diff --git "a/4tE3T4oBgHgl3EQfowri/content/tmp_files/2301.04637v1.pdf.txt" "b/4tE3T4oBgHgl3EQfowri/content/tmp_files/2301.04637v1.pdf.txt"
new file mode 100644--- /dev/null
+++ "b/4tE3T4oBgHgl3EQfowri/content/tmp_files/2301.04637v1.pdf.txt"
@@ -0,0 +1,3551 @@
+DRAFT VERSION JANUARY 12, 2023
+Typeset using LATEX twocolumn style in AASTeX62
+A Systematic Study of Ia-CSM Supernovae from the ZTF Bright Transient Survey
+YASHVI SHARMA,1 JESPER SOLLERMAN,2 CHRISTOFFER FREMLING,1 SHRINIVAS R. KULKARNI,1 KISHALAY DE,3 IDO IRANI,4
+STEVE SCHULZE,5 NORA LINN STROTJOHANN,4 AVISHAY GAL-YAM,4 KATE MAGUIRE,6 DANIEL A. PERLEY,7 ERIC C. BELLM,8
+ERIK C. KOOL,2 THOMAS BRINK,9 RACHEL BRUCH,4 MAXIME DECKERS,6 RICHARD DEKANY,10 ALISON DUGAS,11
+SAMANTHA GOLDWASSER,4 MATTHEW J. GRAHAM,1 MELISSA L. GRAHAM,8 STEVEN L. GROOM,12 MATT HANKINS,13
+JACOB JENCSON,14 JOEL P. JOHANSSON,5 VIRAJ KARAMBELKAR,1 MANSI M. KASLIWAL,1 FRANK J. MASCI,12
+MICHAEL S. MEDFORD,15, 16 JAMES D. NEILL,1 GUY NIR,9 REED L. RIDDLE,10 MICKAEL RIGAULT,17 TASSILO SCHWEYER,2
+JACCO H. TERWEL,6, 18 LIN YAN,1 YI YANG (杨轶) ,9 AND YUHAN YAO1
+1Division of Physics, Mathematics and Astronomy, California Institute of Technology, Pasadena, CA 91125, USA
+2Department of Astronomy, The Oskar Klein Center, Stockholm University, AlbaNova, 10691 Stockholm, Sweden
+3MIT-Kavli Institute for Astrophysics and Space Research, Cambridge, MA 02139, USA
+4Department of Particle Physics and Astrophysics, Weizmann Institute of Science, 234 Herzl St, 76100 Rehovot, Israel
+5Department of Physics, The Oskar Klein Center, Stockholm University, AlbaNova, 10691 Stockholm, Sweden
+6School of Physics, Trinity College Dublin, the University of Dublin, College Green, Dublin, Ireland
+7Astrophysics Research Institute, Liverpool John Moores University, Liverpool Science Park, 146 Brownlow Hill, Liverpool L35RF, UK
+8DIRAC Institute, Department of Astronomy, University of Washington, 3910 15th Avenue NE, Seattle, WA 98195, USA
+9Department of Astronomy, University of California, Berkeley, CA 94720-3411, USA
+10Caltech Optical Observatories, California Institute of Technology, Pasadena, CA 91125, USA
+11Institute for Astronomy, University of Hawai’i, Honolulu, HI 96822, USA
+12IPAC, California Institute of Technology, 1200 E. California Blvd, Pasadena, CA 91125, USA
+13Arkansas Tech University, Russellville, AR 72801, USA
+14Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721-0065, USA
+15Department of Astronomy, University of California, Berkeley, Berkeley, CA 94720
+16Lawrence Berkeley National Laboratory, 1 Cyclotron Rd., Berkeley, CA 94720
+17Universit´e Clermont Auvergne, CNRS/IN2P3, Laboratoire de Physique de Clermont, 63000 Clermont-Ferrand, France
+18Isaac Newton Group (ING), Apt. de correos 321, E-38700, Santa Cruz de La Palma, Canary Islands, Spain
+ABSTRACT
+Among the supernovae (SNe) that show strong interaction with the circumstellar medium, there is a rare
+subclass of Type Ia supernovae, SNe Ia-CSM, that show strong narrow hydrogen emission lines much like SNe
+IIn but on top of a diluted over-luminous Type Ia spectrum. In the only previous systematic study of this class
+(Silverman et al. 2013), 16 objects were identified, 8 historic and 8 from the Palomar Transient Factory (PTF).
+Now using the successor survey to PTF, the Zwicky Transient Facility (ZTF), we have classified 12 additional
+objects of this type through the systematic Bright Transient Survey (BTS). In this study, we present and analyze
+the optical and mid-IR light curves, optical spectra and host galaxy properties of this sample. Consistent with
+previous studies, we find the objects to have slowly evolving light curves compared to normal SNe Ia with peak
+absolute magnitudes between −19.1 and −21, spectra having weak Hβ, large Balmer decrements of ∼ 7 and
+strong Ca NIR emission. Out of 10 SNe from our sample observed by NEOWISE, 9 have 3σ detections, along
+with some showing a clear reduction in red-wing of Hα, indicative of newly formed dust. We do not find our
+SN Ia-CSM sample to have significantly different distribution of equivalent width of He I λ5876 than SNe IIn
+as observed in Silverman et al. (2013). The hosts tend to be late-type galaxies with recent star formation. We
+also derive a rate estimate of 29+27
+−21 Gpc−3 yr−1 for SNe Ia-CSM which is ∼0.02–0.2% of the SN Ia rate. This
+work nearly doubles the sample of well studied Ia-CSM objects in Silverman et al. (2013), increasing the total
+number to 28.
+Corresponding author: Yashvi Sharma
+yssharma@astro.caltech.edu
+arXiv:2301.04637v1  [astro-ph.HE]  11 Jan 2023
+
+2
+Keywords: circumstellar matter – supernovae: general – supernovae: individual (SN 1997cy, SN 2002ic, SN
+2005gj, SN 2005ip, SN 2006jc, SN 2008J, SN 2009ip, SN 2010jl, PTF11kx, SN 2012ca, SN 2013dn,
+SN 2018crl, SN 2018gkx, SN 2018evt, SN 2019agi, SN 2019ibk, SN 2019rvb, SN 2020onv, SN
+2020qxz, SN 2020uem, SN 2020xtg, SN 2020abfe, SN 2020aekp)
+1. INTRODUCTION
+When it comes to supernovae (SNe) interacting with cir-
+cumstellar material (CSM), a number of sub-types of core-
+collapse SNe (CCSNe) show signs of strong interaction, like
+SNe IIn (Schlegel 1990; Filippenko 1997), SNe Ibn (Pas-
+torello et al. 2008; Foley et al. 2007; Chugai 2009; Hos-
+seinzadeh et al. 2017) and most recently SNe Icn (Gal-Yam
+et al. 2021, 2022; Perley et al. 2022). SN IIn progenitors are
+generally thought to be massive stars (like Luminous Blue
+Variables, LBVs) that lose their hydrogen envelopes to wind-
+driven mass loss and outbursts (Gal-Yam et al. 2007; Gal-
+Yam & Leonard 2009; Kiewe et al. 2012; Taddia et al. 2013;
+Smith 2014). Helium-rich but hydrogen-deficient CSM in
+the case of SNe Ibn (Pastorello et al. 2008; Foley et al. 2007;
+Chugai 2009) and both hydrogen and helium deficient CSM
+in SNe Icn (Gal-Yam et al. 2022; Perley et al. 2022; Pel-
+legrino et al. 2022) are thought to arise from high-velocity
+wind mass loss or stripping of the envelope in binary con-
+figurations of massive Wolf-Rayet (WR) like stars. For SNe
+IIn in most cases, the mass-loss rate derived from the CSM
+velocity is consistent with estimates from LBV-like eruptive
+mass loss.
+However, there exists a rare sub-type of thermonuclear su-
+pernovae (SNe Ia) which also interacts strongly with CSM
+i.e. SNe Ia-CSM. This class poses a challenge to the progen-
+itor debate of SNe Ia. There is some consensus on there being
+at least two major progenitor channels for SNe Ia; the double-
+degenerate (DD) channel (Webbink 1984; Iben & Tutukov
+1984) which is the merging of two C/O white dwarfs and
+the single-degenerate (SD) channel (Whelan & Iben 1973)
+where the white dwarf accretes enough material from a non-
+degenerate companion to explode. Although there are more
+arguments for the DD scenario from observations of nearby
+SNe Ia (Nugent et al. 2011; Li et al. 2011; Brown et al. 2012;
+Bloom et al. 2011), the strongest observational evidence for
+the SD scenario are SNe Ia with CSM.
+Indications of CSM around SNe Ia ranges from detec-
+tion of time varying narrow Na ID absorption lines (Patat
+et al. 2007; Blondin et al. 2009; Simon et al. 2009) in high-
+resolution spectra (found in at least 20% of SNe Ia in spiral
+hosts, Sternberg et al. 2011; Maguire et al. 2013; Clark et al.
+2021), to strong intermediate and narrow Balmer emission
+features in the spectra and large deviations of the light curves
+from the standard shape. The latter phenomena have been
+named SNe Ia-CSM (Silverman et al. 2013), but were ear-
+lier referred to as “SNe IIna” or “SNe Ian” due to the strong
+similarity between their spectra and those of SNe IIn. The
+first two examples of this class studied in detail were SNe
+2002ic (Hamuy et al. 2003; Deng et al. 2004; Wang et al.
+2004; Wood-Vasey et al. 2004; Kotak & Meikle 2005; Chugai
+et al. 2004) and 2005gj (Aldering et al. 2006; Prieto et al.
+2007), but for a long time there was ambiguity regarding their
+thermonuclear nature (Benetti et al. 2006). These SNe were
+dominated by interaction from the first spectrum and were
+quite over-luminous compared to normal SNe Ia. The first
+clear example of a thermonuclear SN Ia-CSM was PTF11kx
+(Dilday et al. 2012; Silverman et al. 2013). It looked like a
+luminous SN Ia (99aa-like) at early phases but started show-
+ing interaction at ∼ 60 days from explosion and thereafter
+strongly resembled SNe 2002ic and 2005gj at late times.
+Higher resolution spectra taken at early times indicated mul-
+tiple shells of CSM with some evacuated regions in between.
+Dilday et al. (2012) suggested a symbiotic nova progenitor
+involving a WD and a red giant (similar to RS Ophiuchi)
+could produce such CSM distribution, however later studies
+argued that the massive CSM of PTF11kx was inconsistent
+with the mass-loss rates from symbiotic nova systems (Sil-
+verman et al. 2013; Soker et al. 2013).
+Ever since, a handful of SNe of this class have been stud-
+ied in detail to investigate their progenitors and to distinguish
+them from their spectroscopic cousins, the Type IIn SNe.
+Both SN Ia-CSM and SN IIn spectra share a blue quasi-
+continuum, a strong Hα feature with an intermediate and a
+narrow component, and often a broad Ca NIR triplet fea-
+ture, but they differ with regards to the line strength of Hβ,
+strength/presence of helium and presence of emission lines
+from intermediate mass elements often found in CCSNe.
+There are some individual SNe with unclear type often re-
+ferred to as SN Ia-CSM/IIn, like SN 2012ca for which some
+papers argue for core-collapse (Inserra et al. 2014, 2016) and
+others for a thermonuclear origin (Fox et al. 2015). This am-
+biguity becomes more dominant as the underlying SN flux
+gets smaller compared to the interaction power (Leloudas
+et al. 2015). Silverman et al. (2013, hereafter S13) is the
+only study to analyze a sample of SNe Ia-CSM, 16 objects
+in total including 6 previously known, 3 re-discovered (re-
+classified SNe IIn) and 7 new from the Palomar Transient
+Factory (PTF). Their paper presents the common properties
+of optical light curves, spectra and host galaxies and contrast
+them against SN IIn properties. In this paper, we present
+12 new SNe Ia-CSM discovered as part of the Zwicky Tran-
+sient Facility’s (ZTF; Bellm et al. 2019; Graham et al. 2019;
+
+3
+Dekany et al. 2020) Bright Transient Survey (BTS; Fremling
+et al. 2020; Perley et al. 2020) and analyze their optical light
+curves, spectra, hosts and rates. Throughout this paper, we
+have compared the results derived from our sample to the
+ones in S13.
+This paper is organised as follows; we first discuss the sam-
+ple selection criteria, the photometric and spectroscopic data
+collection in §2, then the analysis of light- and color-curves
+and the bolometric luminosities is done in §3.1. The analysis
+of early and late-time spectra and emission line identification
+is presented in §3.2, and analysis of the host galaxies is pro-
+vided in §3.3. The rates are estimated from the BTS survey
+in §3.4. We end with a discussion about the nature of SN
+Ia-CSM progenitors and a summary in §4 and §5.
+2. OBSERVATIONS AND DATA REDUCTION
+In this section, we outline our selection criteria, and
+present the optical photometry and spectroscopic observa-
+tions of the 12 SNe Ia-CSM in our sample.
+2.1. Selection Criteria
+To carefully curate our sample of SNe Ia-CSM, we used
+the BTS sample and its publicly available BTS Sample Ex-
+plorer1 website to obtain the list of all classified Type Ia sub-
+types during the period 2018-05-01 to 2021-05-01. We then
+filter out oddly behaving Type Ia SNe based on their light-
+curve properties. We used two criteria; the primary being
+rest-frame duration considering flux above 20% of peak flux,
+and the second being change in magnitude after 30 days from
+peak (∆m30). We calculated these two properties from either
+g or r-band light curves (whichever had maximum number of
+detections) grouped into 3-day bins and used Gaussian Pro-
+cess Regression2 to interpolate the light curves where cov-
+erage was missing. For the first filtering, we calculated the
+mean (µ ≈ 35 days) and standard deviation (σ ≈ 16 days)
+of the duration distribution and selected everything that had
+a duration greater than µ + 3σ. Given the large sample size
+(N = 3486), the standard error on the mean is ∼ 0.5 days,
+hence our duration cut of 3σ is suitable. This filtering se-
+lected 41 out of 3486 BTS SNe Ia.
+Then from these 41
+SNe, we calculated the mean and standard deviation of the
+∆m30 distribution and removed SNe that were more than 1σ
+away from the mean on the higher side to reject the relatively
+steeply declining long SNe, which resulted in 35 SNe being
+kept. Again, the mean and standard deviation of ∆m30 dis-
+tribution of these 41 long duration SNe are 0.48 mag and 0.27
+mag respectively and the standard error on mean is ∼ 0.04,
+making our 1σ cut suitable. Finally, we manually inspected
+1 https://sites.astro.caltech.edu/ztf/bts/explorer.php
+2 Pedregosa et al. (2011) https://scikit-learn.org/stable/modules/gaussian
+process.html
+the 35 selected SNe Ia to confirm their classification. 20 out
+of the 35 SNe that passed the above filtering criteria were
+just normal SNe Ia either caught late or missing some post-
+peak coverage in ZTF or had spurious detections that resulted
+in long duration estimates, 2 had incorrect duration estimate
+due to an interpolation error and were recalculated and 1
+(AT2020caa; Soraisam et al. 2021) had some detections be-
+fore the SN explosion which could be connected to a different
+SN (i.e. a sibling; Graham et al. 2022).
+The remaining 12 long-duration SNe Ia all turned out to be
+spectroscopically classified SNe Ia-CSM in BTS, and none
+of the classified BTS SNe Ia-CSM were missed in this fil-
+tering. No other SNe apart from these stood out in particu-
+lar, indicating the classification reliability of the BTS sample.
+During the same period, 9 SNe Ia-CSM were reported to the
+Transient Name Server (TNS), out of which 7 are already in
+our sample, 1 was detected by ZTF but did not meet the BTS
+criteria, and 1 was not detected in ZTF as the transient lo-
+cation fell too close to the field edges and was masked by
+the automated image subtraction pipeline. Yao et al. (2019)
+presented early photometric observations of one SN Ia-CSM
+in our sample, SN 2018crl. Table 1 summarizes the coor-
+dinates, redshifts, peak absolute magnitudes, durations, host
+galaxy information and Milky Way extinction for the 12 SNe
+Ia-CSM in our sample.
+Furthermore, we re-checked the classifications of 142 SNe
+IIn classified in BTS during the same period as above, in case
+any SN Ia-CSM was masquerading among them and found 6
+to have ambiguous classifications. These are discussed fur-
+ther in Appendix A.
