosingle-star models is the inclusion of Roche-lobe overflow (RLOF).For our specific best-fitting model, RLOF turns on in the post-mainsequence phase (i.e., Case B mass transfer; shown with a square inFigure 9), and continues through the end of the primary star’s evo-lution. In particular, mass-loss due to RLOF follows the prescriptionfor common-envelope evolution (CEE) as the radius of the primarystar is smaller than the binary separation throughout post-main se-quence evolution (following prescription in Eldridge et al. 2017). Thebinary separation is only 8 . 𝑅 (cid:12) starting in the post-main sequenceand at the onset of CEE, and so the primary mass-loss rate increasessignificantly to 1–5 × − 𝑀 (cid:12) yr − . This common envelope massloss phase largely determines the final mass and state of the primarystar as it is larger than wind-driven mass loss by a factor of ≈ 𝑀 init > 𝑀 (cid:12) (Eldridgeet al. 2017). Thus stars that terminate near the progenitor candidatein the Hertzsprung-Russell diagram require a specific mass-loss sce-nario where CEE can strip most of the hydrogen envelope but leavea small amount (at least 0 . 𝑀 (cid:12) according to our models), lead-ing to relatively tight constraints on binary mass ratio and periodfor our BPASS fits. However, these parameters are subject to sig-nificant systematic uncertainty in terms of the CEE and mass lossprescriptions assumed. In Section 5.1 we discuss whether these best-fit binary models to the pre-explosion photometry of SN 2019yvr areconsistent with its classification as a type Ib SN. In our BPASS v2.2.1 fits, we examined columns 53–73, representing theabsolute magnitude and colours from the combined flux of both the primaryand companion star.MNRAS , 1–22 (2020) rogenitor of SN 2019yvr Given our analysis in Section 4, we assume throughout this discussionthat the SN 2019yvr pre-explosion counterpart is dominated by emis-sion from the SN progenitor system. From our inferences about thissource above as well as our knowledge of SN 2019yvr, we considerwhat evolutionary pathways could lead to the source observed in the
HST photometry as well as the resulting SN. These pathways needto explain several facts referenced throughout the previous analysis,which we summarize here as:(i) SN 2019yvr was a SN Ib with no evidence for hydrogen inits early-time spectra, starting from 7 days before peak light(Dimitriadis et al. 2019) until well after peak light. Followingmodels in Dessart et al. (2012), this suggests that the progen-itor star must have had < − 𝑀 (cid:12) (but possibly as much as0 . 𝑀 (cid:12) ; Hachinger et al. 2012) of hydrogen remaining in itsenvelope at the time of explosion.(ii) SN 2019yvr began interacting with CSM starting around150 days after explosion and exhibited strong H 𝛼 , radio, and X-ray emission consistent with a shock formed in hydrogen-richmaterial (Auchettl et al. in prep.). Using conservative assump-tions about the SN shock velocity (10,000 km s − ) and velocityof the CSM (100 km s − ), we infer that this material must havebeen ejected at least ≈
44 yr prior to core collapse and implyinga shell or clump of material at > 𝐴 𝑉 ≈ . 𝐸 ( 𝐵 − 𝑉 ) = .
02 mag. We infer from the strong Na i D line at the redshiftof the SN 2019yvr host galaxy NGC 4666 that this extinctionimplies a significant dust column in the NGC 4666 interstellarmedium toward SN 2019yvr and/or a correspondingly largecolumn of circumstellar dust in the environment of the pro-genitor system itself. There is no clear evidence for any massejections or warm circumstellar gas in a compact shell aroundthe progenitor system, either in pre-explosion data or fromcircumstellar interaction once SN 2019yvr exploded.(iv) There is a single, point-like progenitor candidate toSN 2019yvr detected in pre-explosion
HST imaging. Thissource exhibits very little photometric variability over a110 day period from 2.7 to 2.4 yr prior to core collapse. Be-fore applying our host extinction estimate but applying MilkyWay extinction, this progenitor candidate has a red colour of 𝑚 F555W − 𝑚 F814W = .
065 mag. Accounting for the inferredextinction and distance modulus above, the source is relativelyluminous with 𝑀 F555W = − . 𝑉 -band). Thisvalue is consistent with massive stars but low for a stellarcluster (as in Gieles et al. 2006).(v) Accounting for all extinction, the progenitor candidate is con-sistent with a log ( 𝐿 / 𝐿 (cid:12) ) =5.3 ± 𝑇 eff ≈ ≈ 𝑅 (cid:12) at 2.6 yr prior to the SN 2019yvr explosion. Such a star wouldbe closest in temperature and luminosity to yellow supergiantsconfirmed as SN IIb progenitor stars (Figure 9 and green cir-cles for SNe 1993J, 2008ax, 2011dh, 2013df, and 2016gkg).Comparing to stellar SEDs, the spectral type and luminosityclass are best matched to F4–F0 supergiant stars.(vi) Comparing the pre-explosion photometry to BPASS binarystellar evolution tracks in Eldridge et al. (2017), the best-fittingmodel is a 19 𝑀 (cid:12) +1.9 𝑀 (cid:12) system that undergoes commonenvelope evolution and strips most of the material from theprogenitor star. Immediately prior to explosion, the primarystar retains 0.047 𝑀 (cid:12) of hydrogen in its envelope, inconsistentwith the masses for SN Ib systems given above. No otherBPASS models were consistent with both our pre-explosionphotometry and a system that produced a SN explosion. We compare the SN 2019yvr progenitor candidate in a Hertzsprung-Russell diagram to other known progenitor systems in Figure 9,including SNe II (Smartt et al. 2015, and references therein), SNe IIb(SNe 1993J, 2008ax, 2011dh, 2013df, and 2016gkg; Aldering et al.1994; Crockett et al. 2008; Maund et al. 2011; Van Dyk et al. 2014;Kilpatrick et al. 2017), and the progenitor of the SN Ib iPTF13bvn(Cao et al. 2013; Bersten et al. 2014; Folatelli et al. 2016). We alsonote that there is a single SN Ic progenitor candidate for SN 2017ein(Kilpatrick et al. 2018b; Van Dyk et al. 2018), but the source hasa luminosity log ( 𝐿 / 𝐿 (cid:12) )≈ 𝑇 eff > K and sois off our plotting range (and may actually be a very young opencluster). For comparison, we also overplot single-star evolutionarytracks from the Mesa Isochrones & Stellar Tracks code (Choi et al.2016, 2017).The most notable feature of the SN 2019yvr progenitor candi-date in Figure 9 and compared with other stripped-envelope SNprogenitor stars is its relatively cool effective temperature, whichin turn implies an extended photosphere given our constraints onits SED. Assuming a single-star origin and judging solely by thesource’s inferred luminosity and temperature of log ( 𝐿 / 𝐿 (cid:12) ) =5.3 and 𝑇 eff = 𝑀 (cid:12) star in theso-called “Hertzsprung gap” (see, e.g., de Jager & Nieuwenhuijzen1997; Stothers & Chin 2001). This is in sharp contrast to the predictedprogenitors of SNe Ib, which are thought to have low-mass, compactenvelopes, consistent with a star that has almost no hydrogen in itsouter layers (Yoon 2015).Given these facts, we consider the implications of different pro-genitor scenarios. In particular, we emphasize the apparent contra-diction between a SN from a star without a significant hydrogenenvelope mass and a pre-explosion counterpart consistent with amassive star with a significantly extended photosphere, which typi-cally requires a non-negligible hydrogen envelope mass. Combinedwith the SN 2014C-like circumstellar interaction observed at latetimes, SN 2019yvr may offer significant insight into mass loss inlate-stage stellar evolution for stripped-envelope SNe. Here we re-view “standard” single and binary star scenarios and assess whetherthey can explain all of these properties. It is still debated if single massive stars can evolve to the pointwhere they would explode as yellow supergiants, in part because the
MNRAS , 1–22 (2020) Kilpatrick et al. M ⊙ M ⊙ M ⊙ M ⊙ M ⊙ M ⊙ M ⊙ . . . . . .