+2.2. Discovery
+All SNe Ia-CSM were detected by the ZTF (Bellm et al.
+2019; Graham et al. 2019; Dekany et al. 2020) and passed
+the criteria for the BTS (Fremling et al. 2020; Perley et al.
+2020) automatic filtering, i.e. extra-galactic real transients
+with peak magnitudes brighter than 19 mag.
+These were
+saved and classified as part of BTS which aims to classify all
+transients brighter than 18.5 magnitude, and reported to the
+Transient Name Server3 (TNS) during the period 2018-05-
+01 to 2021-05-01. Out of the 12 SNe, 6 were first reported to
+TNS (i.e. discovered) by ZTF (AMPEL, Nordin et al. 2019;
+Soumagnac & Ofek 2018 and BTS), 3 were first reported by
+GaiaAlerts (Hodgkin et al. 2021), 2 by ATLAS (Smith et al.
+2020) and 1 by ASAS-SN (Shappee et al. 2014). For classi-
+fication, 9 were classified by the ZTF group, 1 by ePESSTO
+(Smartt et al. 2015; Stein et al. 2018a), 1 by SCAT (Tucker
+et al. 2018; Payne et al. 2019) and 1 by the Trinity College
+Dublin (TCD) group (Prentice et al. 2020). The follow-up
+spectral series for these SNe were obtained as part of the
+3 https://www.wis-tns.org/
+
+4
+Table 1. Properties of the 12 BTS SNe Ia-CSM
+ZTF Name
+IAU Name
+z
+M peak
+r
+Duration1
+Host Name
+Host Mag2
+(mag)
+(days)
+(mr)
+ZTF18aaykjei
+SN 2018crl
+0.097
+-19.66
+130
+SDSS J161938.90+491104.5
+18.89
+ZTF18abuatfp
+SN 2018gkx
+0.1366
+-20.07
+322
+SDSS J135219.22+553830.2
+18.23
+ZTF18actuhrs
+SN 2018evt
+0.02378
+-19.10
+447
+MCG-01-35-011
+14.07
+ZTF19aaeoqst
+SN 2019agi
+0.0594
+<-18.76
+>303
+SDSS J162244.06+240113.4
+17.82
+ZTF19abidbqp
+SN 2019ibk
+0.04016
+<-17.55
+>576
+SDSS J014611.93-161701.1
+15.55
+ZTF19acbjddp
+SN 2019rvb
+0.1835
+-20.74
+172
+WISEA J163809.90+682746.3
+20.44
+ZTF20abmlxrx
+SN 2020onv
+0.095
+<-20.36
+>154
+WISEA J231646.31-231839.9
+17.95
+ZTF20abqkbfx
+SN 2020qxz
+0.0964
+-20.00
+166
+WISEA J180400.99+740050.0
+17.65
+ZTF20accmutv
+SN 2020uem
+0.041
+<-20.17
+>279
+WISEA J082423.32-032918.6
+15.88
+ZTF20aciwcuz
+SN 2020xtg
+0.0612
+<-19.60
+>336
+SDSS J153317.64+450022.8
+15.42
+ZTF20acqikeh
+SN 2020abfe
+0.093
+-20.24
+171
+SDSS J200003.30+100904.2
+20.18
+ZTF21aaabwzx
+SN 2020aekp
+0.046
+-19.62
+458
+SDSS J154311.45+174843.7
+18.41
+1 Rest frame duration above 20% of r-band peak flux, uncertainty of ±2 − 3 days from ZTF cadence.
+2 Corrected for Galactic extinction.
+BTS classification campaign as many were difficult to clas-
+sify with the ultra-low resolution spectrograph P60/SEDM
+(Blagorodnova et al. 2018) and hence were followed up with
+intermediate resolution spectrographs. The SEDM spectra
+were helpful in determining an initial redshift but the tem-
+plate matches were unclear (matched to SN IIn as well as
+SN Ia-CSM and SN Ia-pec templates, some matched poorly
+to SN Ia/Ic at early times). SNe 2019agi (classification and
+spectrum taken from TNS), 2019rvb, 2020onv, 2020qxz and
+2020uem were classified as Ia-CSM ∼ 1 − 2 month after
+discovery using spectra at phases of 42, 26, 38, 45 and 51
+days respectively. SNe 2018crl, 2018gkx and 2019ibk were
+classified ∼ 2 − 3 months after discovery using spectra at
+phases of 92, 75 and 103 days respectively. SNe 2018evt,
+2020abfe and 2020aekp were classified ∼ 4 − 5 months af-
+ter discovery using the spectra at phases of 144, 146 and 132
+days respectively. SN 2020xtg immediately went behind the
+sun after its first detection in ZTF therefore its first spectrum
+(using SEDM) was taken at 91 days since explosion which
+was dominated by strong Hα emission, and thus SN 2020xtg
+was initially classified as a Type II. As this SN was exhibiting
+a long lasting light curve, an intermediate resolution spec-
+trum was taken at 340 days which matched very well to SN
+Ia-CSM and therefore its classification was updated. SNe
+2020uem and 2020aekp showed peculiar features and were
+followed up for more optical spectroscopy for single object
+studies (to be presented in future papers).
+2.3. Optical photometry
+To assemble our sample light curves, we obtained forced
+PSF photometry via the ZTF forced-photometry service
+(Masci et al. 2019; IRSA 2022) in g, r and i bands and
+also added data from ATLAS (Tonry et al. 2018; Smith et al.
+2020) forced-photometry service in c and o bands. The high
+cadence ZTF partnership survey in i band contributed some
+photometry to SNe 2018crl, 2018gkx, 2019agi, 2019ibk and
+2019rvb. The ZTF and ATLAS data were supplemented with
+data from the Rainbow camera (RC, Ben-Ami et al. 2012)
+on the robotic Palomar 60-inch telescope (P60, Cenko et al.
+2006) and the Optical wide field camera (IO:O) on the Liver-
+pool telescope (LT, Steele et al. 2004). The P60 data was pro-
+cessed with the automatic image subtraction pipeline FPipe
+(Fremling et al. 2016) using reference images from SDSS
+when available, and otherwise from Pan-STARRS1.
+The
+IO:O data was initially reduced with their standard pipeline4
+then image subtraction was carried out using the method
+outlined in Taggart (2020).
+For SN 2018evt, some early
+time data available from ASAS-SN (Shappee et al. 2014;
+Kochanek et al. 2017) in the V band was obtained through
+their Sky Patrol5 interface.
+We corrected all photometry for Milky Way extinction
+with the Python package extinction (Barbary 2016) us-
+ing the dust extinction function from Fitzpatrick (1999), the
+Schlafly & Finkbeiner (2011) dust map, and an RV of 3.1.
+Then we converted all measurements into flux units for anal-
+ysis and considered anything less than a 3σ detection an up-
+per limit. There is moderate to good coverage in g, r, c and
+o bands for all SNe in our sample. Figure 1 shows a multi-
+paneled figure of the light curves of the objects in our sample.
+2.4. Mid-IR photometry
+4 https://telescope.livjm.ac.uk/TelInst/Pipelines/
+5 https://asas-sn.osu.edu/
+
+5
+0
+80
+160
+240
+320
+400
+-15.5
+-16.9
+-18.2
+-19.6
+-21.0
+Co decay
+SN 2018crl
+2.11 mag/100 d
+0.36 mag/100 d
+0
+100
+200
+300
+400
+500
+-15.0
+-16.5
+-18.0
+-19.5
+-21.0
+Co decay
+SN 2018gkx
+1.18 mag/100 d
+0.42 mag/100 d
+0
+150
+300
+450
+600
+-14.0
+-15.6
+-17.2
+-18.9
+-20.5
+Co decay
+SN 2018evt
+0.51 mag/100 d
+1.36 mag/100 d
+0
+80
+160
+240
+320
+400
+-15.5
+-16.9
+-18.2
+-19.6
+-21.0
+Co decay
+SN 2019rvb
+0.93 mag/100 d
+0.5 mag/100 d
+0
+100
+200
+300
+400
+500
+-15.0
+-16.5
+-18.0
+-19.5
+-21.0
+Co decay
+SN 2019agi
+0.96 mag/100 d
+0.45 mag/100 d
+0
+150
+300
+450
+600
+-14.0
+-15.6
+-17.2
+-18.9
+-20.5
+Co decay
+SN 2019ibk
+0.61 mag/100 d
+0.18 mag/100 d
+0
+80
+160
+240
+320
+400
+-15.5
+-16.9
+-18.2
+-19.6
+-21.0
+Co decay
+SN 2020qxz
+1.75 mag/100 d
+0.29 mag/100 d
+0
+100
+200
+300
+400
+500
+-15.0
+-16.5
+-18.0
+-19.5
+-21.0
+D
+Co decay
+SN 2020onv
+1.26 mag/100 d
+1.06 mag/100 d
+0
+150
+300
+450
+600
+-14.0
+-15.6
+-17.2
+-18.9
+-20.5
+Co decay
+SN 2020uem*
+0.6 mag/100 d
+1.27 mag/100 d
+0
+80
+160
+240
+320
+400
+D
+-15.5
+-16.9
+-18.2
+-19.6
+-21.0
+Co decay
+SN 2020abfe
+1.51 mag/100 d
+1.51 mag/100 d
+0
+100
+200
+300
+400
+500
+-15.0
+-16.5
+-18.0
+-19.5
+-21.0
+Co decay
+SN 2020xtg
+0.62 mag/100 d
+1.21 mag/100 d
+0
+150
+300
+450
+600
+-14.0
+-15.6
+-17.2
+-18.9
+-20.5
+Co decay
+SN 2020aekp
+1.52 mag/100 d
+0.07 mag/100 d
+0.66 mag/100 d
+Rest-frame days since explosion
+Absolute magnitude
+P48:ZTF: g
+P48:ZTF: r
+P48:ZTF: i
+ATLAS: c
+ATLAS: o
+P60:RC: g
+P60:RC: r
+P60:RC: i
+LT:IOO: g
+LT:IOO: r
+LT:IOO: i
+LT:IOO: z
+LCO: sdssg
+LCO: sdssr
+LCO: sdssi
+ASASSN: V
+Figure 1. Optical light curves of the ZTF BTS SN Ia-CSM sample. The SNe Ia-CSM have longer duration than the average SN Ia, with some
+variety like bumpy light curves or long plateaus. The one SN marked with an asterisk (SN 2020uem) has an unconstrained explosion time
+estimate (∼ ±50 d). The decline rate from Cobalt decay is marked with black dashed line, the light curve decline rates measured from r-band
+data are shown in the subplot legends.
+
+6
+The transients were observed during the ongoing NEO-
+WISE all-sky mid-IR survey in the W1 (3.4 µm) and W2
+(4.5 µm) bands (Wright et al. 2010a; Mainzer et al. 2014).
+We retrieved time-resolved coadded images of the field cre-
+ated as part of the unWISE project (Lang 2014a; Meisner
+et al. 2018). To remove contamination from the host galax-
+ies, we used a custom code (De et al. 2020) based on the
+ZOGY algorithm (Zackay et al. 2016) to perform image sub-
+traction on the NEOWISE images using the full-depth coadds
+of the WISE and NEOWISE mission (obtained during 2010-
+2014) as reference images. Photometric measurements were
+obtained by performing forced PSF photometry at the tran-
+sient position on the subtracted WISE images until the epoch
+of the last NEOWISE data release (data acquired until De-
+cember 2021). Further analysis of the mid-IR photometry is
+presented in §3.1.4
+2.5. Optical spectroscopy
+The main instruments used for taking spectra and the soft-
+ware used to reduce the data are summarized in Table 2. Ad-
+ditionally, the spectrum Reguitti (2020) obtained using the
+Asiago Faint Object Spectrograph and Camera (AFOSC) on
+the 1.8 m telescope at Cima Ekar, and the spectrum Stein
+et al. (2018b) obtained using the ESO Faint Object Spectro-
+graph and Camera version 2 (EFOSC2) on ESO New Tech-
+nology Telescope (NTT) were taken from TNS.
+The details for all optical spectra (61 for the sample in to-
+tal) presented in this paper are provided in Table 3. Further-
+more, all spectra were corrected for Milky Way extinction
+using extinction and the same procedure as for the pho-
+tometry. The SN redshifts were derived using narrow host
+lines for the objects which did not already have a host red-
+shift available in the NASA/IPAC Extragalactic Database6
+(NED). Photometric calibration was done for all spectra i.e.
+they were scaled such that the synthetic photometry from the
+spectrum matched the contemporaneous host-subtracted ZTF
+r-band data. For SN 2018crl, a host galaxy spectrum taken
+using P200/DBSP was available, which was subtracted from
+the P200/DBSP SN spectrum taken at +92 days. For SN
+2020aekp, more spectra beyond ∼ 350 days were obtained
+but will be presented in a future study of the object (34 addi-
+tional spectra up to ∼600 day).
+These processed spectra were used for the rest of the anal-
+ysis as detailed in §3.2 and will be available on WISeREP7
+(Yaron & Gal-Yam 2012).
+3. ANALYSIS
+3.1. Photometry
+6 https://ned.ipac.caltech.edu/
+7 https://www.wiserep.org/
+Table 2. Description of spectrographs used for follow-up and the
+corresponding data reduction pipelines
+Inst.
+Telescope
+Reduction Software
+SEDM1
+Palomar 60-inch (P60)
+pySEDM2
+ALFOSC3
+Nordic Optical Telescope
+IRAF4, PyNOT14, pypeit
+DBSP5
+Palomar 200-inch (P200)
+IRAF6, DBSP DRP7
+KAST8
+Shane 3-m
+IRAF
+LRIS9
+Keck-I
+LPipe10
+SPRAT11
+Liverpool Telescope
+Barnsley et al. (2012)
+DIS12
+APO13
+IRAF
+1 Spectral Energy Distribution Machine (Blagorodnova et al. 2018)
+2 Rigault et al. (2019)
+3 Andalucia Faint Object Spectrograph and Camera
+4 Tody (1986, 1993)
+5 Double Beam Spectrograph (Oke & Gunn 1982)
+6 Standard pipeline by Bellm & Sesar (2016) used prior to Fall
+2020
+7 pypeit (Prochaska et al. 2020) based pipeline (https://github.
+com/finagle29/dbsp drp) used since Fall 2020
+8 Kast Double Spectrograph (Miller & Stone 1987)
+9 Low Resolution Imaging Spectrometer (Oke et al. 1995)
+10 IDL based automatic reduction pipelinea (Perley 2019)
+11 Spectrograph for the Rapid Acquisition of Transients (Piascik
+et al. 2014)
+12 Dual Imaging Spectrograph
+13 Astrophysics Research Consortium telescope at the Apache
+Point Observatory
+14 https://github.com/jkrogager/PyNOT
+ahttps://sites.astro.caltech.edu/∼dperley/programs/lpipe.html
+3.1.1. Explosion epoch estimates
+For the purpose of this paper, the ‘explosion time’ simply
+refers to the time when optical flux rises above the zero-point
+baseline (i.e. first light). We used pre-peak g, r, i-band ZTF
+photometry and c, o-band ATLAS photometry (binned in 1-
+day bins), when available, for our analysis. For each SN, the
+light curve was interpolated using Gaussian process regres-
+sion to obtain the peak flux epoch, then a power-law (PL)
+model was fit using epochs from baseline to 60% of peak
+brightness in each band following Miller et al. (2020). The
+PL fits converged in at least one band for 6 out of 12 BTS
+SNe Ia-CSM. For the rest, we simply took the middle point
+between the first 5σ detection and the last upper limit before
+this detection as the explosion epoch with half of the separa-
+tion between these two points as the uncertainty.
+The explosion time estimates, light curve bands used for
+the PL fits and the 1σ uncertainties on explosion times are
+listed in Table 4. The unfilled ‘PL fit filters’ column in the
+table are the SNe for which the PL fit did not converge and
+averages were used. For the PL fits this typically constrains
+the time of explosion to within a fraction of a day. Given the
+high cadence of the ZTF survey, even in the cases where we
+
+7
+Table 3. Summary of optical spectra
+SN
+JD
+Epoch
+Telescope/Instrument
+Int
+SN
+JD
+Epoch
+Tel./Instr.
+Int
+(−2450000)
+(days)
+(sec)
+(−2450000)
+(days)
+(sec)
+SN 2018crl
+8282
+9
+APO/DIS
+2400
+SN 2020uem
+9128
+11
+P60/SEDM
+1800
+8288
+15
+P60/SEDM
+2700
+9136
+18
+P60/SEDM
+1800
+8295
+21
+P60/SEDM
+2700
+9170
+51
+Ekar/AFOSC
+1200
+8306
+31
+P60/SEDM
+2700
+9222
+101
+Lick-3m/KAST
+3600
+8373
+92
+P200/DBSP
+600
+9252
+130
+Lick-3m/KAST
+2700
+(Host)
+8627
+324
+P200/DBSP
+900
+9263
+140
+Lick-3m/KAST
+2400
+SN 2018gkx
+8457
+75
+Keck1/LRIS
+300
+9291
+167
+NOT/ALFOSC
+900
+SN 2018evt
+8343
+9
+NTT/EFOSC2
+300
+9481
+349
+P60/SEDM
+2160
+8465
+127
+P60/SEDM
+1200
+9492
+360
+Keck1/LRIS
+600
+8481
+143
+P60/SEDM
+1200
+9583
+448
+P60/SEDM
+2160
+8481
+144
+LT/SPRAT
+1000
+9586
+451
+P60/SEDM
+2160
+8534
+195
+P60/SEDM
+1200
+SN 2020xtg
+9226
+91
+P60/SEDM
+2160
+SN 2019agi
+8547
+42
+UH88/SNIFS
+1820
+9491
+340
+Keck1/LRIS
+600
+SN 2019ibk
+8691
+35
+P60/SEDM
+2250
+9606
+448
+Keck1/LRIS
+1200
+8695
+39
+P60/SEDM
+2250
+SN 2020abfe
+9189
+27
+P60/SEDM
+2700
+8697
+41
+P60/SEDM
+2250
+9319
+146
+Keck1/LRIS
+400
+8748
+90
+P60/SEDM
+2250
+SN 2020aekp
+9224
+19
+P60/SEDM
+2160
+8761
+103
+P200/DBSP
+600
+9342
+132
+P60/SEDM
+2160
+SN 2019rvb
+8766
+14
+P60/SEDM
+2250
+9343
+132
+NOT/ALFOSC
+1200
+8780
+26
+P200/DBSP
+600
+9362
+151
+P60/SEDM
+2700
+SN 2020onv
+9058
+23
+P60/SEDM
+1800
+9381
+169
+NOT/ALFOSC
+2400
+9062
+27
+P60/SEDM
+1800
+9404
+191
+P60/SEDM
+2700
+9069
+33
+P60/SEDM
+1800
+9425
+211
+NOT/ALFOSC
+1800
+9070
+34
+LT/SPRAT
+750
+9434
+220
+P60/SEDM
+2700
+9073
+37
+P60/SEDM
+1800
+9448
+233
+P60/SEDM
+2700
+9074
+38
+NOT/ALFOSC
+450
+9468
+252
+P60/SEDM
+2700
+SN 2020qxz
+9076
+13
+P60/SEDM
+2250
+9569
+348
+P60/SEDM
+2700
+9087
+22
+P60/SEDM
+2250
+9092
+26
+NOT/ALFOSC
+1800
+9098
+32
+P60/SEDM
+2250
+9101
+34
+NOT/ALFOSC
+1200
+9107
+40
+P200/DBSP
+900
+9112
+45
+Keck1/LRIS
+300
+9121
+53
+P60/SEDM
+2250
+9141
+71
+Keck1/LRIS
+399
+use only the last non-detection the uncertainty range is typ-
+ically less than 3 days. Only for SN 2020uem is the date of
+explosion virtually unconstrained (±57 days) as it was be-
+hind the sun at the time of explosion.
+Although for SN 2019ibk the explosion time is formally
+constrained with a ±3 day uncertainty, this estimate was
+derived using only ATLAS o-band data right after the SN
+emerges from behind the sun. There is not a clear rise ob-
+served over a few epochs but two non-detections before a
+5σ detection. It is possible that the actual peak of this SN
+occurred earlier while it was behind the sun and the rising
+o-band points after it emerged are due to a second peak or
+bump (similar to SN 2018evt, in that case the actual rise was
+caught before the SN went behind the sun in ASAS-SN data).
+If the former explosion epoch estimate from o-band is to be
+believed then SN 2019ibk would be the most sub-luminous
+among the SNe Ia-CSM, peaking at −17.5.
+3.1.2. Duration and absolute magnitudes
+Figure 2 shows the SNe Ia-CSM (colored squares) in our
+sample in the duration-luminosity and duration-∆m30 phase
+space. In the top panel, the x-axis is duration above half-max
+and the y-axis is the peak absolute magnitude (see Table 1)
+when we have photometric coverage both pre-peak and post-
+peak. For SNe missing the pre-peak coverage, their discov-
+ery magnitude is taken to be the upper limit to peak absolute
+magnitude and the duration from discovery the lower limit
+
+8
+Table 4. Explosion time epoch estimates derived from pre-peak
+multi-band light curves. For 6 out of 12 SNe Ia-CSM, we were able
+to fit a power-law model to multi-band data following Miller et al.
+(2020). For the remaining 6 SNe, the explosion epoch was estimated
+by taking the mean of the first 5σ detection and last upper-limit
+before the first detection.
+IAU Name
+PL fit filters
+to
+1σ interval
+(MJD)
+(days)
+SN 2018crl
+g, r, o
+58271.83
+[−0.48,+0.38]
+SN 2018gkx
+r, o
+58371.34
+[−0.64,+0.53]
+SN 2018evt
+-
+58334.26
+[−2.00,+2.00]
+SN 2019agi
+-
+58502.48
+[−1.51,+1.51]
+SN 2019ibk
+-
+58654.61
+[−2.99,+2.99]
+SN 2019rvb
+g, r, i, o
+58749.16
+[−0.79,+0.60]
+SN 2020onv
+o
+59032.75
+[−2.49,+1.10]
+SN 2020qxz
+g, r, o
+59063.05
+[−0.51,+0.45]
+SN 2020uem
+-
+59117.03
+[−56.63,+56.63]
+SN 2020xtg
+-
+59130.14
+[−0.04,+0.04]
+SN 2020abfe
+g, r, o
+59159.36
+[−2.16,+2.23]
+SN 2020aekp
+-
+59204.53
+[−5.50,+5.50]
+to duration above half-max (marked by arrows in Figure 2).
+The BTS SN Ia sample is shown in gray points, and we also
+show the SNe Ia-CSM presented in S13 with empty trian-
+gles for comparison in the top panel. In the bottom panel,
+the x-axis is duration above 20% of peak flux (∆t20) and the
+y-axis is ∆m30, the two parameters used in the selection cri-
+teria. Most of the SNe Ia-CSM lie on the longer duration
+and brighter luminosity side, and are even more distinctly
+separated in the ∆t20-∆m30 phase space. This makes the
+SN initial decline rate and duration useful tools for identify-
+ing thermonuclear SNe potentially interacting with CSM, if
+they have not revealed themselves already in their early time
+spectra. The gray points lying in the same phase space as
+SNe Ia-CSM are the false positive cases described in §2.1.
+Also worth noting is that the duration calculated by taking
+the flux above half of peak flux value does not capture the
+true duration of the light curve when the plateau phase falls
+below half-max as is the case for SN 2020aekp (> 500 days
+light curve) but ∆t20 and ∆m30 do.
+3.1.3. Light and color curves
+We have good pre-peak coverage in ZTF data for 8 of
+the 12 SNe in our sample8.
+SN 2018evt was discovered
+by ASAS-SN on JD 2458341.91 (Nicholls & Dong 2018)
+and classified by ePESSTO the next day (Stein et al. 2018a),
+around 115 days before the first detection in ZTF when the
+SN came back from behind the sun. Hence we have only one
+8 except for SNe 2018evt, 2019ibk, 2020onv and 2020uem.
+0
+25
+50
+75
+100
+125
+150
+175
+Rest-Frame duration above half-max (days)
+22
+21
+20
+19
+18
+17
+16
+15
+Peak absolute magnitude
+Ia
+SN2018crl
+SN2018gkx
+SN2018evt
+SN2019agi
+SN2019ibk
+SN2019rvb
+SN2020onv
+SN2020qxz
+SN2020uem
+SN2020xtg
+SN2020abfe
+SN2020aekp
+Silverman 2013
+0
+100
+200
+300
+400
+500
+Rest-frame duration above 20% of max (days)
+0.0
+0.5
+1.0
+1.5
+2.0
+2.5
+3.0
+mpeak + 30d
+mpeak (r-band)
+Figure 2. Top: Location of our 12 SNe Ia-CSM in the peak absolute
+magnitude vs. rest-frame duration above half max phase space. The
+colored points are the BTS SNe Ia-CSM and the gray points are the
+rest of the BTS SNe Ia. Also shown with empty triangles are the
+SNe Ia-CSM from S13. The vertical arrows mark the upper limits
+to peak absolute magnitudes and horizontal arrows mark the lower
+limits to durations of SNe not having pre-peak coverage. Bottom:
+Change in magnitude 30 days after peak (∆m30) vs. rest-frame
+duration above 20% of peak-flux for BTS SNe Ia and SNe Ia-CSM.
+These criteria were used to filter out potential SNe Ia-CSM from all
+SNe Ia and demonstrate that SNe Ia-CSM occupy a distinct portion
+in this phase space. However some gray points (not SN Ia-CSM)
+remain on the longer duration side and are the false positive cases
+described in §2.1.
+epoch of pre-peak photometry and one early spectrum for SN
+2018evt.
+Our mixed bag of SNe Ia-CSM show post-maximum de-
+cline rates ranging from 0.5 to 2.0 mag 100d−1 in the r
+band from peak to ∼ 100 days post peak. The median de-
+cline rate is 1.07 mag 100d−1, which is much slower than
+the decline rates of normal SNe Ia.
+We see a variety of
+changes in decline rates after around 100 days from peak.
+Two SNe (2020onv and 2020abfe) show no change and have
+a constant slow decline throughout.
+Four SNe (2018gkx,
+2019agi, 2019ibk and 2019rvb) evolve to a shallower slope
+going from ∼ 0.6–1 mag 100d−1 to ∼ 0.2–0.5 mag 100d−1.
+Three SNe (2018crl, 2020qxz and 2020aekp) show a ma-
+jor change in decline rate with the light curves becoming
+
+9
+almost flat, and SN 2020aekp shifts back to a slow de-
+cline from this plateau after ∼ 200 days. In three cases,
+the decline rate actually becomes steeper, SN 2018evt goes
+from 0.52 mag 100d−1 to 1.4 mag 100d−1, SN 2020uem goes
+from 0.52 mag 100d−1 to 1.25 mag 100d−1 and SN 2020xtg
+seems to go from 0.61 mag 100d−1 to 1.35 mag 100d−1 (even
+though there is only one epoch at late times to measure
+this change).
+The 3 SNe with fastest initial decline rates
+(≳ 1.5 mag 100d−1 in the r band) are similar to SN 2002ic
+(initial decline of 1.66 mag 100d−1 in V ) and PTF11kx (ini-
+tial decline of 3.3 mag 100d−1 in R) and coincidentally are
+also the ones that evolve into a plateau.
+The rest of the
+sample have initial decline rates comparable to SN 1997cy
+(0.75 mag 100d−1) and SN 2005gj (0.88 mag 100d−1) (In-
+serra et al. 2016). From these observations, we can conclude
+that SNe Ia-CSM exhibit a range of slow evolution indicat-
+ing that there exists a continuum of phases at which strong
+CSM interaction begins to dominate the powering of the light
+curves for these SNe. It is, however, difficult to pinpoint
+the exact phase when interaction starts from the light curve
+without modeling. CSM interaction could be affecting the
+peak brightness significantly even in cases where interaction
+only appears to dominate after a few weeks (SNe 2018crl,
+2020qxz 2020aekp). Considering the average peak phase to
+be ∼ 20 days past explosion from the light curves and as-
+suming an ejecta velocity of ∼ 20000 km s−1, the CSM is
+located at ∼ 3.5 × 1015 cm. This estimate can be refined
+by considering the phase of the earliest spectrum that shows
+interaction signatures (see §3.2). At late times, all the de-
+cline rates are slower than that expected from Cobalt decay
+(0.98 mag 100d−1), confirming that the power from CSM in-
+teraction dominates the light curve behaviour for a long time.
+Figure 3 shows the g − r color evolution of our sam-
+ple SNe as a function of phase (rest-frame days from r-
+band maximum), comparing them with some famous SNe
+Ia-CSM (SNe 2005gj, 1997cy, 1999E), and SNe 2012ca (Ia-
+CSM/IIn), 2010jl (IIn) and 1991T (over-luminous Type Ia).
+The color evolution of normal SNe Ia from ZTF (Dhawan
+et al. 2022) is shown in grey lines. We use g − r colors
+when available, otherwise we estimate the g − r color by
+fitting Planck functions to estimate the black-body tempera-
+tures from the V − R colors. Our SNe Ia-CSM show simi-
+lar color evolution as the older Type Ia-CSM/IIn interacting
+SNe, i.e. the g − r color increases gradually for about 100
+days and then settles onto a plateau or slowly declines, and
+one object (SN 2019ibk) becomes redder at late times similar
+to SN 2012ca. The interacting SNe are redder at late times
+compared to the normal SNe Ia.
+3.1.4. Mid-IR brightness comparison
+Out of 12 SNe in our sample, only one observed (SN
+2020abfe) did not have 3σ detections post explosion in the
+unWISE difference photometry light curves and two (SNe
+2019rvb and 2020qxz) did not have coverage post explosion.
+The unWISE light curves for the rest of the SNe Ia-CSM
+having > 3σ detections in W1 (3.3 µm) and W2 (4.6 µm)
+bands are shown in Figure 4 (black and red stars) along with
+Spitzer IRAC survey data of SN 2008cg (indigo and ma-
+genta empty triangles), SN 2008J (indigo and magenta empty
+squares) (both Ia-CSM) and some SNe IIn (blue and orange
+crosses) taken from Fox et al. (2011). The most nearby SN
+in our sample, SN 2018evt, is among the brightest (∼ 17 AB
+mag) in MIR at least until ∼1000 days after explosion and
+has a bumpy light curve. SNe 2019ibk and 2018crl however
+are the most luminous with an absolute magnitude of −18.7
+mag in the W1 band. The brightness of the BTS SNe Ia-CSM
+is comparable with other interacting SNe and span a similar
+range (−16 to −19). However, SNe IIn have been detected
+until even later epochs (up to 1600 days) than SNe Ia-CSM,
+probably due to the larger number of SNe IIn at closer dis-
+tances. SN 2020abfe has upper limits around ∼ −18 in W1
+band and ∼ −18.5 in W2 band up to ∼300 days post explo-
+sion shown with upside down filled triangles. As the mid-IR
+luminosity can be fainter than these limits for SNe Ia-CSM
+(as can be seen for other nearby SNe in this sample) and SN
+2020abfe is at a redshift of 0.093, it might just be out of reach
+for WISE.
+This brightness of SNe Ia-CSM in mid-IR can be indica-
+tive of existing or newly formed dust. A clear signature of
+new dust is reduced flux in the red wing of the Hα emission
+line at late phases as the new dust formed in the cold dense
+shell behind forward shock absorbs the far-side (redshifted)
+intermediate and narrow line emission (see bottom panel of
+Fig. 7). For our sample, this reduction in Hα red wing is the
+most pronounced for SN 2018evt.
+3.1.5. Bolometric luminosity
+As the SN Ia-CSM luminosity is dominated by CSM inter-
+action, their spectra comprise of a pseudo-continuum on the
+blue side and strong Hα emission on the red side, hence a
+blackbody fit to multi-band photometric data is not appropri-
+ate to estimate the bolometric luminosity. Instead we calcu-
+late a pseudo-bolometric luminosity from the available multi-
+band optical data by linearly interpolating the flux between
+the bands and integrating over the optical wavelength range
+spanned by the ATLAS and ZTF bands. The individual band
+light curves are first interpolated using Gaussian process re-
+gression to fill in the missing epochs. This estimate places a
+strict lower limit on the bolometric luminosity.