75 log ( T eff / K ) . . . . . . . l og ( L / L ⊙ ) SN IISN IIbSN Ib/c M ⊙ . . . . . . ( T eff / K ) . . . . . l og ( L / L ⊙ ) SNRLOF > − M ⊙ yr − Yoon et al. 2017 IIbYoon et al. 2017 Ib
Figure 9. ( Top ): A Hertzpsrung-Russell diagram showing the location ofthe SN 2019yvr progenitor candidate (blue star) with comparison to SN IIbprogenitor stars (green squares), iPTF13bvn (blue diamond; Cao et al. 2013;Bersten et al. 2014; Folatelli et al. 2016), and SN II progenitor stars (redsquares; Smartt et al. 2015). We overplot MIST single-star evolutionarytracks from Choi et al. (2016, 2017) for comparison.
Bottom : A 19+1.9 𝑀 (cid:12) binary star evolution track from BPASS v2.2 (Eldridge et al. 2017), whichis consistent with the pre-explosion photometry of the SN 2019yvr progeni-tor candidate. We highlight the location on the track where the primary starbegins Roche-lobe overflow (RLOF; square), reaches its minimum hydrogenenvelope mass (0 . 𝑀 (cid:12) , circle), and terminates as a supernova (star). Wealso show binary star models from Yoon (2017) with outcomes predicted fortype Ib (blue) and IIb SNe (green). Hertzsprung gap is typically a short-lived and transitional phase inthe post-main sequence. Standard single-star evolution would sug-gest that a star with an effective temperature of 𝑇 eff = ( 𝐿 / 𝐿 (cid:12) ) =5.3 retains a massive hydrogen envelope: a 30 𝑀 (cid:12) ini-tial mass star would retain a 60% surface hydrogen mass fraction onits first passage through the Hertzsprung gap (following the struc-ture of model stars in Choi et al. 2016). Thus, in the context ofstripped-envelope SNe, any single-star model would likely enter theyellow supergiant phase after evolving through the RSG branch and shedding the remainder of its hydrogen envelope. Georgy (2012)demonstrate that such evolution is possible if RSG mass loss ratesare increased by approximately an order of magnitude towards theend of their nuclear lives.While the physics of an enhanced late-stage mass loss is unclear(see, e.g. Yoon & Cantiello 2010), there do exist a number of lu-minous yellow supergiants (termed “yellow hypergiants”) which arehypothesized to be such post-RSG stars. Many yellow hypergiantsare extremely variable with high mass-loss rates, such as 𝜌 Cas(Smith 2014; Lobel et al. 2015) and V509 Cas (Percy & Zsoldos1992), and are located toward the end of the luminous blue variable(LBV) bistability track (Smith et al. 2004). This variability involvesa rapid (years to decades) evolution between quiescent, hot phasesand erupting, cool phases with extreme mass-loss episodes (e.g.,10 − 𝑀 (cid:12) yr − as in the yellow hypergiants 𝜌 Cas and HR 8752; deJager 1998; Humphreys et al. 2002).Yellow hypergiants undergoing extreme mass loss have been pro-posed as candidates for type Ib/c progenitors as their LBV-like massejections provide an efficient way to rid the star of its hydrogen andhelium envelope in the years before core collapse (as in SN 2006jc;Foley et al. 2007; Smith & Conti 2008). The presence of these starsin the gap and the fact that some of them explode as type IIb SNe(the bluest SN IIb progenitor stars span this gap, e.g., Crockett et al.2008; Kilpatrick et al. 2017) strengthens the association between ex-treme mass loss and stripped-envelope SNe (de Jager 1998; Stothers& Chin 1999). While it is debated if such enhanced mass-loss ratesare possible for single stars on the RSG branch (e.g., Beasor et al.2020), mass-loss episodes and variability of stars near this point onthe Hertzpsrung-Russell diagram suggests it is possible to rapidlyshed their hydrogen envelopes and increase in temperature overtimescales of years (as observed with HR 8752, for example; deJager & Nieuwenhuijzen 1997). However, yellow hypergiants thatwe observe in this region of the Hertzsprung-Russell diagram havemassive, hydrogen-rich envelopes and their winds are known to behydrogen rich (Smith et al. 2004). Assuming such a star exploded asa SN Ib only 2.6 yr after being observed in this evolutionary phase,it would either need to shed its remaining envelope in the final 2.6 yror remain cool after retaining only a trace hydrogen envelope. Theformer scenario will be discussed in Section 5.2, below.Alternatively, it is worth considering if a single star with virtuallyno remaining hydrogen in its envelope could inflate to a radius of320 𝑅 (cid:12) while exhibiting a photospheric temperature of 6800 K—asrequired for SN 2019yvr and its progenitor candidate. The coolestknown helium stars are only ≈ 𝑇 eff = ,
000 K (Drilling et al. 1984; Schoenberner& Drilling 1984), KS Per is 10,000 K (Woolf 1973; Drilling &Schonberner 1982), and 𝜈 Sgr is 11,800 K (Frame et al. 1995). Thecounterpart we observe is only this hot assuming our SED modelingis inconsistent with the true temperature at > 𝜎 or if there is > ( 𝐿 / 𝐿 (cid:12) )≥ ( 𝐿 / 𝐿 (cid:12) ) =4.6 (see pre-vious references and Dudley 1992). Thus an anomalously high-massand luminous helium star would be needed to match to SN 2019yvr,which in general is not allowed by standard single star stellar evolu-tion models. MNRAS000
000 K (Drilling et al. 1984; Schoenberner& Drilling 1984), KS Per is 10,000 K (Woolf 1973; Drilling &Schonberner 1982), and 𝜈 Sgr is 11,800 K (Frame et al. 1995). Thecounterpart we observe is only this hot assuming our SED modelingis inconsistent with the true temperature at > 𝜎 or if there is > ( 𝐿 / 𝐿 (cid:12) )≥ ( 𝐿 / 𝐿 (cid:12) ) =4.6 (see pre-vious references and Dudley 1992). Thus an anomalously high-massand luminous helium star would be needed to match to SN 2019yvr,which in general is not allowed by standard single star stellar evolu-tion models. MNRAS000 , 1–22 (2020) rogenitor of SN 2019yvr As shown above, the best-fitting binary systems for our SN 2019yvrpre-explosion photometry are short-period, low mass ratio systemsthat undergo CEE. While such systems are expected to have complexcircumstellar environments, possibly consistent with SN 2019yvr,we emphasize that all of the BPASS models terminate with finalhydrogen-envelope masses of > 𝑀 (cid:12) . This is inconsistent witha classification as a SN Ib via the models of Dessart et al. (2012) andHachinger et al. (2012). In addition, there is debate as to whetherthe evolution of such high mass ratio binaries within BPASS is rep-resentative of actual stellar systems. Neugent et al. (2018, 2020)argue that any star that will eventually explode as a core-collapseSN evolves off the main sequence fast enough that companion starswith initial masses (cid:46) 𝑀 (cid:12) would not have enough time to completetheir contraction phase and thus would still be protostars. This is notaccounted for in current BPASS models, which initialize all stellarmasses on the ZAMS simultaneously.More broadly, the progenitor systems of SNe Ib are predicted to below-mass helium stars that evolve via binary evolution. While theycan expand to moderately large radii due to shell burning, they typi-cally remain < 𝑅 (cid:12) (Kleiser et al. 2018; Laplace et al. 2020). Theirphotospheric temperatures are thus significantly hotter than the HST photometry of the SN 2019yvr progenitor candidate implies. Thelatter allows at most 11,000 K as opposed to the ≈ 𝑇 eff > < few day) orbits wherethe primary has been stripped to 3–10 𝑀 (cid:12) , likely through Case Bmass transfer (Yoon et al. 2012; Yoon 2015, 2017). The companioncan span a wide range of luminosities and evolutionary states (e.g.,WR7a exhibits a 0.204 day binary orbital period, but no secondary isobserved implying a very low-mass star or a compact object; Pereiraet al. 1998; Oliveira et al. 2003). The primary stars are left completelystripped of a hydrogen envelope. This is in contrast to the best-fittingBPASS model, which undergoes extreme mass loss due to CEE andRLOF during the post-main sequence but stops before the hydrogenenvelope is completely depleted (Delgado & Thomas 1981; Ivanovaet al. 2013; Eldridge et al. 2017).Thus, as with the single star models, there remains significanttension as to whether any standard binary evolution model can re-produce both the progenitor candidate and a system that explodes asa SN Ib. We are left needing to invoke some additional mechanismthat can account for both the cool photospheric temperature 2.6 yrprior to explosion and the negligible hydrogen envelope mass at thetime of core collapse. In the following sections we discuss two po-tential resolutions to this paradox (Section 5.2 and Section 5.3) andhighlight some evolutionary scenarios that may be allowed whilealso explaining the dense CSM shell observed around SN 2019yvr(Section 5.4). One way to resolve the apparent conflict between the extended pro-genitor radius and lack of hydrogen in the ejecta of SN 2019yvr wouldbe if the progenitor star did possess a envelope with ∼ (cid:12) of hydrogen (similar to the yellow supergiant progenitor stars ofSNe IIb) at the time of the HST observations, but somehow man- aged to lose this material in the intervening 2.6 yr. Such a scenariois not unprecedented: episodic mass ejections have been invoked toexplain stripped-envelope SN progenitor systems that exhibit denseshells or clumps of CSM (see, e.g., Chugai & Chevalier 2006; Bi-etenholz et al. 2014; Chandra 2018; Mauerhan et al. 2018; Pooleyet al. 2019; Sollerman et al. 2020; Tartaglia et al. 2020). This was alsoobserved directly for SN 2006jc, a SN Ib with a pre-explosion out-burst 2 yr before explosion (Foley et al. 2007; Smith & Conti 2008;Pastorello et al. 2008; Maund et al. 2016), which later manifestedas circumstellar interaction. Indeed, SN 2019yvr shows evidence forrelatively narrow H 𝛼 , radio, and X-ray emission ≈
150 days after corecollapse, providing evidence for episodic mass ejections throughoutthe progenitor’s final evolutionary stages.In this scenario, this hydrogen-rich material should be locatedat some radius around the progenitor star. A key question forSN 2019yvr is whether the shell of H-rich material encountered bythe SN ejecta ∼
150 days post-explosion could be the remnants ofsuch an ejection that occurred between the
HST observations andexplosion. While the location and timing of this ejection will be dis-cussed in detail in Auchettl et al. (in prep.), we perform an orderof magnitude calculation to investigate this possibility here. For aCSM wind velocity ( 𝑣 𝑤 ) and a SN shock velocity ( 𝑣 𝑠 ) the detectionof CSM interaction starting ∼
150 days after core collapse implies amass-loss event that occurred: 𝑡 mle ≈ (cid:18) 𝑣 𝑠 kms − (cid:19) (cid:18) 𝑣 w
100 kms − (cid:19) − yrs (4)before core collapse. We have scaled our results to an average shockvelocity of 10,000 km s − and wind velocity of 100 km s − . Theformer is roughly consistent with the velocity inferred from heliumabsorption in our spectra of SN 2019yvr near maximum light, whichis a lower limit for the shock velocity. While the latter is less con-strained, an ejection speed of ∼
100 km s − is approximately theescape speed for a star with a radius of 320 𝑅 (cid:12) (as inferred from ourprogenitor candidate) and a mass of 5 −
10 M (cid:12) . For these assump-tions, the mass loss event would have occurred significantly earlierthan the time of the
HST observations.Therefore, if the hydrogen we observe in CSM is the same mate-rial inferred from the progenitor star photosphere, the wind velocitymust be at least 1800 km s − such that the progenitor star couldeject it after the HST data were obtained. A wind at this speed canonly be achieved by a compact and massive WR-like star (similarto those presented in Rochowicz & Niedzielski 1995; van der Hucht2001), inconsistent with our pre-explosion photometry. Such a highvelocity would therefore require and ejection mechanism capable ofaccelerating material to (cid:38) × the stellar escape speed. This would bein contrast to current theoretical models for both wave-driven massloss in hydrogen-poor stars (which have terminal velocities of a fewhundred km s − Fuller & Ro 2018) and common-envelope ejections(which tend to proceed at roughly the escape velocity).Unless a stronger ejection mechanism can be identified, while themass-loss event that led to the material at ≈ HST imaging.Multiple eruptive mass-loss events would be needed to explain boththe CSM and the final depletion of the progenitor star’s hydrogenenvelope assuming the pre-explosion counterpart, the source of thematerial around SN 2019yvr, and the SN 2019yvr progenitor starare all the same. The second ejection would place material closerto the progenitor star, which is not detected in our light curves orspectra (e.g., via enhanced emission due to CSM interaction or flash
MNRAS , 1–22 (2020) Kilpatrick et al. ionisation features similar to Gal-Yam et al. 2014). It is possible thatthis material was missed due to a delay between the explosion ofSN 2019yvr and its discovery, and detailed analysis of these datawill be carried out in Auchettl et al. (in prep.)to assess whether ornot this could be the case.