+In Figure 5 we show the pseudo-bolometric luminosity
+of our SN Ia-CSM sample in comparison with SN 1991T
+(Type Ia), SNe 1997cy, 1999E, 2002ic, 2005gj, 2013dn and
+PTF11kx (Ia-CSM). Multi-band photometric data were taken
+from the Open Supernova Catalog (Guillochon et al. 2017)
+
+10
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2018crl
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2018gkx
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2018evt
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2019agi
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2019ibk
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2019rvb
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+D
+SN 2020onv
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2020qxz
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2020uem
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2020xtg
+0
+150
+300
+450
+600
+D
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2020abfe
+0
+150
+300
+450
+600
+-0.5
+0.0
+0.5
+1.0
+1.5
+2.0
+SN 2020aekp
+Rest-frame days from r-band maximum
+g
+r color (mag)
+SN2005gj
+SN2012ca
+SN1997cy
+SN1999E
+SN1991T
+SN2010jl
+SN Ia
+Figure 3. Color evolution (g − r) of BTS SNe Ia-CSM from r-band maximum (plotted in black) compared with SNe 2005gj, 1997cy, 1999E
+(Ia-CSM), SN 2012ca (IIn/Ia-CSM), SN 2010jl (IIn), SN 1991T (SN Ia) and ZTF SNe Ia (gray lines). As can be seen for up to ∼ 150 days,
+our SNe Ia-CSM tend to be redder than SNe Ia and at late times develop a plateau similar to other interacting SNe (IIn/Ia-CSM).
+
+11
+0
+250
+500
+750
+1000
+1250
+1500
+1750
+Days since explosion
+19.5
+19.0
+18.5
+18.0
+17.5
+17.0
+16.5
+16.0
+15.5
+Absolute Magnitude (AB)
+BTS SN Ia-CSM: W1
+BTS SN Ia-CSM: W2
+SN 2008J: 3.6 m
+SN 2008J: 4.5 m
+SN 2008cg: 3.6 m
+SN 2008cg: 4.5 m
+SN IIn: 3.6 m
+SN IIn: 4.5 m
+Figure 4. unWISE detections in the W1 and W2 bands of BTS
+SNe Ia-CSM. The W1 and W2 points are marked with black and
+red filled stars respectively. Spitzer IRAC photometry of SNe IIn
+(blue and orange crosses) and two SNe Ia-CSM from Fox et al.
+(2011) (SNe 2008cg and 2008J in empty triangle and square) are
+also shown for comparison. 9 out of 12 BTS SNe Ia-CSM are as
+bright in mid-IR as other interacting SNe (∼ −16 to ∼ −19). The
+upper limits for SN 2020abfe are shown in black and red filled up-
+side down triangles.
+for SN 1991T (Filippenko et al. 1992; Ford et al. 1993;
+Schmidt et al. 1994) to generate the bolometric luminos-
+ity light curve through black body fitting.
+The pseudo-
+bolometric luminosity light curve for SN 1997cy was ob-
+tained from Germany et al. (2000), for SN 2013dn from
+Fox et al. (2015) and for SNe 2002ic, 2005gj, 1999E and
+PTF11kx from Inserra et al. (2016).
+All BTS SNe Ia-CSM show a slow evolution in bolomet-
+ric luminosity, inconsistent with the decay of 56Co to 56Fe.
+The sample’s overall luminosity decline rates are comparable
+to those of SNe 1997cy and 2013dn, as shown in Figure 5.
+Only SNe 2018crl and 2020aekp seem to show early decline
+in their pseudo-bolometric light curves similar to SN 1991T
+for about 40 days after peak like SN 2002ic and PTF11kx.
+Another BTS interacting SN Ia, ZTF20aatxryt (Kool et al.
+2022), was found to follow the PTF11kx light-curve evo-
+lution very closely and as its light curve fell into a plateau
+the SN started showing signs of interaction with a helium-
+rich CSM and evolved into a helium-rich SN Ia-CSM. We
+have excluded ZTF20aatxryt from the sample as we focus on
+typical SNe Ia-CSM interacting with hydrogen-rich CSM in
+this study. At late phases (∼ 300 days), the SNe Ia-CSM
+are approximately 100 times brighter than normal SNe Ia
+at the same epoch. Therefore, at these late phases, the lu-
+minosity and spectral features of SNe Ia-CSM are entirely
+dominated by CSM-interaction with little emergent SN flux.
+From the pseudo-bolometric light curves, we place a lower
+limit on the total radiated energy for SNe Ia-CSM to be 0.1–
+1.5 ×1050erg. This is well below the thermonuclear budget
+(Ekin ∼ 1051 erg), but as this is a lower limit and some SNe
+in the sample have unconstrained peaks, the true total radia-
+tive energy might come close to the thermonuclear budget,
+requiring high conversion efficiency to achieve their lumi-
+nosity.
+0
+80
+160
+240
+320
+400
+Rest-frame days since explosion
+41.0
+41.5
+42.0
+42.5
+43.0
+43.5
+44.0
+log(Lopt (erg/s))
+SN2013dn
+SN1997cy
+SN1991T
+PTF11kx
+SN2002ic
+SN2005gj
+SN1999E
+SN2018crl
+SN2018gkx
+SN2018evt
+SN2019agi
+SN2019ibk
+SN2019rvb
+SN2020onv
+SN2020qxz
+SN2020uem
+SN2020xtg
+SN2020abfe
+SN2020aekp
+Figure 5. Pseudo-bolometric luminosity light curves of BTS SNe
+Ia-CSM compared with pseudo-bolometric light curves of SNe
+1991T, 1997cy, 1999E, 2002ic, 2005gj, 2013dn, and PTF11kx from
+literature. The light curves in each filter having more than 10 epochs
+were interpolated using Gaussian process regression to fill in the
+missing epochs, and at each epoch the fluxes between the bands
+were linearly interpolated and integrated over the optical wave-
+length range spanned by ZTF and ATLAS filters to get the pseudo-
+bolometric luminosity. For BTS SNe, the phases are with respect
+to the estimated explosion epochs, while for comparison SNe the
+phases are with respect to discovery.
+3.2. Spectroscopy
+Figure 6 displays the spectral series obtained for the BTS
+SNe Ia-CSM. Most of the early time spectra were taken
+with the SEDM, the BTS workhorse instrument (R ∼100),
+which is not able to resolve the narrow CSM lines. There-
+fore, these SNe were followed up with higher resolution in-
+struments to get more secure classifications. For each spec-
+trum in Figure 6, the phase is provided with respect to the
+explosion epoch estimate given in Table 4. We have spec-
+tra ranging from a few to around 470 days from explo-
+sion. Considering the well constrained explosion times of
+SN 2018evt, presence of narrow Hα in its first spectrum at
+8 days since explosion and assuming a typical ejecta veloc-
+ity of ∼20000 km s−1, this implies that the CSM interaction
+start as close as ∼1.4×1015 cm.
+Figure 7 shows the early time (left) and late time (right)
+spectral behaviour of the BTS SNe Ia-CSM together with a
+few historical SNe for comparison, namely SNe Ia-CSM SN
+2011jb (Silverman et al. 2013), SN 2005gj and PTF11kx, the
+Type Ia SN 1991T and the well-observed Type IIn SN 2010jl.
+Vertical gray regions mark typical SN Ia absorption features
+
+12
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+2
+0
+2
+4
+6
+8
+10
+12
+14
+9 d
+15 d
+21 d
+31 d
+92 d
+H
+H
+H
+He I
+H
+Ca II
+SN2018crl
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+1
+0
+1
+2
+3
+4
+75 d
+H
+H
+H
+He I
+H
+Ca II
+SN2018gkx
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+0
+2
+4
+6
+8
+10
+9 d
+127 d
+143 d
+144 d
+195 d
+H
+H
+H
+He I
+H
+Ca II
+SN2018evt
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+0
+2
+4
+6
+8
+10
+12
+14
+16
++ Constant
+35 d
+39 d
+41 d
+90 d
+103 d
+H
+H
+H
+He I
+H
+Ca II
+SN2019ibk
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+0
+1
+2
+3
+42 d
+H
+H
+H
+He I
+H
+Ca II
+SN2019agi
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+0
+1
+2
+3
+4
+5
+14 d
+26 d
+H
+H
+H
+He I
+H
+Ca II
+SN2019rvb
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+0
+2
+4
+6
+8
+10
+12
+Normalized Flux
+23 d
+27 d
+33 d
+34 d
+37 d
+38 d
+H
+H
+H
+He I
+H
+Ca II
+SN2020onv
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+0.0
+2.5
+5.0
+7.5
+10.0
+12.5
+15.0
+17.5
+20.0
+13 d
+22 d
+26 d
+32 d
+34 d
+40 d
+45 d
+53 d
+71 d
+H
+H
+H
+He I
+H
+Ca II
+SN2020qxz
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+Rest Wavelength (Å)
+0
+5
+10
+15
+20
+25
+19 d
+132 d
+132 d
+151 d
+169 d
+191 d
+211 d
+220 d
+233 d
+252 d
+348 d
+H
+H
+H
+He I
+H
+Ca II
+SN2020aekp
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+0
+2
+4
+6
+8
+10
+12
+27 d
+146 d
+H
+H
+H
+He I
+H
+Ca II
+SN2020abfe
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+0
+5
+10
+15
+20
+25
+11 d
+18 d
+51 d
+101 d
+130 d
+140 d
+167 d
+349 d
+360 d
+448 d
+451 d
+H
+H
+H
+He I
+H
+Ca II
+SN2020uem
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+0
+5
+10
+15
+20
+91 d
+340 d
+448 d
+H
+H
+H
+He I
+H
+Ca II
+SN2020xtg
+P60/SEDM
+NOT/ALFOSC
+P200/DBSP
+Keck1/LRIS
+LT/SPRAT
+Lick-3m/KAST
+Ekar/AFOSC
+NTT/EFOSC2
+UH88/SNIFS
+APO/DIS
+Figure 6. Spectral series of all SNe Ia-CSM presented in this paper. The rest-frame phases are shown alongside the spectra in each subplot and
+have been calculated using the explosion epoch estimate. The colors depict different instruments used to obtain this data. Major emission lines
+are marked with vertical dashed lines.
+
+13
+and [Fe II/III] line regions, and vertical dashed lines mark the
+Balmer emission lines. The sample spectra have been mul-
+tiplied by a constant factor to magnify relevant spectral fea-
+tures. In the following paragraphs, we compare the observa-
+tions of some of the spectral features with previous analysis
+of this class (Silverman et al. 2013; Fox et al. 2015; Inserra
+et al. 2016).
+A few of our early time SNe Ia-CSM show underlying SN
+Ia absorption features like PTF11kx and SN 2002ic (most
+are, however, quite diluted and also affected by the low res-
+olution and signal-to-noise ratio (SNR) of the SEDM spec-
+tra), the most notable being SNe 2018evt, 2020qxz and
+2020aekp. SNe 2020qxz and 2020aekp also have among the
+fastest initial post-peak decline rates in the sample, similar to
+PTF11kx, while coverage around peak is not available for SN
+2018evt. On the other hand, SNe with slower decline rates
+similar to SN 1997cy and SN 2005gj have more SN IIn-like
+early time spectra dominated by blue pseudo-continuum and
+Balmer emission. The faster decline rate suggests we are still
+seeing some of the emission from the ejecta at those phases.
+To unveil the nature of the progenitor of interacting SNe, it is
+therefore necessary to obtain some spectroscopic follow-up
+before peak light. Spectroscopic data at the phase of tran-
+sition to interaction-dominated luminosity would also help
+in deducing the extent and density structure of the optically
+thick CSM.
+Late time spectra of SNe Ia-CSM look very similar to those
+of SNe IIn, heavily dominated by Hα emission. The CSM in-
+teraction masks the underlying SN signature and we instead
+see late-time spectra riddled with photoionized CSM lines.
+In some cases, the photosphere might lie in an optically thick
+cold dense shell (CDS) formed between the forward and re-
+verse shocks which obscures the ejecta completely (Smith
+et al. 2008; Chugai et al. 2004). The continuum is also en-
+shrouded under a blue quasi-continuum from a forest of iron-
+group element lines (S13) as identified and analyzed for SNe
+2012ca and 2013dn by Fox et al. (2015).
+The blue quasi-continuum blend of iron lines ([Fe III] lines
+around ∼4700 ˚A and [Fe II] around ∼5200 ˚A) in the spectra
+of the BTS SN Ia-CSM sample (see Figure 7 top right panel)
+is the dominant feature blue-ward of 5500 ˚A but the ratio
+of [Fe III]/[Fe II] is much weaker compared to for SNe Ia
+(like SN 1991T). This feature is more apparent in the SNe Ia-
+CSM like PTF11kx and SN 2002ic that became interaction-
+dominated later than for other SNe Ia-CSM such as SNe
+1997cy, 1999E and SN 2012ca (SN Ia-CSM/IIn, for which
+a clear type has not been established). Inserra et al. (2014)
+argues for a core-collapse origin for SN 2012ca given this
+low amount of [Fe III] along with the detection of blueshifted
+Carbon and Oxygen lines (which however, were later argued
+to be [Fe II] lines by Fox et al. 2015). S13 instead argues
+in favor of a thermonuclear origin given the presence of this
+blue quasi-continuum, despite [Fe III] being weaker. Fox
+et al. (2015) points out that a similarly suppressed ratio of
+[Fe III]/[Fe II] is observed in some SNe Ia, particularly the
+super-Chandra candidate SN 2009dc, for which the expla-
+nation was suggested to be a low ionization nebular phase
+owing to high central ejecta density and low expansion ve-
+locities (Taubenberger et al. 2013). Fox et al. (2015) argue
+that in the case of SNe Ia-CSM, a lower ionization state could
+arise owing to the deceleration of ejecta by the dense CSM
+explaining the Fe line ratio suppression. Since Ca has lower
+first and second ionization potentials than Fe, the detection
+of [Ca II] λλ7291, 7324 would be consistent with this low
+ionization, which Fox et al. (2015) confirms for SNe 2012ca
+and 2013dn. Indeed, we find clear evidence of [Ca II] emis-
+sion for 8 out of 12 SNe in our sample and moderate to weak
+signal for the remaining 4.
+Although this does favor the
+argument for a thermonuclear origin, a similar blue quasi-
+continuum is also observed in other interacting SN types like
+SNe Ibn (SN 2006jc, Foley et al. 2007) and SNe IIn (SNe
+2005ip and 2009ip), making Fe an incomplete indicator of
+the progenitor nature (see detailed discussion in Fox et al.
+2015).
+We do not find strong evidence of O I λ7774 or [O I]
+λλ6300, 6364 emission in our sample, although they might
+be present at very weak levels in some SNe (e.g. SN
+2020uem). SN 2020uem has strong emission lines at 6248,
+7155 and 7720 ˚A which are consistent with being iron lines
+and were also observed in SNe 2012ca, 2013dn and 2008J.
+S13 note that the very broad emission around 7400 ˚A can be
+due to a blend of [Ca II] λλ7291, 7324 and [O II] λλ7319,
+7330, however we note that this broad emission is likely to
+be from calcium as O II is harder to excite than O I which is
+either very weak or absent in our spectra. The broad Ca NIR
+triplet feature resulting from electron scattering is the next
+strongest feature after the Balmer emission and is present in
+all mid to late-time spectra of the SNe in our sample where
+the wavelength coverage is available. We observe it increas-
+ing in relative strength with phase, at least for a year, after
+which we no longer have spectral coverage.
+The bottom panel of Figure 7 shows the line profile of Hα,
+with the blue side reflected over the red side at the maxi-
+mum flux after continuum removal. We do see evidence of
+diminished flux in the red wing of Hα at late phases in some
+SNe (most notable in SNe 2018evt and 2020uem), which can
+indicate formation of new dust in the post-shock CSM. S13
+claim to observe this for all non-PTF SNe Ia-CSM in their
+sample starting at ∼75–100 days, while for the PTF SNe
+Ia-CSM they do not have spectra available post that phase
+range. For some BTS SNe Ia-CSM, we also do not have
+spectra available post 100 days which limits any analysis of
+this phenomenon for a large enough sample.