If the scenario discussed in Section 5.2 is not viable, then theSN 2019yvr progenitor candidate must have contained virtually nohydrogen 2.6 yr before explosion ( < − 𝑀 (cid:12) as required for a typeIb classification by Dessart et al. 2012; Hachinger et al. 2012). Inthis case, we would require that the envelope was inflated by someprocess not accounted for in standard models of stellar evolution.Here, we consider two scenarios in which a star with only a tracehydrogen envelope could exhibit a photospheric radius of ≈ 𝑅 (cid:12) or roughly 1.5 AU: (a) formation of a pseudo-photosphere in a densestellar wind, and (b) inflation due to an additional heat/energy sourcethat produces a radiation pressure supported envelope. For certain mass-loss rates, a stellar wind can become optically thickand appear at a radius well beyond that of the underlying star (see,e.g., Gallagher 1992). The radius at which this quasi-photosphereforms ( 𝑅 𝜏 ) depends on the density ( 𝜌 ) and opacity ( 𝜅 ) in the wind,with optical depth ( 𝜏 ) expressed as 𝜏 = ∫ ∞ 𝑅 𝜏 𝜌𝜅𝑑𝑟. (5)From this equation, a quasi-photosphere will form when 𝜏 (cid:38) 𝑟 − for aconstant mass-loss rate ( (cid:164) 𝑀 ), these conditions are largely dependenton the properties of the wind and thus the underlying star. Fromequation (16) in de Koter et al. (1996), where the wind opacityis modeled as temperature-dependent bound-free opacity from thePaschen continuum, the quasi-photosphere constraint above implies (cid:164) 𝑀 > . × − 𝑀 (cid:12) yr − (cid:18) 𝑇 eff K (cid:19) / (cid:18) 𝑅 𝜏 𝑅 (cid:12) (cid:19) / (cid:18) 𝑣 𝑤
100 km s − (cid:19) / (cid:18) 𝑐 𝑠
10 km s − (cid:19) / (6)with 𝑣 𝑤 the wind velocity and 𝑐 𝑠 the local sound speed.To estimate mass loss rate required to explain the photosphere ob-served for the progenitor candidate of SN 2019yvr, we assume thatthe wind could be as fast as 100 km s − ,as described in § 5.2 (a de-tailed analysis of the CSM properties will be presented in Auchettl etal. (in prep.)). We also use our constraint on the effective temperatureof the observed photosphere 𝑇 eff = 𝑅 = 𝑅 (cid:12) above, which gives a local sound speed 𝑐 𝑠 ≈ − and implies (cid:164) 𝑀 > . × − 𝑀 (cid:12) yr − .This mass-loss rate is extreme for a star with log ( 𝐿 / 𝐿 (cid:12) ) = . (cid:164) 𝑀 ≈ − 𝑀 (cid:12) yr − (Humphreys & Davidson 1994). However, the total mass of materialneeded to form a quasi-photosphere is only (cid:164) 𝑀𝑣 − 𝑤 𝑅 𝜏 ≈ − 𝑀 (cid:12) .This material can easily be shed from the star’s remaining hydro-gen/helium envelope over a brief period of strong mass loss, althoughthe timescale for such a shell to form would be only 𝑅 𝜏 / 𝑣 𝑤 =
26 days. However, we do not observe any emission lines or “flash spec-troscopy” features in our early spectra of SN 2019yvr as may beexpected if dense wind material is present. This scenario would alsorequire a significant change in the mass loss from the SN 2019yvrprogenitor timed (cid:46)
An alternative is that the progenitor star retained a small amountof hydrogen ( < − 𝑀 (cid:12) as required for a type Ib classification byDessart et al. 2012; Hachinger et al. 2012) with an envelope that hasrelaxed to a steady-state configuration following some injection ofadditional energy. Assuming that radiation pressure dominates overgas pressure, the radial density profile in the wind will follow 𝑟 − (Loeb & Ulmer 1997). Hydrostatic equilibrium holds up to the radiusat which the radiation is no longer trapped, and so we associate theouter radius of this envelope with the photospheric condition 𝜏 = 𝑅 𝜏 ≈ (cid:18) 𝑀 env − 𝑀 (cid:12) (cid:19) / 𝑅 (cid:12) , (7)where we assume Thompson opacity and a hydrogen envelope witha Solar mass fraction. We note that an inflated progenitor could existprovided that the mass contained in the envelope is > − 𝑀 (cid:12) , whichis both reconcilable with a type Ib classification for SN 2019yvr andin general is less restrictive than the condition given by Equation 6.However, even in this case several mysteries remain as to the natureof the progenitor system. In particular, for a complete description ofSN 2019yvr, the progenitor system would still need to eject a shellof hydrogen-rich material at ∼ In the sections above, we have argued that reconciling the extendedprogenitor radius and lack of hydrogen in the ejecta of SN2019yvrrequires a system that either:(i) ejected the remainder of its hydrogen envelope in the final 2.6yrs pre-explosion (Section 5.2).(ii) undergoes a process not accounted for in standard models ofevolution and that leads to additional inflation or has sufficientmaterial around the progenitor to form a quasi-photosphere inthe CSM (Section 5.3).In addition, in either case, a viable progenitor scenario forSN 2019yvr must also explain a shell of hydrogen-rich material ob-served at ∼ One evolutionary pathway that naturally explains many of the ob-servables for SN 2019yvr is a series of eruptions from a massivestar in a LBV phase. These eruptions are known to precede many
MNRAS000
MNRAS000 , 1–22 (2020) rogenitor of SN 2019yvr SNe, some of which are thought to be core-collapse explosions (aswas argued for SN 2009ip Mauerhan et al. 2013), although this in-terpretation of LBVs as a phase immediately preceding core collapseremains controversial (Margutti et al. 2014).Many types of progenitor systems exhibit pre-SN eruptions in thefinal years to weeks before explosion, notably for the progenitorsof SNe IIn (SNe with narrow Balmer lines in their optical spectra,indicative of interaction between ejecta and a dense mass of CSM;Smith 2017). These systems must eject from 0.1–10 𝑀 (cid:12) over shorttimescales (Smith & McCray 2007; Fox et al. 2010, 2011) requiringan extreme mass-loss rate and variability on the timescale of theeruptions. Their progenitor systems have also been observed in theliterature, notably for SN 2009ip whose progenitor star had an initialmass > 𝑀 (cid:12) (Smith et al. 2009; Foley et al. 2011). Although there issome ambiguity whether these events are actually terminal explosionsof massive stars (Mauerhan et al. 2013; Margutti et al. 2014), theeruptive mechanism that fills their environments with dense shells ofCSM may be more common among stars at a wide range of initialmasses. Indeed, so-called SN impostors (lower luminosity, likelynon-terminal explosions of massive stars that resemble SNe IIn; VanDyk et al. 2000; Maund et al. 2006; Pastorello et al. 2010; Ofeket al. 2016) come from systems with initial masses possibly as lowas 20 𝑀 (cid:12) (Kilpatrick et al. 2018a). Such an eruption for SN 2019yvrand SN 2014C-like events (hypothesized by Milisavljevic et al. 2015)could explain the source of the CSM, although most SN IIn and SNimpostor progenitor stars retain some hydrogen leading to broadBalmer lines in their spectra.For SN 2019yvr, we would require a scenario analogous to SNe IInand Ibn with strong circumstellar interaction soon after explosion(Smith 2017; Hosseinzadeh et al. 2017) but involving multiple mass-loss episodes timed years or even decades ahead of core collapseinstead of months to years. The main caveat in invoking this pro-genitor scenario is whether episodic mass ejections can fully stripthe progenitor star’s hydrogen envelope on the timescale required byour HST observations. While LBV eruptions provide a compellingmass-loss scenario as progenitor systems extend nearly to the pa-rameter space of the Hertzsprung-Russell diagram where we find theSN 2019yvr counterpart (see SN 2009ip, 2015bh, Gaia16cfr; Smithet al. 2009; Mauerhan et al. 2013; Elias-Rosa et al. 2016; Thöne et al.2017; Kilpatrick et al. 2018a), these events tend to result in hydrogen-rich explosions with long-lived Balmer emission. The final eruptionsof SN 2019yvr would need to strip a sufficiently small amount ofhydrogen (potentially hundredths of a Solar mass as in Section 4.5)that interaction with this material would not be observed in early timespectra. Detailed analysis of the initial observations of SN 2019yvrcould place stronger constraints on whether this scenario occurred.