+
+14
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+Rest Wavelength (Å)
+Normalized Flux + Constant
+15 d
+9 d
+14 d
+23 d
+13 d
+11 d
+27 d
+19 d
++23 d
++6 d
++11 d
++21 d
+H
+He I
+H
+Ca II
+Mg II
+Si II
+Early time spectra comparison
+SN2018crl
+SN2018gkx
+SN2018evt
+SN2019agi
+SN2019ibk
+SN2019rvb
+SN2020onv
+SN2020qxz
+SN2020uem
+SN2020xtg
+SN2020abfe
+SN2020aekp
+SN2011jb
+SN2005gj
+PTF11kx
+SN1991T
+SN2010jl
+3000
+4000
+5000
+6000
+7000
+8000
+9000
+10000
+Rest Wavelength (Å)
+Normalized Flux + Constant
+92 d
+75 d
+144 d
+42 d
+103 d
+26 d
+38 d
+71 d
+140 d
+340 d
+146 d
+169 d
++185 d
++91 d
++328 d
++21 d
++83 d
+H
+He I
+H
+[O I]
+O I
+He I
+Ca II IR
+[Fe II/III]
+[Fe III]
+Late time spectra comparison
+8000
+4000
+0
+4000
+8000
+Velocity km s
+1
+0
+1
+2
+3
+4
+5
+6
+7
+8
+Normalized flux + Constant
+92.0 d
+75.0 d
+144.0 d
+42.0 d
+103.0 d
+26.0 d
+38.0 d
+71.0 d
+8000
+4000
+0
+4000
+8000
+Velocity km s
+1
+0
+1
+2
+3
+4
+5
+6
+7
+8
+Normalized flux + Constant
+167.0 d
+360.0 d
+340.0 d
+448.0 d
+146.0 d
+132.0 d
+169.0 d
+211.0 d
+SN2018crl
+SN2018gkx
+SN2018evt
+SN2019agi
+SN2019ibk
+SN2019rvb
+SN2020onv
+SN2020qxz
+SN2020uem
+SN2020xtg
+SN2020abfe
+SN2020aekp
+Figure 7. Top left: Early-time spectra of BTS SNe Ia-CSM with phases between 0 and 30 days since explosion compared to spectra of SNe
+2011jb, 2005gj, 1991T and PTF11kx (phases in days since discovery). Top right: Late-time spectra of BTS SNe Ia-CSM (phases ranging from
+40 to 370 days since explosion) compared to spectra of SNe 2011jb, 2005gj, 2010jl and PTF11kx (phases in days since discovery).
+Bottom left and right: Hα line profiles (post continuum removal) with the blue side reflected across the peak flux, marked by dashed lines.
+SNe 2020aekp, 2020abfe, 2020xtg and 2020uem in the right panel, and SNe 2018crl, 2018gkx, 2018evt, 2019agi, 2019ink, 2019rvb, 2020onv,
+2020qxz in left panel.
+
+15
+The spectra were reduced and processed as outlined in §2.5
+for the emission line analysis, the results of which are de-
+scribed in the next section. We used only good SNR SEDM
+spectra and intermediate resolution spectra for line identifi-
+cation and analysis.
+3.2.1. Hα, Hβ and He I emission lines
+To analyze the Hα line emission, we first fit the con-
+tinuum level using the fit continuum function of the
+specutils Python package, where the continuum is es-
+timated by a cubic function fitted on regions on each side of
+the line. We remove this continuum level and then fit the Hα
+line with a broad and a narrow component Gaussian func-
+tion using the fit lines function of specutils which
+returns the best fit Gaussian model and the 1σ uncertainty
+on the model parameters. We generate 1000 sample mod-
+els within 1σ uncertainties of the parameters centered around
+the best-fit values and calculate the intensity, flux and veloc-
+ity (FWHM) of the broad and narrow components for each
+model. Then we take the median and standard deviation of
+the intensity, flux and velocity FWHM distributions to get
+their final best value and 1σ uncertainty.
+The equivalent
+width was also calculated for the Hα line using the model
+fit as well as directly from the data, and the difference be-
+tween the values derived from model and data is reported as
+the error on the EW. All values are reported in Table 5. For 3
+SNe in our sample, we have a series of intermediate resolu-
+tion spectra through which we can trace the evolution of the
+Hα line with phase. Figure 8 shows this trend of the Hα line
+parameters (integrated flux in the top panel and equivalent
+width in the bottom panel) versus phase for all SNe in our
+sample. The un-filled markers represent the narrow emis-
+sion while the filled markers represent the broad emission.
+For SNe where this analysis could be done on multiple spec-
+tra, we see that the Hα equivalent width generally increase
+over time, with some SNe showing fluctuations up to 100
+days possibly due to interaction of ejecta with multiple CSM
+shells of varying density. For SN 2018evt, Yang et al. (2022)
+analyzed Hα line properties from a comprehensive spectral
+series data, which are plotted in Figure 8 in gray circles and
+seem to agree well with our analysis at comparable epochs.
+From the Gaussian profile line fitting analysis of the Hα
+emission line, we found that the broader component has ve-
+locities ranging from ∼1000 to ∼4000 km s−1 (intermedi-
+ate width) and the narrow component has velocities of about
+∼200 km s−1 to ∼1000 km s−1 (see Figure 9). The narrow
+component could only be resolved down to ∼300 km s−1
+limited by the mediocre resolution of the spectrographs used
+(KeckI/LRIS R∼800, P200/DBSP R∼1000, NOT/ALFOSC
+has R∼360). While we know that the narrow lines originate
+in the unshocked ionized CSM, the exact origin of the inter-
+mediate components is uncertain. They could arise from the
+post-shock gas behind the forward shock or from the shocked
+dense clumps in the CSM (Chugai & Danziger 1994).
+The luminosities of the Hα line measured from the BTS
+SNe Ia-CSM lie in the range 2.5–37×1040 erg s−1 which are
+comparable to the values from S13 who reported most of
+their SNe in the 1–10×1040 erg s−1 range except one object
+that had a luminosity of 39×1040 erg s−1. From the broad
+Hα luminosity, we did a simple estimate of the mass-loss rate
+assuming spherically symmetric CSM deposited by a station-
+ary wind ρ ∝ r−2 having velocity vw (Chugai 1991; Sala-
+manca et al. 1998). The mass-loss rate ˙M can be related to
+the broad Hα luminosity LBroad
+Hα
+as (Salamanca et al. 1998,
+their Eq. 2)
+LBroad
+Hα
+= 1
+4ϵHα
+˙M
+vw
+v3
+s
+where vs is the shock velocity (obtained from the broad
+component velocity of the Hα line). We used a value of
+100 km s−1 considering previous high resolution spectral
+studies of SNe Ia-CSM (Kotak & Meikle 2005; Aldering
+et al. 2006; Dilday et al. 2012) for vw as we cannot fully
+resolve the narrow component and a maximum value of 0.1
+for the efficiency factor ϵHα (Salamanca et al. 1998). The
+mass-loss rates were estimated from the available spectra
+and are shown in Figure 10 as a function of years before
+explosion (tw =
+vst
+vw , where t is the phase of the spectra).
+For most SNe in the sample, the mass-loss rates lie be-
+tween 0.001–0.02 M⊙ yr−1, except for SN 2019rvb which
+has ∼0.07 M⊙ yr−1 lost within 2 years prior the explosion.
+These rates are much higher than what could be attained
+from a red giant superwind (∼ 3 × 10−4 M⊙ yr−1) but are
+comparable to previous estimates (calculated through mul-
+tiple methods) for SNe Ia-CSM and require some unusual
+mechanism to reach such persistently higher mass-loss rates
+in the decades prior to explosion. Also to consider is that the
+simplistic assumption of spherical symmetry likely does not
+apply for SNe Ia-CSM. Evidence of multiple thin shells and
+asymmetric CSM was observed for PTF11kx (Dilday et al.
+2012) and light curve modeling of SNe 1997cy and 2002ic
+suggested a better fit to a flat density profile rather than sta-
+tionary wind (Chugai & Yungelson 2004). An asymmetric or
+clumpy CSM might be the norm for SNe Ia-CSM (and some
+SNe IIn) rather than the exception.
+The same analysis as for the Hα line was also carried out
+for Hβ and He I λ5876 with a one component Gaussian fit.
+For cases where a Gaussian model could not fit the data, we
+integrate the flux value in a 100 ˚A region centered at 5876 ˚A
+for He I. The Na ID absorption lines are also prevalent in
+some spectra and blend with the He I line, resulting in posi-
+tive EWs for some SNe. The cumulative distributions of Hβ
+and He I equivalent widths are shown in the top and bottom
+panels of Figure 11 respectively.
+
+16
+Table 5. Summary of Hα line properties obtained from two-component Gaussian fitting.
+SN Name
+Phase
+Broad Flux
+Narrow Flux
+Total Flux
+Broad Velocity
+Narrow Velocity
+(days)
+(10−16 erg s−1 cm−2)
+(10−16 erg s−1 cm−2)
+(10−16 erg s−1 cm−2)
+FWHM (km s−1)
+FWHM (km s−1)
+SN 2018crl
+92
+135.4±10.0
+32.8±2.0
+168.2±12.0
+4137±312
+< 214
+SN 2018gkx
+75
+9.9±0.7
+3.9±0.2
+13.7±0.9
+2640±398
+< 375
+SN 2018evt
+144
+2020.3±128.5
+1247.4±52.8
+3267.7±181.3
+6465±997
+1816±973
+SN 2019agi
+42
+52.7±3.6
+23.7±1.1
+76.4±4.7
+3836±349
+464±301
+SN 2019ibk
+103
+85.6±1.7
+17.0±0.5
+102.6±2.3
+2431±217
+272±214
+SN 2019rvb
+26
+22.0±3.0
+10.4±1.0
+32.5±4.1
+2321±298
+374±216
+SN 2020onv
+38
+32.8±5.2
+33.3±2.0
+66.1±7.2
+2714±879
+<834
+SN 2020qxz
+26
+76.6±6.2
+13.8±1.7
+90.4±7.9
+11294±1106
+< 836
+SN 2020qxz
+34
+55.1±5.0
+10.8±1.8
+65.9±6.8
+8252±1039
+1070±845
+SN 2020qxz
+40
+12.9±1.7
+7.6±0.5
+20.5±2.2
+2049±284
+245±215
+SN 2020qxz
+45
+20.7±1.6
+9.1±0.4
+29.8±2.1
+3429±419
+< 375
+SN 2020qxz
+71
+39.1±1.3
+10.4±0.4
+49.5±1.7
+5013±395
+400±375
+SN 2020uem
+51
+246.3±47.2
+151.1±16.8
+397.4±64.0
+6520±1163
+1178±840
+SN 2020uem
+101
+655.2±28.9
+241.2±9.6
+896.4±38.4
+7456±309
+1066±217
+SN 2020uem
+130
+552.9±17.6
+281.8±6.2
+834.8±23.8
+7465±265
+1269±215
+SN 2020uem
+140
+545.4±20.0
+283.4±6.8
+828.8±26.7
+7457±275
+1308±216
+SN 2020uem
+167
+424.3±19.0
+312.0±7.7
+736.3±26.6
+6852±854
+1439±834
+SN 2020uem
+360
+179.8±4.0
+77.4±1.4
+257.2±5.4
+5377±382
+1170±375
+SN 2020xtg
+340
+129.2±4.2
+52.1±1.6
+181.3±5.8
+4242±382
+1258±376
+SN 2020xtg
+448
+131.7±7.7
+96.3±3.2
+228.0±10.9
+4452±395
+1566±377
+SN 2020abfe
+146
+33.6±1.1
+3.0±0.3
+36.6±1.4
+4411±389
+< 376
+SN 2020aekp
+132
+149.5±4.0
+33.0±1.0
+182.5±5.0
+7728±846
+< 833
+SN 2020aekp
+169
+231.0±4.5
+32.3±1.3
+263.3±5.8
+6775±839
+< 834
+SN 2020aekp
+211
+251.0±9.5
+58.6±3.4
+309.6±12.8
+7422±852
+1342±836
+The Hβ median EW measured from the BTS SN Ia-CSM
+sample is 7.1 ˚A , close to the S13 value of ∼6 ˚A and quite
+weak compared to what S13 measured for SNe IIn (∼13 ˚A ).
+The overall cumulative distribution of Hβ EW is also compa-
+rable to the S13 SNe Ia-CSM rather than to the S13 SNe IIn.
+For the He I λ5876 line, the median EW measured for our
+BTS SN Ia-CSM sample, considering only significant emis-
+sion features, is 2.4 ˚A . This is close to the value of ∼2 ˚A
+reported in S13, and again significantly different from their
+SN IIn value of ∼6 ˚A (∼4 ˚A with upper limits), however
+the overall distribution seems to be closer to the S13 SNe IIn
+(but still weaker) rather than to the S13 SNe Ia-CSM. This
+indicates that perhaps He I is not as good a discriminant be-
+tween the populations compared to Hβ. Among the most He-
+rich SNe in our sample are SNe 2019ibk, 2020uem, 2020xtg,
+2020aekp and 2018evt, and these SNe also have the higher
+Hα equivalent widths in the sample.
+Figure 12 plots the cumulative distribution of the Balmer
+decrements ( FHα
+FHβ ) measured for our sample SNe. The higher
+Balmer decrement values (>15) have large errors associated
+to them because of low SNR of the spectra from which they
+were derived, particularly near the Hβ line. Consistent with
+the results of S13, the SNe Ia-CSM from this sample also
+have a high median Balmer decrement value of ∼7 (∼5 in
+S13), indicating that the emission line mechanism is prob-
+ably collisional excitation or self-absorption rather than re-
+combination, from which the expected Balmer decrement
+value is ∼3. In the case of SNe Ia-CSM, if the CSM distri-
+bution consists of multiple shells as suggested for PTF11kx,
+moderately high densities could be created when fast moving
+ejecta overtake slowly moving thin dense CSM shells creat-
+ing large enough optical depth in the Hα line which results
+in the Hβ transition decaying as Paα + Hα (Xu et al. 1992).
+For some individual SNe where multiple spectra are avail-
+able, the Balmer decrement is observed to first increase and
+later on decrease with phase.
+3.3. Host galaxies
+We retrieved science-ready co-added images from the
+Galaxy Evolution Explorer (GALEX) general release 6/7
+(Martin et al. 2005), the Sloan Digital Sky Survey DR 9
+(SDSS; Ahn et al. 2012), the Panoramic Survey Telescope
+and Rapid Response System (Pan-STARRS, PS1) DR1
+(Chambers et al. 2016), the Two Micron All Sky Survey
+(2MASS; Skrutskie et al. 2006), and preprocessed WISE im-
+
+17
+1040
+1041
+Line luminosity (erg s
+1)
+SN2018evt (Yang et al.)
+SN2018crl
+SN2018gkx
+SN2018evt
+SN2019agi
+SN2019ibk
+SN2019rvb
+SN2020onv
+SN2020qxz
+SN2020uem
+SN2020xtg
+SN2020abfe
+SN2020aekp
+100
+200
+300
+400
+Phase (days)
+102
+Equivalent Width (Å)
+Figure 8. Integrated fluxes and equivalent widths of Hα emission
+line with respect to SN phases for the BTS SN Ia-CSM sample.
+Broad component values are shown with filled markers and narrow
+component values with un-filled markers. SN 2018evt Hα lumi-
+nosities and EWs presented in Yang et al. (2022) are also shown in
+gray circles.
+100
+200
+300
+400
+Phase (days)
+102
+103
+104
+Line Velocity (km s
+1)
+SN2018crl
+SN2018gkx
+SN2018evt
+SN2019agi
+SN2019ibk
+SN2019rvb
+SN2020onv
+SN2020qxz
+SN2020uem
+SN2020xtg
+SN2020abfe
+SN2020aekp
+Figure 9. Velocity of Hα emission line with respect to SN phases
+for the BTS SN Ia-CSM sample.
+Broad component values are
+shown with filled markers and narrow component values with un-
+filled markers.
+ages (Wright et al. 2010b) from the unWISE archive (Lang
+2014b)9.
+We used the software package LAMBDAR (Lambda
+Adaptive Multi-Band Deblending Algorithm in R) (Wright
+et al. 2016) and tools presented in Schulze et al. (2021),
+to measure the brightness of the host galaxy. The spectral
+energy distribution (SED) was modelled with the software
+9 http://unwise.me
+0
+10
+20
+30
+40
+50
+Years before explosion
+10
+3
+10
+2
+10
+1
+Mass loss rate (M
+ yr
+1)
+SN2018crl
+SN2018gkx
+SN2018evt
+SN2019agi
+SN2019ibk
+SN2019rvb
+SN2020onv
+SN2020qxz
+SN2020uem
+SN2020xtg
+SN2020abfe
+SN2020aekp
+Figure 10. Mass-loss rates estimated from the luminosity of the
+broad component of Hα for the BTS SNe Ia-CSM. A value of
+100 km s−1 was assumed for the wind velocity.
+package Prospector10 (Johnson et al. 2021).
+We assumed
+a linear-exponential star-formation history, the Chabrier
+(2003) initial mass function, the Calzetti et al. (2000) at-
+tenuation model, and the Byler et al. (2017) model for the
+ionized gas contribution. The priors were set as described in
+Schulze et al. (2021).
+Figure 13 shows the log of star formation rate (SFR) as a
+function of stellar mass for hosts of BTS SNe Ia-CSM. We
+also use a Galaxy-zoo (Lintott et al. 2011) sample of ellip-
+tical and spiral galaxies (randomly sampled in the redshift
+range z = 0.015−0.05), and BTS SN Ia hosts as comparison
+samples collected by and used for comparison in Irani et al.
+(2022). We find the SN Ia-CSM host galaxy population to
+be consistent with late-type spirals and irregulars with recent
+star formation history. 4 out of 12 SNe have clearly spiral
+hosts, 3 have edge-on host galaxies, 4 seem to have irregu-
+lars as hosts and 1 has an unclear host type. Host galaxies
+of 10 out of 12 SNe have w2 − w3 measurements available
+which are all > 1 mag, putting them in late-type category
+(Irani et al. 2022), 1 (SN 2019rvb) does not have W3 mea-
+surement but has NUV − PS1r ∼ 1 mag again putting it
+towards late-type and 1 (SN 2020abfe) does not have any of
+the above information available except the PS1r band mag-
+nitude of 20.766, which is the faintest host galaxy (absolute
+SDSS r-band magnitude of −17.4) in our BTS SN Ia-CSM
+sample. As noted in S13, the SN Ia-CSM hosts of their sam-
+ple had generally low luminosities (−19.1 < Mr < −17.6)
+except MW like spiral hosts. Our BTS SN Ia-CSM host lu-
+minosities lie in the range of −21.8 < Mr < −17.4 covering
+low to MW like luminosities.
+3.4. Rates
+Following the methodology for calculating the volumetric
+rate of transients found in the Bright Transient Survey from
+Perley et al. (2020), we use their equation 2 to calculate the
+10 https://github.com/bd-j/prospector version 0.3
+
+18
+0
+20
+40
+60
+80
+100
+H  Equivalent Width (Å)
+0.0
+0.2
+0.4
+0.6
+0.8
+1.0
+Cumulative Fraction of objects
+BTS Ia-CSM
+S13 Ia-CSM
+S13 IIn
+MedianBTS
+Ia
+CSM = 7.1
+MedianS13
+Ia
+CSM = 6.0
+MedianS13
+IIn  = 13.0
+0
+10
+20
+30
+40
+50
+He I 5876 Equivalent Width (Å)
+0.0
+0.2
+0.4
+0.6
+0.8
+1.0
+Cumulative Fraction of objects
+BTS Ia-CSM
+S13 Ia-CSM
+S13 IIn
+MedianBTS
+Ia
+CSM = 2.4
+MedianS13
+Ia
+CSM = 2
+MedianS13
+IIn  = 4
+Figure 11. Cumulative distributions of equivalent width of Hβ and
+He I λ5876 emission lines calculated from the BTS SNe Ia-CSM
+(in grey) compared with the respective distributions presented in
+S13 for SNe Ia-CSM (blue) and SNe IIn (red). Vertical dashed lines
+mark the median EW of the distributions.
+SN Ia-CSM rate:
+R = 1
+T
+N
+�
+i=1
+1
+( 4π
+3 D3
+max,i)fskyfextfrecfcl,i
+where T is the duration of the survey, N is the number of
+transients that pass the quality cut, Dmax,i is the distance out
+to which the ith transient with peak absolute magnitude Mi
+0
+5
+10
+15
+20
+25
+Balmer Decrement
+0.0
+0.2
+0.4
+0.6
+0.8
+1.0
+Cumulative Fraction of objects
+BTS Ia-CSM
+S13 Ia-CSM
+S13 IIn
+MedianBTS
+Ia
+CSM = 7.2
+MedianS13
+Ia
+CSM = 5
+MedianS13
+IIn  = 3
+Figure 12.
+Cumulative distribution of Hα/Hβ intensity ratio
+(Balmer decrement) calculated from intermediate resolution spec-
+tra of BTS SN Ia-CSM sample (grey shaded region). The red line is
+the distribution of Balmer decrement of SNe IIn measured in S13,
+the blue line is the SN Ia-CSM Balmer decrement distribution from
+S13. The black circles are a few representative points indicating the
+high Balmer decrement values and the uncertainties on them. The
+vertical dashed line is the median Balmer decrement measured from
+BTS SNe Ia-CSM.
+can be detected above the survey magnitude limit mlim (=19
+mag for BTS SNe Ia-CSM) at peak light without any extinc-
+tion, fsky is the average active survey coverage as a fraction
+of full sky, fext is average reduction in effective survey vol-
+ume due to Galactic extinction, frec is the average recovery
+efficiency for a detectable transient within the survey cover-
+age area, and fcl,i is the classification efficiency dependent
+on apparent magnitude.
+The duration of the survey in which these 12 SNe Ia-CSM
+were detected is from 2018-05-01 to 2021-05-01, i.e. T = 3
+years. We calculate fsky during this time period by averaging
+the sky area coverage of the public MSIP survey consider-
+ing 3 day cadence for ZTF Phase I (2018-05-01 to 2020-10-
+31) and 2 day cadence for ZTF Phase II (since 2020-11-01),
+which turns out to be 12505 deg2 for Phase I and 14831 deg2
+for Phase II, giving a mean fsky = 0.32. We use the same
+value of 0.82 for fext as calculated in Perley et al. (2020)
+given there has not been any change in the number and posi-
+tions of ZTF fields.
+To estimate frec, we consider SNe Ia-CSM brighter than
+−18.5 peak absolute magnitude and brighter than 18 appar-
+ent magnitude (total 5) of which 4 pass the quality cut, giving
+an frec of 0.8. We take classification completeness of 0.75 at
+
+19
+7.0
+7.5
+8.0
+8.5
+9.0
+9.5
+10.0
+10.5
+11.0
+11.5
+log M/M
+2.0
+1.5
+1.0
+0.5
+0.0
+0.5
+1.0
+1.5
+log(SFR (M
+/yr))
+sSFR = 10
+8 M
+yr
+1
+sSFR = 10
+9 M
+yr
+1
+sSFR = 10
+10 M
+yr
+1
+sSFR = 10
+11 M
+yr
+1
+sSFR = 10
+12 M
+yr
+1
+Galaxy zoo ellipticals
+Galaxy zoo spirals
+BTS SN Ia
+BTS SN Ia-CSM
+Figure 13. Host galaxies of BTS SN Ia-CSM (black circles) on
+SFR vs stellar mass plot with Galaxy-zoo spiral (blue contours) and
+elliptical (red contours) galaxies for comparison. BTS SN Ia hosts
+are also shown for comparison in green circles. Equal sSFR lines
+are marked with grey dashed lines.
+19 mag, 0.9 at 18.5 mag and 1 at 17.2 mag and linearly inter-
+polate in between these values to get fcl,i.
+Then using H0 = 70 km s−1 Mpc−1, ignoring cosmolog-
+ical effects11 as in Perley et al. (2020) and applying a uni-
+form K-correction (K = 2.5×log10(1 + z)), we get a rate of
+29.35+27.53
+−21.37 Gpc−3 yr−1 for SNe Ia-CSM. We also calculate
+a SN Ia rate of 2.88+0.28
+−0.25 × 104 Gpc−3 yr−1 from SNe Ia ob-
+served in the same period following the same method, which
+is close to the value of 2.35×104 Gpc−3 yr−1 calculated in
+Perley et al. (2020). This puts SNe Ia-CSM to be 0.02–0.2%
+of SNe Ia. However this rate estimate should be considered a
+lower limit given various caveats in the correct identification
+of SNe Ia-CSM (see discussion §4.3). If the ambiguous clas-
+sification cases outlined in Appendix A are considered to be
+SN Ia-CSM and included in the rate calculation, we obtain
+a rate upper limit of 97.7+135.8
+−77.3 Gpc−3 yr−1, which is 0.07–
+0.8% of SNe Ia.
+3.5. Precursor rates
+The ZTF precursor rates were calculated following the
+method in Strotjohann et al. (2021) which studied the fre-
+quency of precursors in interacting SNe found in ZTF.
+11 Contraction of control time window approximately compensated by
+increase in the star-formation rate density in the low redshift regime for red-
+shift dependent SN rates.
+Strotjohann et al. (2021) included 6 of the SNe Ia-CSM
+presented in this paper in addition to 4 other SNe Ia-CSM
+not in this paper (see Appendix A for details) for their search
+but did not find any robust 5σ precursor detections. This
+non-detection was concluded to be due to the small sample
+size of SNe Ia-CSM (or that they are more distant) compared
+to the SN IIn sample, so even if the precursors were as bright
+or frequent as for SNe IIn, it would be difficult to detect
+them.
+The same search was here carried out for our larger sam-
+ple by taking the ZTF forced photometry multi-band (g, r, i)
+light curves generated by the pipeline outlined in Masci
+et al. (2019) and stacking them in 1, 3 and 7-day long bins
+to search for faint outbursts. There were 7389 total avail-
+able pre-explosion epochs for BTS SNe Ia-CSM, the earliest
+epoch being 1012 days prior to the explosion and the median
+phase 340 days prior. Hence the results are valid for typical
+SN Ia-CSM progenitors at about ∼1 year before the SN. We
+did not find any robust 5σ precursor detections. The upper
+limits for the precursor rates in different bands are shown in
+Figure 14, where the solid lines indicate up to what fraction
+of the time a precursor of a given brightness could have been
+detected while being consistent with the ZTF non-detections.
+A precursor of −15 magnitude could occur as frequently as
+∼10% of the time given the ZTF non-detections. A continu-
+ous search for the precursors as more SNe Ia-CSM are found
+and classified and their sample size increases could yield a
+detection if the precursors are as frequent and bright as for
+SNe IIn. The dense and massive CSM around these objects
+is close enough to have been deposited within decades prior
+to the SN but the lack of precursors within 1 year indicates
+that there is likely no violent event that ejects a lot of mass
+in that period. Probing for precursors could potentially con-
+strain the progenitor in at least some cases. For example,
+Soker et al. (2013) predicts for their core degenerate (CD)
+model for PTF11kx-like SNe release of significant energy
+(∼1049 erg) before explosion over timescale of several years,
+implying a precursor 3–7 magnitudes fainter than the SN ex-
+plosion spread over several years, peaking in the near-IR.
+4. DISCUSSION
+4.1. Fraction of SNe Ia-CSM with delayed interaction
+The fastest declining SNe in our sample (SNe 2018crl,
+2020qxz and 2020aekp) are also the ones that develop a
+plateau and show relatively stronger SN Ia-like absorption
+features in their early spectra. They seem to have a delayed
+start for the interaction like PTF11kx but not as fast a decline,
+and thus bridge the gap between PTF11kx and the rest of the
+strongly interacting SNe Ia-CSM. It remains to be seen how
+many SNe Ia are weakly interacting where the CSM inter-
+action starts in earnest at timescales of ∼year or more after
+explosion, this requires searching for faint detections in care-
+
+20
+20
+19
+18
+17
+16
+15
+14
+13
+Absolute precursor magnitude
+10
+3
+10
+2
+10
+1
+100
+Fraction of time
+Type Ia-CSM SNe
+Figure 14.
+Precursor rate as a function of magnitude calcu-
+lated from BTS SN Ia-CSM pre-explosion ZTF forced photometry
+stacked in 7-day bins. The different colored shaded regions corre-
+spond to different ZTF bands (r-red, g-green, i-grey). The solid
+lines depict the upper limits on fraction of the time a precursor of
+the corresponding magnitude would have been detected which is
+consistent with the ZTF non-detections.
+fully calibrated forced photometry light curves (stacked to
+go fainter), a study currently undertaken by Terwel et al. (in
+prep). From the current sample, it appears that in addition to
+SNe Ia-CSM being intrinsically rare, delayed interaction SNe
+Ia-CSM are even rarer and only constitute about a quarter of
+all SNe Ia-CSM. This delayed interaction behaviour could
+also be an effect of asymmetric or clumpy CSM wherein part
+of the SN ejecta shine through depending on the viewing an-
+gle. Observational campaigns that capture the inner bound-
+ary of the CSM and the geometry robustly could shed light
+on the distribution of the inner CSM radius and reveal if it is
+a continuous distribution or if there are multiple progenitor
+scenarios within the SN Ia-CSM class.
+4.2. Implications for progenitor based on observed mass
+loss
+From Figure 10, the estimated mass-loss rates from a sim-
+ple spherical treatment of the CSM and a stationary wind lie
+between ∼ 10−3 to 10−1 M⊙ yr−1 over a period of less than
+∼ 60 years before explosion. That gives a total mass loss of
+∼ 0.1 to ∼ 1 M⊙. Dilday et al. (2012) estimated ∼ 5 M⊙
+of CSM around PTF11kx while Graham et al. (2017) revised
+it to be ∼ 0.06 M⊙. Light curve modeling of SN 1997cy
+and SN 2002ic by Chugai & Yungelson (2004) resulted in
+∼ 5 M⊙ estimates for both SNe. Inserra et al. (2016) also fit
+analytical models to some SNe Ia-CSM and found the CSM
+mass to lie between 0.4 and 4.4 M⊙. Since from Figure 5,
+the pseudo-bolometric luminosities of our SNe Ia-CSM lie
+somewhere between PTF11kx and SNe 1997cy, 2002ic and
+2005gj, with SN 1999E somewhere in the middle, we can
+say that the total CSM mass in our sample of SN Ia-CSM
+should also be several solar masses. A WD+AGB star sys-
+tem has typically been suggested for historical SNe Ia-CSM
+to explain this massive CSM. The WD could either gain
+mass through Roche Lobe overflow (RLOF) from the com-
+panion that drives an optically thick wind (OTW) or merge
+with the core of the AGB star that then explodes in or soon
+after the common envelope phase. Meng & Podsiadlowski
+(2019) model WD+MS systems for their common envelope
+wind (CEW) model and find ∼ 1 M⊙ CSM around SNe Ia-
+CSM. Thus, given the large observed CSM mass range, the
+nature of the companion cannot be solely determined from
+total mass lost. High resolution spectroscopy that can resolve
+the narrow unshocked CSM wind velocity is also needed to
+determine the compactness of the companion.
+4.3. Implications for progenitor based observed volumetric
+rate
+Robust observed rate estimates for SNe Ia-CSM have been
+few and far between. Dilday et al. (2010) found 1 interacting
+SN Ia (SN 2005gj) in a sample of 79 SNe Ia at z < 0.15
+in the SDSS-II SN survey, giving a rate of ∼1%. After the
+PTF11kx discovery in the Palomar Transient Factory (PTF)
+survey, the SN Ia-CSM rate was estimated to be ∼0.1% (1
+in 1000 classified SNe Ia; Dilday et al. 2012) but without
+spectroscopic completeness determination. S13 identified 7
+more SNe Ia-CSM from the PTF SN IIn sample, bumping
+up the estimate to ∼0.8%. With this sample we have im-
+proved the rate estimate, providing a robust value (along with
+an uncertainty estimate on that value) from an unbiased sur-
+vey with high spectroscopic completeness up to 18.5 mag-
+nitude. However this rate quite possibly still underestimates
+the true value for two reasons. The first being possible ther-
+monuclear SNe that are enshrouded so completely by CSM
+interaction that they are misclassified as SNe IIn in the ab-
+sence of good early time data. In the BTS SN IIn sample,
+we found 6 SNe IIn to have ambiguous classifications which
+could possibly be SNe Ia-CSM and these are described in
+Appendix A. Including these ambiguous cases in rate esti-
+mation results in a rate upper limit of 0.07–0.8% for strongly
+interacting thermonuclear SNe, while excluding them gives
+an underestimated rate of 0.02-0.2%.
+The second issue with the rates is if there is indeed a con-
+tinuum of delayed interaction SNe Ia-CSM like PTF11kx, in-
+teraction in SNe Ia may present itself hundreds of days later
+at magnitudes fainter than ZTF’s limit (∼20.5) resulting in
+those SNe not being counted when they may be sharing the
+same progenitor as the rest of the interacting SNe Ia-CSM.