An alternative evolutionary pathway in the final decades of evolutionfor the SN 2019yvr, and put forward for SN 2014C in Margutti et al.(2017), is CEE leading to ejection of the primary star’s hydrogen en-velope <
100 yr before explosion. In general, this is not expected asRLOF and CEE are commonly associated with mass transfer muchearlier in the primary star’s life cycle, for example, immediately afterhelium core contraction as discussed in Section 1. The progenitorwould need to be inflated in a later evolutionary stage after heliumburning to restart mass transfer (i.e., through Case C mass trans-fer, see Schneider et al. 2015). This is predicted to occur for only ∼ −
6% of binary system with primary masses between 8 −
20 M (cid:12) (Podsiadlowski et al. 1992).However, if the primary star’s envelope can inflate during this phase, the companion will spiral inward and start CEE, resulting inthe ejection of a significant fraction of the envelope over < <
100 km s − depending on the envelope structure of the primarystar at onset of CEE. In this way, the material could survive longenough that the SN ejecta can encounter it within the first year aftercore collapse, as observed at ∼
150 days in SN 2019yvr.The key question for SN 2019yvr is how the post-CEE systemevolves in the final years before core collapse such that it resem-bles the pre-explosion counterpart and also explodes as a SN Ib.Case-C CEE is an appealing solution to this problem as the remain-ing envelope is predicted to become unstable, leading to dynamicalpulsations and steady mass loss in the remaining years before corecollapse (especially for extremely luminous stars; Heger et al. 1997).Thus a post-CEE star could still form a quasi-photosphere assumingthe mass-loss rate exceeded 10 − 𝑀 (cid:12) yr − as in Section 5.3. Alterna-tively, the inspiral during CE phase itself supplies a source of heat inthe stellar envelope. Thus, after the ejection of most of its hydrogenenvelope, the resulting post-CEE hydrogen-deficient envelope couldremain partly inflated. If the resulting star was dynamically unsta-ble and losing mass significantly faster than the ≈ − 𝑀 (cid:12) yr − predicted for yellow hypergiants (Humphreys & Davidson 1994), itcould rapidly shed even the 10 − –10 − 𝑀 (cid:12) that is predicted to re-main in the envelope of comparable stars from BPASS. This wouldsimultaneously explain the circumstellar material as the result ofCEE, the apparently cool photosphere, and the lack of hydrogen inthe primary star’s envelope at core collapse. SNe Ib with late-time circumstellar interaction similar to SN 2019yvrare not unprecedented and may in fact represent a large fraction ofstripped-envelope SNe overall. As discussed in Section 1, SN 2014Cbegan interacting with its circumstellar environment a few weeksafter explosion (Margutti et al. 2017), and both SN 2019yvr andSN 2014C have relatively deep progenitor constraints consideringboth have pre-explosion
HST imaging and occurred at ≈
15 Mpc(Milisavljevic et al. 2015). More broadly, there are numerous exam-ples of stripped-envelope SNe with circumstellar interaction soonafter explosion (Chugai & Chevalier 2006; Bietenholz et al. 2014;Chandra 2018; Mauerhan et al. 2018; Pooley et al. 2019; Sollermanet al. 2020; Tartaglia et al. 2020), and so any mechanism that we in-voke above may need to explain the presence of CSM for a significantfraction of all core-collapse SNe (Margutti et al. 2017).A compact star cluster toward the position of SN 2014C in pre-explosion imaging (Milisavljevic et al. 2015) had a best-fitting age inthe range 30–300 Myr, although it could be as young as 10 Myr. Thisage implies a main-sequence turnoff mass of 3 . . 𝑀 (cid:12) dependingon the metallicity of the cluster. Assuming this cluster hosted theSN 2014C progenitor star and considering the fact that stars in thismass range either do not explode as core-collapse SNe or are thoughtto lead to SNe II, Milisavljevic et al. (2015) inferred that a morelikely explosion scenario for SN 2014C was through binary starchannels for a star with a LBV-like phase and 𝑀 ZAMS > 𝑀 (cid:12) (implying that the cluster is younger than its colours suggest). Giventhe photometric and spectroscopic similarity between SNe 2014Cand 2019yvr and the presence of a strong density gradient in thecircumstellar environments of both, it is possible that both systemscould be explained through eruptive, LBV-like mass loss or CEE asdiscussed above. MNRAS , 1–22 (2020) Kilpatrick et al.
However, these evolutionary phases would require that theSN 2019yvr progenitor star had extreme photometric variability,and the lack of any such variability in the pre-explosion photom-etry indicates that these episodes would have occurred outside theshort window in which we constrain the progenitor star’s evolution.Assuming that the mass-loss episode was driven by an instability inlate-stage nuclear burning (e.g., Arnett et al. 2014), the star couldalso be inflated temporarily. Smith & Arnett (2014) suggest thisinflation would occur on a timescale comparable to the orbital pe-riod in binary systems, but we see no signature of this variabilityon 30–100 day timescales. Deep, high-cadence limits, such as thosefrom the Young Supernova Experiment (Jones et al. 2019, 2020) orthe upcoming Vera C. Rubin Observatory Legacy Survey of Spaceand Time (down to r ≈ We present pre-explosion imaging, photometry, and spectroscopy ofthe SN Ib 2019yvr. We find:(i) SN 2019yvr was a SN Ib with a large line-of-sight extinctionspectroscopically similar to iPTF13bvn. SN 2019yvr exhibitedsignatures of interaction 150 days after discovery consistentwith a shock between SN ejecta and dense, hydrogen-rich CSM(Auchettl et al. in prep.) and similar to SN 2014C (Milisavljevicet al. 2015; Margutti et al. 2017). This interaction suggests thatthe SN 2019yvr progenitor star underwent an eruptive mass-loss episode at least 44 years before explosion.(ii) There is a single source in