+Lastly in some rare cases, the SN might appear normal in its
+light curve shape and duration (and thus would be missed by
+the selection criteria used in this paper) but seem to have pe-
+culiar narrow Hα in its spectrum or bright mid-IR flux (like
+in the case of SN 2020aaym; Th´evenot et al. 2021).
+
+21
+Han & Podsiadlowski (2006) predicted a rate of 0.1–1%
+for 02ic-like events for their delayed dynamical instability
+SD model but could not naturally explain the delayed interac-
+tion and multiple CSM shells in PTF11kx (which is relevant
+for some SNe in our sample). A symbiotic nova-like progen-
+itor was suggested by Dilday et al. (2012) for PTF11kx and
+they quoted the theoretical rates for the same to lie between
+1–30%, however the model could not explain the massive
+CSM. Soker et al. (2013) suggested a core degenerate (CD)
+scenario in which the explosion is set by the violent prompt
+merger of the core of the giant companion on to the WD and
+could naturally explain the massive CSM of PTF11kx (Livio
+& Riess 2003). Soker et al. (2013) estimated the occurrence
+of such SNe (Mcore+ MW D ≳ 2 M⊙ and Menv ≳ 4 M⊙)
+through population synthesis and found it to be 0.002 per
+1000 M⊙ stars formed. Assuming ∼1–2 SNe Ia occur per
+1000 M⊙ stars formed (Maoz et al. 2012), this corresponds
+to 0.1–0.2%, which compares well with our observed rate es-
+timate.
+The CEW model by Meng & Podsiadlowski (2019) pre-
+dicts that the SNe Ia-CSM like objects could arise in the SD
+CEE scenario when CONe White Dwarfs (WD) steadily ac-
+crete material at the base of the CE without quickly spiral-
+ing in due to the driving of a CEW wind (10–100 km s−1).
+The WD explodes when it reaches the Chandrasekhar mass
+(1.38 M⊙) and could possibly explode within the CE before it
+is ejected. The CEW model predicts that 25–40% of the SNe
+Ia from CONe WD in Common envelope evolution with a
+Main Sequence (MS) companion will show SN Ia-CSM like
+properties. Meng & Podsiadlowski (2019) also give the ratio
+of SNe Ia from CONe WDs to normal SNe Ia from CO WDs
+to be between 1/9 and 1/5 (considering normal SNe Ia only
+come from CO WD + MS systems). Combining that with the
+estimate that roughly 10–20% of all SNe Ia may come from
+the SD scenario (Hayden et al. 2010; Bianco et al. 2011),
+SNe Ia-CSM from CONe WD according to the CEW model
+should be 0.28% to 1.6% of all SNe Ia. A spin-down be-
+fore explosion of the WD (Justham 2011; Di Stefano & Kilic
+2012) could also explain the time delay between explosion
+and interaction.
+Soker (2022) estimated the common envelope to explo-
+sion delay time distribution (CEEDTD) shortly after the CEE
+(tCEED < 104 yr) from SN in planetary nebula rates and
+SN Ia-CSM observed rates to be roughly constant rather than
+having a t−1 dependence, that is the SN explosion could oc-
+cur very soon after the CEE as well. Our observed rates are
+on the lower side compared to these theoretical model esti-
+mates but compare well within the observational uncertain-
+ties, though the CEW model seems to best account for the
+overall SNe Ia-CSM properties.
+5. SUMMARY
+In this paper, we have presented optical and mid-IR pho-
+tometry, optical spectra and detailed analysis of 12 new SNe
+Ia-CSM identified in the Zwicky Transient Facility Bright
+Transient Survey, nearly doubling the total number of such
+objects discussed previously by Silverman et al. (2013). The
+properties of the sample extracted in this paper agree very
+well with similar analysis conducted in S13, particularly the
+median EW of Hβ is found to be significantly weaker in
+SNe Ia-CSM compared with SNe IIn and consequently the
+Balmer decrements are ubiquitously higher in SNe Ia-CSM.
+The brightness of SNe Ia-CSM in mid-IR is comparable to
+SNe IIn and observations of reduced flux in the red side of
+the Hα wing together with the mid-IR brightness points to
+formation of new dust in the cooling post-shock gas. The
+host galaxies of SNe Ia-CSM lie towards late-type galaxies
+with recent star formation. Unlike SNe IIn, no precursors
+were found within ∼1000 days before explosion for SNe Ia-
+CSM, which could be an observational bias (less number of
+SNe Ia-CSM compared to SNe IIn). We provide a robust
+rate estimate of 0.02–0.2% of all SNe Ia for SNe Ia-CSM on
+account of the BTS survey being unbiased and spectroscop-
+ically highly complete. The simple mass-loss rate estimates
+from broad Hα luminosity of ∼ 10−2 M⊙ yr−1 are similar to
+previous estimates from various methods and indicate several
+solar masses of CSM around these SNe. The observed rate
+agrees well within the observational uncertainties with the
+CEW model by Meng & Podsiadlowski (2019) which can
+also explain the interaction delay and massive CSM.
+There are still many unanswered questions about the nature
+of the progenitors and if we are accurately identifying all po-
+tential members of this class. As ZTF Phase II continues, we
+are identifying more and more SNe Ia-CSM (interacting with
+hydrogen rich and helium rich CSM) and looking further to
+the future, if ZTF continues for a Phase III and when LSST
+survey operations begins, a larger sample would further im-
+prove upon the observed rate calculation. However, individ-
+ual object studies are as important and detailed spectroscopic
+and multi-wavelength follow-up is essential to capture the
+CSM configuration and mass.
+6. ACKNOWLEDGMENT
+Based on observations obtained with the Samuel Oschin
+Telescope 48-inch and the 60-inch Telescope at the Palo-
+mar Observatory as part of the Zwicky Transient Facility
+project. ZTF is supported by the National Science Founda-
+tion under Grants No. AST-1440341 and AST-2034437 and
+a collaboration including current partners Caltech, IPAC,
+the Weizmann Institute of Science, the Oskar Klein Cen-
+ter at Stockholm University, the University of Maryland,
+Deutsches Elektronen-Synchrotron and Humboldt Univer-
+sity, the TANGO Consortium of Taiwan, the University of
+Wisconsin at Milwaukee, Trinity College Dublin, Lawrence
+
+22
+Livermore National Laboratories, IN2P3, University of
+Warwick, Ruhr University Bochum, Northwestern Uni-
+versity and former partners the University of Washington,
+Los Alamos National Laboratories, and Lawrence Berke-
+ley National Laboratories.
+Operations are conducted by
+COO, IPAC, and UW. The ZTF forced-photometry ser-
+vice was funded under the Heising-Simons Foundation grant
+#12540303 (PI: Graham). This work was supported by the
+GROWTH project (Kasliwal et al. 2019) funded by the Na-
+tional Science Foundation under PIRE Grant No 1545949.
+The Oskar Klein Centre was funded by the Swedish Re-
+search Council. Partially based on observations made with
+the Nordic Optical Telescope, operated by the Nordic Op-
+tical Telescope Scientific Association at the Observatorio
+del Roque de los Muchachos, La Palma, Spain, of the In-
+stituto de Astrofisica de Canarias.
+Some of the data pre-
+sented here were obtained with ALFOSC. Some of the data
+presented herein were obtained at the W. M. Keck Observa-
+tory, which is operated as a scientific partnership among
+the California Institute of Technology, the University of
+California, and NASA; the observatory was made possi-
+ble by the generous financial support of the W. M. Keck
+Foundation.
+The SED Machine is based upon work sup-
+ported by the National Science Foundation under Grant No.
+1106171. This work has made use of data from the Aster-
+oid Terrestrial-impact Last Alert System (ATLAS) project.
+The Asteroid Terrestrial-impact Last Alert System (ATLAS)
+project is primarily funded to search for near earth asteroids
+through NASA grants NN12AR55G, 80NSSC18K0284, and
+80NSSC18K1575; byproducts of the NEO search include
+images and catalogs from the survey area. The ATLAS sci-
+ence products have been made possible through the contri-
+butions of the University of Hawaii Institute for Astronomy,
+the Queen’s University Belfast, the Space Telescope Sci-
+ence Institute, the South African Astronomical Observatory,
+and The Millennium Institute of Astrophysics (MAS), Chile.
+This research has made use of the NASA/IPAC Infrared Sci-
+ence Archive, which is funded by the National Aeronautics
+and Space Administration and operated by the California In-
+stitute of Technology. The Liverpool Telescope is operated
+on the island of La Palma by Liverpool John Moores Univer-
+sity in the Spanish Observatorio del Roque de los Muchachos
+of the Instituto de Astrofisica de Canarias with financial sup-
+port from the UK Science and Technology Facilities Council.
+Y. Sharma thanks the LSSTC Data Science Fellowship
+Program, which is funded by LSSTC, NSF Cybertraining
+Grant #1829740, the Brinson Foundation, and the Moore
+Foundation; her participation in the program has benefited
+this work.
+S. Schulze acknowledges support from the G.R.E.A.T re-
+search environment, funded by Vetenskapsr˚adet, the Swedish
+Research Council, project number 2016-06012.
+This work has been supported by the research project
+grant “Understanding the Dynamic Universe” funded by
+the Knut and Alice Wallenberg Foundation under Dnr KAW
+2018.0067,
+The research of Y. Yang is supported through a Bengier-
+Winslow-Robertson Fellowship.
+Fritz (van der Walt et al. 2019; Duev et al. 2019) and
+GROWTH marshal (Kasliwal et al. 2019) (dynamic collab-
+orative platforms for time-domain astronomy) were used in
+this work.
+Software: LAMBDAR (Wright et al. 2016), Prospector
+(Johnson et al. 2021), pySEDM (Rigault et al. 2019), IRAF
+(Tody 1986, 1993), pyNOT (https://github.com/jkrogager/
+PyNOT), LPipe (Perley 2019), pypeit (Prochaska et al.
+2020), extinction (Barbary 2016), pyraf-dbsp (Bellm & Sesar
+2016), FPipe (Fremling et al. 2016), DBSP DRP (dbs 2021),
+ztfquery (Rigault 2018), astropy (Astropy Collaboration et al.
+2013, 2018, 2022), matplotlib (Hunter 2007) .
+REFERENCES
+2021, DBSP DRP, GitHub. https://github.com/finagle29/dbsp drp
+Ahn, C. P., Alexandroff, R., Allende Prieto, C., et al. 2012, ApJS,
+203, 21
+Aldering, G., Antilogus, P., Bailey, S., et al. 2006, ApJ, 650, 510
+Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013,
+A&A, 558, A33
+Astropy Collaboration, Price-Whelan, A. M., Sip˝ocz, B. M., et al.
+2018, AJ, 156, 123
+Astropy Collaboration, Price-Whelan, A. M., Lim, P. L., et al.
+2022, apj, 935, 167
+Barbary, K. 2016, extinction v0.3.0, Zenodo,
+doi:10.5281/zenodo.804967.
+https://doi.org/10.5281/zenodo.804967
+Barnsley, R. M., Smith, R. J., & Steele, I. A. 2012, Astronomische
+Nachrichten, 333, 101
+Bellm, E. C., & Sesar, B. 2016, pyraf-dbsp: Reduction pipeline for
+the Palomar Double Beam Spectrograph, Astrophysics Source
+Code Library, record ascl:1602.002, ,
+Bellm, E. C., Kulkarni, S. R., Barlow, T., et al. 2019, PASP, 131,
+068003
+Ben-Ami, S., Konidaris, N., Quimby, R., et al. 2012, in Proc. SPIE,
+Vol. 8446, Ground-based and Airborne Instrumentation for
+Astronomy IV, 844686. http://dx.doi.org/10.1117/12.926317
+Benetti, S., Cappellaro, E., Turatto, M., et al. 2006, ApJL, 653,
+L129
+Bianco, F. B., Howell, D. A., Sullivan, M., et al. 2011, ApJ, 741, 20
+
+23
+Blagorodnova, N., Neill, J. D., Walters, R., et al. 2018, PASP, 130,
+035003
+Blondin, S., Prieto, J. L., Patat, F., et al. 2009, ApJ, 693, 207.
+https://doi.org/10.1088/0004-637x/693/1/207
+Blondin, S., & Tonry, J. L. 2007, ApJ, 666, 1024
+Bloom, J. S., Kasen, D., Shen, K. J., et al. 2011, ApJ, 744, L17.
+https://doi.org/10.1088/2041-8205/744/2/l17
+Brown, P. J., Dawson, K. S., de Pasquale, M., et al. 2012, ApJ, 753,
+22. https://doi.org/10.1088/0004-637x/753/1/22
+Byler, N., Dalcanton, J. J., Conroy, C., & Johnson, B. D. 2017,
+ApJ, 840, 44
+Calzetti, D., Armus, L., Bohlin, R. C., et al. 2000, ApJ, 533, 682
+Cenko, S. B., Fox, D. B., Moon, D.-S., et al. 2006, PASP, 118,
+1396
+Chabrier, G. 2003, PASP, 115, 763
+Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016, arXiv
+e-prints, arXiv:1612.05560
+Chugai, N. N. 1991, MNRAS, 250, 513
+—. 2009, MNRAS, 400, 866
+Chugai, N. N., Chevalier, R. A., & Lundqvist, P. 2004, MNRAS,
+355, 627
+Chugai, N. N., & Danziger, I. J. 1994, MNRAS, 268, 173
+Chugai, N. N., & Yungelson, L. R. 2004, AstL, 30, 65
+Clark, P., Maguire, K., Bulla, M., et al. 2021, MNRAS, 507, 4367
+De, K., Hankins, M. J., Kasliwal, M. M., et al. 2020, PASP, 132,
+025001
+Dekany, R., Smith, R. M., Riddle, R., et al. 2020, PASP, 132,
+038001. https://doi.org/10.1088/1538-3873/ab4ca2
+Deng, J., Kawabata, K. S., Ohyama, Y., et al. 2004, ApJL, 605, L37
+Dhawan, S., Goobar, A., Smith, M., et al. 2022, MNRAS, 510,
+2228
+Di Stefano, R., & Kilic, M. 2012, ApJ, 759, 56
+Dilday, B., Smith, M., Bassett, B., et al. 2010, ApJ, 713, 1026
+Dilday, B., Howell, D. A., Cenko, S. B., et al. 2012, Science, 337,
+942
+Duev, D. A., Mahabal, A., Masci, F. J., et al. 2019, Monthly
+Notices of the Royal Astronomical Society, 489, 3582
+Filippenko, A. V. 1997, ARA&A, 35, 309
+Filippenko, A. V., Richmond, M. W., Matheson, T., et al. 1992,
+ApJL, 384, L15
+Fitzpatrick, E. L. 1999, PASP, 111, 63
+Foley, R. J., Smith, N., Ganeshalingam, M., et al. 2007, ApJ, 657,
+L105. https://doi.org/10.1086/513145
+Ford, C. H., Herbst, W., Richmond, M. W., et al. 1993, AJ, 106,
+1101
+F¨orster, F., Cabrera-Vives, G., Castillo-Navarrete, E., et al. 2021,
+AJ, 161, 242
+Fox, O. D., Chevalier, R. A., Skrutskie, M. F., et al. 2011, ApJ,
+741, 7
+Fox, O. D., Silverman, J. M., Filippenko, A. V., et al. 2015,
+MNRAS, 447, 772
+Fremling, C., Sollerman, J., Taddia, F., et al. 2016, A&A, 593,
+A68. https://doi.org/10.1051/0004-6361/201628275
+Fremling, C., Miller, A. A., Sharma, Y., et al. 2020, ApJ, 895, 32.
+https://doi.org/10.3847/1538-4357/ab8943
+Gal-Yam, A., & Leonard, D. C. 2009, Nature, 458, 865
+Gal-Yam, A., Yaron, O., Pastorello, A., et al. 2021, TNSAN, 76, 1
+Gal-Yam, A., Leonard, D. C., Fox, D. B., et al. 2007, ApJ, 656, 372
+Gal-Yam, A., Bruch, R., Schulze, S., et al. 2022, Nature, 601, 201
+Germany, L. M., Reiss, D. J., Sadler, E. M., Schmidt, B. P., &
+Stubbs, C. W. 2000, ApJ, 533, 320
+Graham, M. J., Kulkarni, S. R., Bellm, E. C., et al. 2019, PASP,
+131, 078001
+Graham, M. L., Harris, C. E., Fox, O. D., et al. 2017, ApJ, 843, 102
+Graham, M. L., Fremling, C., Perley, D. A., et al. 2022, MNRAS,
+511, 241
+Guillochon, J., Parrent, J., Kelley, L. Z., & Margutti, R. 2017, ApJ,
+835, 64
+Hamuy, M., Phillips, M. M., Suntzeff, N. B., et al. 2003, Nature,
+424, 651
+Han, Z., & Podsiadlowski, P. 2006, MNRAS, 368, 1095
+Hayden, B. T., Garnavich, P. M., Kasen, D., et al. 2010, ApJ, 722,
+1691
+Hodgkin, S. T., Harrison, D. L., Breedt, E., et al. 2021, A&A, 652,
+A76
+Hosseinzadeh, G., Arcavi, I., Valenti, S., et al. 2017, ApJ, 836, 158.