Hubble Space Telescope imagingobtained ≈ ( 𝐿 / 𝐿 (cid:12) ) = . ± . 𝑇 eff = + − K, and thus close( < 𝑀 (cid:12) . However, the cool effective temperature and high lu-minosity implies a remaining hydrogen envelope mass of atleast 0 . 𝑀 (cid:12) in the binary star model, which is inconsistentwith the lack of hydrogen in spectra of SN 2019yvr.(iii) Comparison to SN Ib progenitor candidates indicates thatSN 2019yvr is much cooler than what is predicted for ahydrogen-stripped star, much more similar to the identifiedprogenitors of type IIb SNe. Overall, the progenitor candi-date appears cool and inflated relative to the progenitor ofiPTF13bvn and helium stars.(iv) We infer that an extreme and episodic mass-loss scenario isrequired to produce both a stripped-envelope SN progenitorsystem and the luminous, cool progenitor candidate. The bi-nary evolution scenarios discussed above do not incorporatephysical scenarios that can lead to extreme or eruptive massloss soon before explosion, and barring such a mass-loss sce-nario they do not produce a star whose hydrogen envelope isconsistent with the SN Ib classification. We propose that LBV-like mass ejections or CEE provide natural explanations forthe stellar classification, the lack of a massive hydrogen enve-lope, and the presence of dense CSM. We hypothesize that ifthis mass-loss mechanism occurs, the star could have formeda quasi-photosphere from CSM in its environment, requiring either a mass loss at a rate > . 𝑀 (cid:12) yr − or a radiation sup-ported hydrogen envelope with a mass > − 𝑀 (cid:12) at 2.6 yrbefore core collapse. ACKNOWLEDGMENTS
We thank J.J. Eldridge and H. Stevance for helpful comments about ourBPASS analysis, J. A. Vilchez, A. Campillay, Y. K. Riveros and N. Ulloafor help with the Swope observations, as well as R. Carrasco for supportof our Gemini-S/GSAOI programme. C.D.K. acknowledges support throughNASA grants in support of
Hubble Space Telescope programmes GO-15691and AR-16136. M.R.D. acknowledges support from the NSERC throughgrant RGPIN-2019-06186, the Canada Research Chairs Program, the Cana-dian Institute for Advanced Research (CIFAR), and the Dunlap Institute atthe University of Toronto. K.A. is supported by the Danish National Re-search Foundation (DNRF132) and VILLUM FONDEN Investigator grant(project number 16599). Parts of this research were supported by the Aus-tralian Research Council Centre of Excellence for All Sky Astrophysics in3 Dimensions (ASTRO 3D), through project number CE170100013. D.O.J.is supported by a Gordon and Betty Moore Foundation postdoctoral fel-lowship at the University of California, Santa Cruz. Support for this workwas provided by NASA through the NASA Hubble Fellowship grant HF2-51462.001 awarded by the Space Telescope Science Institute, which is oper-ated by the Association of Universities for Research in Astronomy, Inc., forNASA, under contract NAS5-26555. The UCSC team is supported in partby NASA grant NNG17PX03C, NASA grants in support of
Hubble SpaceTelescope programmes AR-14296 and GO-16239 through STScI, NSF grantAST-1815935, the Gordon & Betty Moore Foundation, the Heising-SimonsFoundation, and by a fellowship from the David and Lucile Packard Founda-tion to R.J.F. J.H. was supported by a VILLUM FONDEN Investigator grant(project number 16599). W.J-G is supported by the National Science Founda-tion Graduate Research Fellowship Program under Grant No. DGE-1842165.R.M. acknowledges support by the National Science Foundation under AwardNo. AST-1909796. R.M. acknowledges support by the National Science Foun-dation under Awards No. AST-1909796 and AST-1944985. R.M. is a CIFARAzrieli Global Scholar in the Gravity & the Extreme Universe Program,2019 and a Alfred P. Sloan Fellow in Physics, 2019. The Margutti teamat Northwestern is partially funded by the Heising-Simons Foundation undergrant
Hubble Space Telescope , obtained from the data archive atthe Space Telescope Science Institute. STScI is operated by the Associationof Universities for Research in Astronomy, Inc. under NASA contract NAS5-26555. This work is based in part on observations made with the
SpitzerSpace Telescope , which was operated by the Jet Propulsion Laboratory, Cal-ifornia Institute of Technology under a contract with NASA. (AST-1911206and AST-1852393).MNRAS000
SpitzerSpace Telescope , which was operated by the Jet Propulsion Laboratory, Cal-ifornia Institute of Technology under a contract with NASA. (AST-1911206and AST-1852393).MNRAS000 , 1–22 (2020) rogenitor of SN 2019yvr Facilities : Gemini (GSAOI),
HST (WFC3), Keck (LRIS), LCO (FLOYDS,Sinistro),
Spitzer (IRAC), Swope (Direct/4K × DATA AVAILABILITY
The photometry in this article are presented in the article. Imaging andspectroscopy data presented in this article are available upon request.The
Hubble Space Telescope and
Spitzer Space Telescope data are pub-licly available and can be accessed from the Mikulski Archive for SpaceTelescopes ( https://archive.stsci.edu/hst/ ) and Spitzer Her-itage Archive ( http://sha.ipac.caltech.edu/applications/Spitzer/SHA/ ), respectively.
Table 1. SN 2019yvr Photometry
MJD (Epoch) Filter Magnitude (Uncertainty) Source58849.333 (-4.947) 𝐵 𝐵 𝐵 𝐵 𝐵 𝐵 𝐵 𝐵 𝐵 𝐵 𝑉 𝑉 𝑉 𝑉 𝑉 𝑉 𝑉 𝑉 𝑉 𝑉 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑔 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 MNRAS , 1–22 (2020) Kilpatrick et al.
Table 1 (cont’d)
MJD (Epoch) Filter Magnitude (Uncertainty) Source58867.425 (13.145) 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑖 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 𝑢 𝑢 𝑢 𝑢 𝑢 𝑢 𝑢 𝑢 𝑢 All magnitudes were calibrated using Pan-STARRS DR2 photometry stan-dards following procedures described in Kilpatrick et al. (2018a) and Sec-tion 2.3. All 𝑢𝐵𝑉 𝑔𝑟𝑖 magnitudes are on the AB system. All epochs arereported relative to 𝑉 -band maximum. REFERENCES
Aldering G., Humphreys R. M., Richmond M., 1994, AJ, 107, 662Allison J. R., Sadler E. M., Meekin A. M., 2014, MNRAS, 440, 696Amanullah R., et al., 2015, MNRAS, 453, 3300Arnett W. D., Meakin C., Viallet M., 2014, AIP Advances, 4, 041010Bastian N., Gieles M., Lamers H. J. G. L. M., Scheepmaker R. A., de GrijsR., 2005, A&A, 431, 905Beasor E. R., Davies B., Smith N., van Loon J. T., Gehrz R. D., Figer D. F.,2020, MNRAS, 492, 5994Benetti S., et al., 2011, MNRAS, 411, 2726Bersten M. C., et al., 2014, AJ, 148, 68 Bertin E., 2010, SWarp: Resampling and Co-adding FITS Images Together(ascl:1010.068)Bertin E., Arnouts S., 1996, A&AS, 117, 393Bietenholz M. F., De Colle F., Granot J., Bartel N., Soderberg A. M., 2014,MNRAS, 440, 821Burrows A., Hayes J., Fryxell B. A., 1995, ApJ, 450, 830Burrows A., Dessart L., Livne E., Ott C. D., Murphy J., 2007, ApJ, 664, 416Cao Y., et al., 2013, ApJ, 775, L7Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245Chandra P., 2018, Space Sci. Rev., 214, 27Choi J., Dotter A., Conroy C., Cantiello M., Paxton B., Johnson B. D., 2016,ApJ, 823, 102Choi J., Conroy C., Byler N., 2017, ApJ, 838, 159Chugai N. N., Chevalier R. A., 2006, ApJ, 641, 1051Claeys J. S. W., de Mink S. E., Pols O. R., Eldridge J. J., Baes M., 2011,A&A, 528, A131Crockett R. M., et al., 2008, MNRAS, 391, L5Crowther P. A., 2007, ARA&A, 45, 177Delgado A. J., Thomas H. C., 1981, A&A, 96, 142Deng J. S., Qiu Y. L., Hu J. Y., Hatano K., Branch D., 2000, ApJ, 540, 452Dessart L., Livne E., Waldman R., 2010, MNRAS, 405, 2113Dessart L., Hillier D. J., Livne E., Yoon S.-C., Woosley S., Waldman R.,Langer N., 2011, MNRAS, 414, 2985Dessart L., Hillier D. J., Li C., Woosley S., 2012, MNRAS, 424, 2139Dessart L., Hillier D. J., Woosley S., Livne E., Waldman R., Yoon S.-C.,Langer N., 2015, MNRAS, 453, 2189Dimitriadis G., Foley R. J., Siebert M. R., Kilpatrick C. D., Corbett H. T.,2019, The Astronomer’s Telegram, 13375, 1Divine N., 1965, ApJ, 142, 824Dolphin A., 2016, DOLPHOT: Stellar photometry (ascl:1608.013)Drake A. J., et al., 2009, ApJ, 696, 870Dressel L., 2012, Wide Field Camera 3 Instrument Handbook for Cycle 21 v.5.0. STScIDrilling J. S., Schonberner D., 1982, A&A, 113, L22Drilling J. S., Landolt A. U., Schonberner D., 1984, ApJ, 279, 748Drout M. R., et al., 2011, ApJ, 741, 97Dudley R. E., 1992, PhD thesis, University of St. Andrews (United Kingdom)Eldridge J. J., Maund J. R., 2016, MNRAS, 461, L117Eldridge J. J., Fraser M., Smartt S. J., Maund J. R., Crockett R. M., 2013,MNRAS, 436, 774Eldridge J. J., Stanway E. R., Xiao L., McClelland L. A. S., Taylor G., Ng M.,Greis S. M. L., Bray J. C., 2017, Publ. Astron. Soc. Australia, 34, e058Elias-Rosa N., et al., 2016, MNRAS, 463, 3894Filippenko A. V., 1997, ARA&A, 35, 309Filippenko A. V., Matheson T., Ho L. C., 1993, ApJ, 415, L103Flewelling H. A., et al., 2020, ApJS, 251, 7Folatelli G., et al., 2016, ApJ, 825, L22Foley R. J., Smith N., Ganeshalingam M., Li W., Chornock R., FilippenkoA. V., 2007, ApJ, 657, L105Foley R. J., Berger E., Fox O., Levesque E. M., Challis P. J., Ivans I. I., RhoadsJ. E., Soderberg A. M., 2011, ApJ, 732, 32Foley R. J., et al., 2016, MNRAS, 461, 1308Fox O. D., Chevalier R. A., Dwek E., Skrutskie M. F., Sugerman B. E. K.,Leisenring J. M., 2010, ApJ, 725, 1768Fox O. D., et al., 2011, ApJ, 741, 7Fox O. D., et al., 2014, ApJ, 790, 17Frame D. J., Cottrell P. L., Gilmore A. C., Kilmartin P. M., Lawson W. A.,1995, MNRAS, 276, 383Fryer C. L., et al., 2007, PASP, 119, 1211Fuller J., Ro S., 2018, MNRAS, 476, 1853Gal-Yam A., et al., 2014, Nature, 509, 471Galbany L., et al., 2016a, MNRAS, 455, 4087Galbany L., et al., 2016b, A&A, 591, A48Gallagher J. S., 1992, in Drissen L., Leitherer C., Nota A., eds, Astronom-ical Society of the Pacific Conference Series Vol. 22, Nonisotropic andVariable Outflows from Stars. p. 388Georgy C., 2012, A&A, 538, L8Gieles M., Portegies Zwart S. F., 2011, MNRAS, 410, L6MNRAS , 1–22 (2020) rogenitor of SN 2019yvr Gieles M., Larsen S. S., Bastian N., Stein I. T., 2006, A&A, 450, 129Götberg Y., de Mink S. E., Groh J. H., 2017, A&A, 608, A11Götberg Y., de Mink S. E., Groh J. H., Leitherer C., Norman C., 2019, A&A,629, A134Graur O., Bianco F. B., Huang S., Modjaz M., Shivvers I., Filippenko A. V.,Li W., Eldridge J. J., 2017a, ApJ, 837, 120Graur O., Bianco F. B., Modjaz M., Shivvers I., Filippenko A. V., Li W.,Smith N., 2017b, ApJ, 837, 121Graur O., Zurek D. R., Cara M., Rest A., Seitenzahl I. R., Shappee B. J.,Shara M. M., Riess A. G., 2018, ApJ, 866, 10Groh J. H., Damineli A., Hillier D. J., 2008, in de Koter A., Smith L. J., WatersL. B. F. M., eds, Astronomical Society of the Pacific Conference SeriesVol. 388, Mass Loss from Stars and the Evolution of Stellar Clusters.p. 173Guillochon J., Parrent J., Kelley L. Z., Margutti R., 2017, ApJ, 835, 64Habets G. M. H. J., 1986, A&A, 167, 61Hachinger S., Mazzali P. A., Taubenberger S., Hillebrandt W., Nomoto K.,Sauer D. N., 2012, MNRAS, 422, 70Heger A., Jeannin L., Langer N., Baraffe I., 1997, A&A, 327, 224Heger A., Fryer C. L., Woosley S. E., Langer N., Hartmann D. H., 2003, ApJ,591, 288Hora J. L., et al., 2012, in Proc. SPIE. p. 844239, doi:10.1117/12.926894Hosseinzadeh G., et al., 2017, ApJ, 836, 158Humphreys R. M., Davidson K., 1994, PASP, 106, 1025Humphreys R. M., Davidson K., Smith N., 2002, AJ, 124, 1026Ivanova N., et al., 2013, A&ARv, 21, 59Ivezić Ž., et al., 2019, ApJ, 873, 111Izzard R. G., Ramirez-Ruiz E., Tout C. A., 2004, MNRAS, 348, 1215Jacobson-Galán W. V., et al., 2020, arXiv e-prints, p. arXiv:2005.01782James S., Baron E., 2010, ApJ, 718, 957Jones D. O., et al., 2019, Transient Name Server AstroNote, 148, 1Jones D. O., et al., 2020, ApJ, in prep.Kilpatrick C. D., Foley R. J., 2018, MNRAS, 481, 2536Kilpatrick C. D., et al., 2017, MNRAS, 465, 4650Kilpatrick C. D., et al., 2018a, MNRAS, 473, 4805Kilpatrick C. D., et al., 2018b, MNRAS, 480, 2072Kiminki D. C., Kobulnicky H. A., 2012, ApJ, 751, 4Kippenhahn R., Weigert A., 1967, Z. Astrophys., 65, 251Kleiser I., Fuller J., Kasen D., 2018, MNRAS, 481, L141Kochanek C. S., Khan R., Dai X., 2012, ApJ, 759, 20Kochanek C. S., et al., 2017, PASP, 129, 104502Langer