+https://doi.org/10.3847/1538-4357/836/2/158
+Hunter, J. D. 2007, Computing in Science & Engineering, 9, 90
+Iben, I., J., & Tutukov, A. V. 1984, ApJS, 54, 335
+Inserra, C., Smartt, S. J., Scalzo, R., et al. 2014, MNRAS, 437, L51
+Inserra, C., Fraser, M., Smartt, S. J., et al. 2016, MNRAS, 459,
+2721
+Irani, I., Prentice, S. J., Schulze, S., et al. 2022, ApJ, 927, 10
+IRSA. 2022, Zwicky Transient Facility Image Service, IPAC,
+doi:10.26131/IRSA539. https:
+//catcopy.ipac.caltech.edu/dois/doi.php?id=10.26131/IRSA539
+Johnson, B. D., Leja, J., Conroy, C., & Speagle, J. S. 2021, ApJS,
+254, 22
+Justham, S. 2011, ApJL, 730, L34
+Kasliwal, M. M., Cannella, C., Bagdasaryan, A., et al. 2019,
+Publications of the Astronomical Society of the Pacific, 131,
+038003. https://doi.org/10.1088%2F1538-3873%2Faafbc2
+Kiewe, M., Gal-Yam, A., Arcavi, I., et al. 2012, ApJ, 744, 10
+Kochanek, C. S., Shappee, B. J., Stanek, K. Z., et al. 2017, PASP,
+129, 104502
+Kool, E., Johansson, J., Sollerman, J., & et al. 2022, Nature
+Astronomy
+
+24
+Kotak, R., & Meikle, W. P. S. 2005, in Astronomical Society of the
+Pacific Conference Series, Vol. 342, 1604-2004: Supernovae as
+Cosmological Lighthouses, ed. M. Turatto, S. Benetti,
+L. Zampieri, & W. Shea, 264
+Lang, D. 2014a, AJ, 147, 108
+—. 2014b, AJ, 147, 108
+Leloudas, G., Hsiao, E. Y., Johansson, J., et al. 2015, A&A, 574,
+A61
+Li, W., Bloom, J. S., Podsiadlowski, P., et al. 2011, Nature, 480,
+348
+Lintott, C., Schawinski, K., Bamford, S., et al. 2011, MNRAS,
+410, 166
+Livio, M., & Riess, A. G. 2003, ApJL, 594, L93
+Maguire, K., Sullivan, M., Patat, F., et al. 2013, MNRAS, 436, 222
+Mainzer, A., Bauer, J., Cutri, R. M., et al. 2014, ApJ, 792, 30
+Maoz, D., Mannucci, F., & Brandt, T. D. 2012, MNRAS, 426, 3282
+Martin, D. C., Fanson, J., Schiminovich, D., et al. 2005, ApJL,
+619, L1
+Masci, F. J., Laher, R. R., Rusholme, B., et al. 2019, PASP, 131,
+018003
+Meisner, A. M., Lang, D., & Schlegel, D. J. 2018, AJ, 156, 69
+Meng, X., & Podsiadlowski, P. 2019, IAU Symposium, 343, 470
+Miller, A. A., Yao, Y., Bulla, M., et al. 2020, ApJ, 902, 47
+Miller, J., & Stone, R. 1987, The CCD Cassegrain Spectrograph at
+the Shane Reflector, Santa Cruz: Lick Observatory Techical
+Reports No. 48
+Nicholls, B., & Dong, S. 2018, TNSCR, 2018-1144, 1
+Nordin, J., Brinnel, V., van Santen, J., et al. 2019, A&A, 631,
+A147. http://dx.doi.org/10.1051/0004-6361/201935634
+Nugent, P. E., Sullivan, M., Cenko, S. B., et al. 2011, Nature, 480,
+344
+Oke, J. B., & Gunn, J. E. 1982, PASP, 94, 586
+Oke, J. B., Cohen, J. G., Carr, M., et al. 1995, PASP, 107, 375
+Pastorello, A., Mattila, S., Zampieri, L., et al. 2008, MNRAS, 389,
+113
+Patat, F., Chandra, P., Chevalier, R., et al. 2007, Science, 317, 924
+Payne, A. V., Tucker, M. A., Do, A., Shappee, B. J., & Huber,
+M. E. 2019, TNSCR, 2019-343, 1
+Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, JMLR, 12,
+2825
+Pellegrino, C., Howell, D. A., Terreran, G., et al. 2022, arXiv
+e-prints, arXiv:2205.07894
+Perley, D. A. 2019, PASP, 131, 084503
+Perley, D. A., Fremling, C., Sollerman, J., et al. 2020, ApJ, 904,
+35. http://dx.doi.org/10.3847/1538-4357/abbd98
+Perley, D. A., Sollerman, J., Schulze, S., et al. 2022, ApJ, 927, 180
+Piascik, A. S., Steele, I. A., Bates, S. D., et al. 2014, in Proc. SPIE,
+Vol. 9147, Ground-based and Airborne Instrumentation for
+Astronomy V, 91478H
+Prentice, S., Maguire, K., Magee, M. R., & Deckers, M. 2020,
+TNSCR, 2020-2285, 1
+Prieto, J. L., Garnavich, P. M., Phillips, M. M., et al. 2007, arXiv
+e-prints, arXiv:0706.4088
+Prochaska, J. X., Hennawi, J. F., Westfall, K. B., et al. 2020,
+Journal of Open Source Software, 5, 2308.
+https://doi.org/10.21105/joss.02308
+Reguitti, A. 2020, TNSCR, 2020-3825, 1
+Rigault, M. 2018, ztfquery, a python tool to access ZTF data,
+Zenodo, vdoi, Zenodo, doi:10.5281/zenodo.1345222
+Rigault, M., Neill, J. D., Blagorodnova, N., et al. 2019, A&A, 627,
+A115
+Salamanca, I., Cid-Fernandes, R., Tenorio-Tagle, G., et al. 1998,
+MNRAS, 300, L17
+Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103.
+https://doi.org/10.1088/0004-637x/737/2/103
+Schlegel, E. M. 1990, MNRAS, 244, 269
+Schmidt, B. P., Kirshner, R. P., Leibundgut, B., et al. 1994, ApJL,
+434, L19
+Schulze, S., Yaron, O., Sollerman, J., et al. 2021, ApJS, 255, 29
+Shappee, B., Prieto, J., Stanek, K. Z., et al. 2014, in AAS Meeting
+Abstracts, Vol. 223, 236.03
+Silverman, J. M., Nugent, P. E., Gal-Yam, A., et al. 2013, ApJS,
+207, 3. https://doi.org/10.1088/0067-0049/207/1/3
+Silverman, J. M., Nugent, P. E., Gal-Yam, A., et al. 2013, ApJ, 772,
+125
+Simon, J. D., Gal-Yam, A., Gnat, O., et al. 2009, ApJ, 702, 1157.
+https://doi.org/10.1088/0004-637x/702/2/1157
+Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131,
+1163
+Smartt, S. J., Valenti, S., Fraser, M., et al. 2015, A&A, 579, A40
+Smith, K. W., Smartt, S. J., Young, D. R., et al. 2020, PASP, 132,
+085002
+Smith, N. 2014, ARA&A, 52, 487.
+https://doi.org/10.1146/annurev-astro-081913-040025
+Smith, N., Chornock, R., Li, W., et al. 2008, ApJ, 686, 467
+Soker, N. 2022, RAA, 22, 035025
+Soker, N., Kashi, A., Garc´ıa-Berro, E., Torres, S., & Camacho, J.
+2013, MNRAS, 431, 1541
+Soraisam, M., Matheson, T., & Lee, C.-H. 2021, RNAAS, 5, 62
+Soumagnac, M. T., & Ofek, E. O. 2018, PASP, 130, 075002
+Steele, I. A., Smith, R. J., Rees, P. C., et al. 2004, in Society of
+Photo-Optical Instrumentation Engineers (SPIE) Conference
+Series, Vol. 5489, Ground-based Telescopes, ed. J. Oschmann,
+Jacobus M., 679–692
+Stein, R., Callis, E., Kostrzewa-rutkowska, Z., Fraser, M., &
+Yaron, O. 2018a, TNSCR, 2018-1168, 1
+—. 2018b, TNSCR, 2018-1168, 1
+Sternberg, A., Gal-Yam, A., Simon, J. D., et al. 2011, Science, 333,
+856
+
+25
+Strotjohann, N. L., Ofek, E. O., Gal-Yam, A., et al. 2021, ApJ, 907,
+99
+Taddia, F., Stritzinger, M. D., Sollerman, J., et al. 2013, A&A, 555,
+A10
+Taggart, K. 2020, PhD thesis, Liverpool John Moores University,
+doi:10.24377/LJMU.t.00014104.
+http://researchonline.ljmu.ac.uk/id/eprint/14104/
+Taubenberger, S., Kromer, M., Hachinger, S., et al. 2013, MNRAS,
+432, 3117
+Th´evenot, M., Kabatnik, M., & Gantier, J. M. 2021, Transient
+Name Server AstroNote, 212, 1
+Tody, D. 1986, in Proc. SPIE, Vol. 627, Instrumentation in
+astronomy VI, ed. D. L. Crawford, 733
+Tody, D. 1993, in Astronomical Society of the Pacific Conference
+Series, Vol. 52, Astronomical Data Analysis Software and
+Systems II, ed. R. J. Hanisch, R. J. V. Brissenden, & J. Barnes,
+173
+Tonry, J. L., Denneau, L., Heinze, A. N., et al. 2018, PASP, 130,
+064505
+Tucker, M. A., Huber, M., Shappee, B. J., et al. 2018, The
+Astronomer’s Telegram, 11444, 1
+van der Walt, S. J., Crellin-Quick, A., & Bloom, J. S. 2019, Journal
+of Open Source Software, 4, doi:10.21105/joss.01247.
+http://joss.theoj.org/papers/10.21105/joss.01247
+Wang, L., Baade, D., Hflich, P., et al. 2004, ApJ, 604, L53.
+https://doi.org/10.1086/383411
+Webbink, R. F. 1984, ApJ, 277, 355
+Whelan, J., & Iben, Icko, J. 1973, ApJ, 186, 1007
+Wood-Vasey, W. M., Wang, L., & Aldering, G. 2004, ApJ, 616,
+339. https://doi.org/10.1086/424826
+Wright, A. H., Robotham, A. S. G., Bourne, N., et al. 2016,
+MNRAS, 460, 765
+Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010a,
+AJ, 140, 1868
+Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010b,
+AJ, 140, 1868
+Xu, Y., McCray, R., Oliva, E., & Randich, S. 1992, ApJ, 386, 181
+Yang, Y., Baade, D., Hoeflich, P., et al. 2022, MNRAS,
+arXiv:2211.04423
+Yao, Y., Miller, A. A., Kulkarni, S. R., et al. 2019, ApJ, 886, 152
+Yaron, O., & Gal-Yam, A. 2012, PASP, 124, 668
+Zackay, B., Ofek, E. O., & Gal-Yam, A. 2016, ApJ, 830, 27
+
+26
+APPENDIX
+A. AMBIGUOUS SN IA-CSM/IIN IN BTS
+To identify potential SNe Ia-CSM hiding in the SN IIn sample classified by BTS, we rechecked all SNe IIn classifications (total
+142) using SuperNova IDentification (SNID; Blondin & Tonry 2007) software. SNe IIn spectra were processed through SNID,
+and any SN having ≥ 3 matches to a SN Ia-CSM in the top 10 matches were manually checked. The SNe having ambiguous
+classifications are described below.
+A.1. SN 2019smj
+Discovered by ZTF and reported to TNS by ALeRCE (F¨orster et al. 2021) on 2019-10-13 11:28:42.000, SN 2019smj
+(ZTF19aceqlxc) was classified as a Type IIn by BTS at z = 0.06. It peaked at apparent magnitude 17.1 in r band (∼ −20.1)
+and then developed a weaker but broader bump. The spectra showed very weak Hβ, barely any He I λ5876, no O I λ7774
+or [O I] lines but showed some iron group lines, Ca NIR emission and [Ca II]. SNID best matches were to SNe 1997cy and
+2005gj. The early spectra from P60/SEDM have some matches to SN 2005gj but are too noisy and of ultra-low resolution to
+conclusively provide a Ia-CSM classification. From these observations, SN 2019smj is most likely a Type Ia-CSM but given the
+lack of confirmation we have excluded it from the main sample.
+A.2. SN 2018dfa
+Discovered and reported to TNS by ATLAS on 2018-07-05 08:51:21.000, SN 2018dfa was classified initially as a Type IIP
+by BTS but later spectra revealed it to be a Type IIn at z = 0.128. It peaked at apparent magnitude of 17.5 in r band (−20.2)
+and showed a minor bump before main peak in the light curve. The spectra showed weak Hβ and He I λ5876, no O I λ7774 or
+[O I] lines. SNID best matches were to SNe 2002ic and 2005gj along with SNe Ia-norm/91T. The earliest spectra with good SNR
+from P200/DBSP had one match to SN 2005gj but could not provide a robust Ia-CSM classification. From these observations,
+SN 2018dfa is most likely a Type Ia-CSM but given the lack of confirmation we have excluded it from the main sample.
+A.3. SN 2019vpk
+Discovered by ZTF and reported to TNS by ALeRCE on 2019-11-25 06:33:38.000, SN 2019vpk was classified as a Type IIn
+by BTS at z = 0.1. It peaked at apparent magnitude of ∼ 18 in r band (∼ −20.5). The early spectra were too noisy and the
+only spectrum with good SNR was obtained with P200/DBSP nearly 6 weeks after discovery which showed weak Hβ, no clear
+He I emission but possibly Si II λ5958 emission (which is unlike any other SN Ia-CSM). SNID top matches were to SN 2005gj
+but visually did not look entirely convincing, and some matches were also to Type IIn. We conclude SN 2019vpk does not have
+enough data for a robust Ia-CSM classification.
+A.4. SN 2019wma
+Discovered by ZTF and reported to TNS by ALeRCE on 2019-12-13 13:35:26.000, SN 2019wma was classified as a Type IIn
+by BTS at z = 0.088. It peaked at apparent magnitude of ∼ 18.5 in r band (∼ −19.5). The spectra obtained were either from
+P60/SEDM or LT/SPRAT hence of low resolution and showed weak Hβ and He I emission. SNID top matches to earliest SEDM
+spectrum were to SN 2005gj at the correct redshift but given the lack of intermediate resolution spectra and absence of late time
+follow-up we did not assign a Type Ia-CSM classification to SN 2019wma and excluded it from the main sample.
+A.5. SN 2019kep
+Discovered and reported to TNS by ATLAS on 2019-07-02 14:13:55.000, SN 2019kep was classified as a Type IIn by BTS
+at z = 0.02388. It peaked at apparent magnitude of 18.2 in r band (−17). Most early spectra were too noisy for classification
+but matched to SN 2005gj. A good SNR P200/DBSP spectrum showed narrow P-Cygni Hα with absorption minimum at ∼
+2500 km s−1 but overall matched to a Type II SN. From these observations, we could not determine a robust classification for SN
+2019kep and excluded it from the main sample.
+A.6. SN 2018ctj
+Discovered and reported to TNS by ZTF on 2018-04-21 08:36:57.000, SN 2018ctj was classified as a Type IIn by BTS at
+z = 0.0378. It peaked at apparent magnitude of 18.4 in r band (−17.8) and was also detected in unWISE data. Only one
+P60/SEDM spectrum was obtained that matched well to SNe 1997cy and 2005gj. Given the lack of intermediate resolution
+spectra this SN remains classified as Type IIn and excluded from the main sample.
+