N., Hamann W. R., Lennon M., Najarro F., Pauldrach A. W. A., PulsJ., 1994, A&A, 290, 819Laplace E., Götberg Y., de Mink S. E., Justham S., Farmer R., 2020, A&A,637, A6Li W., et al., 2011, MNRAS, 412, 1441Lobel A., de Jager C., Nieuwenhuijzen H., van Genderen A. M., Oudmaijer R.,2015, in EAS Publications Series. pp 279–280, doi:10.1051/eas/1571062Loeb A., Ulmer A., 1997, ApJ, 489, 573Maeder A., Meynet G., 2000, A&A, 361, 159Margutti R., et al., 2014, ApJ, 780, 21Margutti R., et al., 2017, ApJ, 835, 140Mauerhan J. C., et al., 2013, MNRAS, 430, 1801Mauerhan J. C., Filippenko A. V., Zheng W., Brink T. G., Graham M. L.,Shivvers I., Clubb K. I., 2018, MNRAS, 478, 5050Maund J. R., Ramirez-Ruiz E., 2016, MNRAS, 456, 3175Maund J. R., et al., 2006, MNRAS, 369, 390Maund J. R., et al., 2011, ApJ, 739, L37Maund J. R., Pastorello A., Mattila S., Itagaki K., Boles T., 2016, ApJ, 833,128McCully C., Volgenau N. H., Harbeck D.-R., Lister T. A., Saunders E. S.,Turner M. L., Siiverd R. J., Bowman M., 2018, in Guzman J. C., Ibsen J.,eds, Society of Photo-Optical Instrumentation Engineers (SPIE) Confer-ence Series Vol. 10707, Software and Cyberinfrastructure for AstronomyV. p. 107070K ( arXiv:1811.04163 ), doi:10.1117/12.2314340McGregor P., et al., 2004, in Moorwood A. F. M., Iye M., eds, Society ofPhoto-Optical Instrumentation Engineers (SPIE) Conference Series Vol.5492, Proc. SPIE. pp 1033–1044, doi:10.1117/12.550288 Meakin C. A., Arnett D., 2006, ApJ, 637, L53Milisavljevic D., et al., 2015, ApJ, 815, 120Murguia-Berthier A., Batta A., Janiuk A., Ramirez-Ruiz E., Mandel I., NobleS. C., Everson R. W., 2020, arXiv e-prints, p. arXiv:2005.10212Neichel B., Lu J. R., Rigaut F., Ammons S. M., Carrasco E. R., Lassalle E.,2014, MNRAS, 445, 500Neugent K. F., Levesque E. M., Massey P., 2018, AJ, 156, 225Neugent K. F., Levesque E. M., Massey P., Morrell N. I., Drout M. R., 2020,ApJ, 900, 118Nomoto K., Suzuki T., Shigeyama T., Kumagai S., Yamaoka H., Saio H.,1993, Nature, 364, 507Ofek E. O., et al., 2016, ApJ, 824, 6Oliveira A. S., Steiner J. E., Cieslinski D., 2003, MNRAS, 346, 963Onken C. A., et al., 2020, arXiv e-prints, p. arXiv:2008.10359Pan Y. C., Foley R. J., Jones D. O., Filippenko A. V., Kuin N. P. M., 2020,MNRAS, 491, 5897Pastorello A., et al., 2008, MNRAS, 389, 131Pastorello A., et al., 2010, MNRAS, 408, 181Percy J. R., Zsoldos E., 1992, A&A, 263, 123Pereira C. B., Landaberry S. J. C., da Conceição F., 1998, AJ, 116, 1971Phillips M. M., et al., 2013, ApJ, 779, 38Pickles A., Depagne É., 2010, PASP, 122, 1437Podsiadlowski P., 1993, Space Sci. Rev., 66, 439Podsiadlowski P., Joss P. C., Hsu J. J. L., 1992, ApJ, 391, 246Pooley D., et al., 2019, ApJ, 883, 120Poznanski D., Prochaska J. X., Bloom J. S., 2012, MNRAS, 426, 1465Ramirez-Ruiz E., García-Segura G., Salmonson J. D., Pérez-Rendón B., 2005,ApJ, 631, 435Rest A., et al., 2005, ApJ, 634, 1103Richmond M. W., Treffers R. R., Filippenko A. V., Paik Y., Leibundgut B.,Schulman E., Cox C. V., 1994, AJ, 107, 1022Rigaut F., et al., 2014, MNRAS, 437, 2361Rochowicz K., Niedzielski A., 1995, Acta Astron., 45, 307Sana H., et al., 2012, Science, 337, 444Schechter P. L., Mateo M., Saha A., 1993, PASP, 105, 1342Schlafly E. F., Finkbeiner D. P., 2011, ApJ, 737, 103Schneider F. R. N., Izzard R. G., Langer N., de Mink S. E., 2015, ApJ, 805,20Schoenberner D., Drilling J. S., 1984, ApJ, 276, 229Scolnic D., et al., 2015, ApJ, 815, 117Shappee B. J., et al., 2014, ApJ, 788, 48Shappee B. J., et al., 2016, ApJ, 826, 144Shivvers I., et al., 2017a, PASP, 129, 054201Shivvers I., et al., 2017b, MNRAS, 471, 4381Siebert M. R., Dimitriadis G., Polin A., Foley R. J., 2020, ApJ, 900, L27Smartt S. J., et al., 2015, A&A, 579, A40Smith N., 2014, ARA&A, 52, 487Smith N., 2017, Interacting Supernovae: Types IIn and Ibn. Cham, Switzer-land : Springer, p. 403, doi:10.1007/978-3-319-21846-5_38Smith N., Arnett W. D., 2014, ApJ, 785, 82Smith N., Conti P. S., 2008, ApJ, 679, 1467Smith N., McCray R., 2007, ApJ, 671, L17Smith N., Vink J. S., de Koter A., 2004, ApJ, 615, 475Smith N., et al., 2009, ApJ, 697, L49Smith N., Li W., Filippenko A. V., Chornock R., 2011, MNRAS, 412, 1522Smith N., et al., 2017, MNRAS, 466, 3021Smith K. W., et al., 2019, Transient Name Server AstroNote, 161, 1Sollerman J., et al., 2020, A&A, 643, A79Srivastav S., Anupama G. C., Sahu D. K., 2014, MNRAS, 445, 1932Stothers R. B., Chin C.-w., 1999, ApJ, 522, 960Stothers R. B., Chin C.-w., 2001, ApJ, 560, 934Stritzinger M., et al., 2009, ApJ, 696, 713Stritzinger M. D., et al., 2018, A&A, 609, A135Sukhbold T., Ertl T., Woosley S. E., Brown J. M., Janka H.-T., 2016, ApJ,821, 38Sun N.-C., Maund J. R., Crowther P. A., 2020, MNRAS, 497, 5118Tartaglia L., et al., 2020, arXiv e-prints, p. arXiv:2009.03331Thöne C. C., et al., 2017, A&A, 599, A129 MNRAS , 1–22 (2020) Kilpatrick et al.
Tinyanont S., et al., 2016, ApJ, 833, 231Tinyanont S., et al., 2019, ApJ, 887, 75Valenti S., et al., 2011, MNRAS, 416, 3138Van Dyk S. D., Peng C. Y., King J. Y., Filippenko A. V., Treffers R. R., LiW., Richmond M. W., 2000, PASP, 112, 1532Van Dyk S. D., et al., 2014, AJ, 147, 37Van Dyk S. D., et al., 2018, ApJ, 860, 90Vanbeveren D., de Grève J. P., van Dessel E. L., de Loore C., 1979, A&A,73, 19Woolf N. J., 1973, ApJ, 185, 229Woosley S. E., Eastman R. G., Weaver T. A., Pinto P. A., 1994, ApJ, 429, 300Wu S., Everson R. W., Schneider F. R. N., Podsiadlowski P., Ramirez-RuizE., 2020, ApJ, 901, 44Yan L., et al., 2017, ApJ, 848, 6Yoon S.-C., 2015, Publ. Astron. Soc. Australia, 32, e015Yoon S.-C., 2017, MNRAS, 470, 3970Yoon S.-C., Cantiello M., 2010, ApJ, 717, L62Yoon S.-C., Woosley S. E., Langer N., 2010, ApJ, 725, 940Yoon S.-C., Gräfener G., Vink J. S., Kozyreva A., Izzard R. G., 2012, A&A,544, L11Yoon S.-C., Dessart L., Clocchiatti A., 2017, ApJ, 840, 10de Jager C., 1998, A&ARv, 8, 145de Jager C., Nieuwenhuijzen H., 1997, MNRAS, 290, L50de Koter A., Lamers H. J. G. L. M., Schmutz W., 1996, A&A, 306, 501van der Hucht K. A., 2001, New Astron. Rev., 45, 135MNRAS000