The Lick Observatory Supernova Search follow-up program: photometry data release of 70 SESNe

ABSTRACT

1 INTRODUCTION

It is well established that massive stars (i.e. those having M ≳ 8 M_⊙) have short lives that end in catastrophic explosions known as core-collapse supernovae (CCSNe). Among CCSNe, those whose spectra show features of hydrogen are classified as Type II SNe (see, e.g. Filippenko 1997; Gal-Yam 2017 for reviews of SN classification). In contrast, hydrogen-poor CCSNe are classified as Type Ib or Ic, depending on whether their optical spectra contain obvious helium features (Matheson 2001). The progenitor stars of hydrogen-poor CCSNe have their outer envelopes stripped away before explosion by strong winds during the Wolf–Rayet phase (e.g. Conti 1975; Smith & Owocki 2006; Gal-Yam et al. 2014), By interaction with a binary companion (e.g. Podsiadlowski, Joss & Hsu 1992; Sana et al. 2012; Eldridge et al. 2013), or some combination of these two modes. If the envelope-stripping process is highly efficient, the helium shell is also removed before explosion, leading to the differentiation between SNe Ib (He-rich) and SNe Ic (He-poor). Hydrogen-poor CCSNe are generally referred to as stripped-envelope supernovae (SESNe).

SESNe are found to be observationally heterogeneous. For example, in some cases, the stripping process is incomplete and thus the envelope is left with some fraction of hydrogen. These SNe typically show H lines at early times that rapidly disappear after maximum light (Filippenko 1988; Filippenko, Matheson & Ho 1993), and their spectra resemble SNe Ib at late times. A small subset of SN Ib-like events show evidence of interaction with dense circumstellar material (CSM); having relatively narrow spectral emission lines (e.g. Foley et al. 2007; Pastorello et al. 2007; Hosseinzadeh et al. 2017), these objects have been dubbed SNe Ibn (‘n’ for ‘narrow lines’). In addition, a subset of SNe Ic characterised by the presence broad spectral lines which indicate extremely high ejecta velocities (≳ 15 000 km s⁻¹) are designated as SNe Ic-BL (e.g. Modjaz et al. 2014). Objects within this subclass have been found to be associated with long-duration gamma-ray bursts (e.g. Woosley & Bloom 2006). Recently, a new class, SNe Icn, has been proposed by Gal-Yam et al. (2021) based on the prototype SN 2019hgp, followed by SNe 2021csp (Fraser et al. 2021; Perley et al. 2022) and 2021ckj (Pastorello et al. 2021). The early-time spectra of these objects are dominated by narrow lines with profiles similar to those seen in SNe Ibn, but originating from carbon and oxygen rather than He. In any case, all of the aforementioned SN classifications (Ib, IIb, Ibn, Ic, Ic-BL, and Icn) are, to some extent, related to the envelope of their progenitor star being stripped. Accordingly, we consider all of them to be SESNe in the analysis presented herein.

Owing to the efforts of various SN surveys spanning the globe, the study of SESNe with large light-curve samples has proliferated. Li et al. (2011) presented a set of roughly two dozen unfiltered SESN light curves within ∼60 Mpc. In the same year, Drout et al. (2011) presented V- and R-band light curves of 25 SESNe from the Palomar 60 inch telescope. Bianco et al. (2014) published multiband light curves of 64 SESNe obtained by the Harvard-Smithsonian Center for Astrophysics (CfA) SN group, and Taddia et al. (2015) presented expanded sets of multiband light curves of 20 SESNe from the Sloan Digital Sky Survey (SDSS) SN Survey II. In addition, Stritzinger et al. (2018a) published 34 SESN light curves from the first phase of the Carnegie Supernova Project (CSP-I). Despite these impressive efforts, the state of large-scale photometric studies of SESNe substantially lags that of other SNe (e.g. SNe Ia, which are routinely studied photometrically at the hundreds-of-objects scale).

Over the past two decades, our Lick Observatory Supernova Search (LOSS; Filippenko et al. 2001) program has invested considerably in both discovering and monitoring all kinds of SNe, including SESNe. Large light-curve samples have already been published by Ganeshalingam et al. (2010) and Stahl et al. (2019) for SNe Ia, and by de Jaeger et al. (2019) for SNe II. In this paper, we release the light curves of 70 SESNe observed by LOSS since 2003. In the remainder of the paper, we describe the sample (Section 2) and our data-reduction strategies (Section 3) before presenting an analysis of the light curves (Section 4) and offering our conclusions (Section 5).

2 DATA SAMPLE

The Berkeley SESN sample consists of 70 objects observed between 2003 and 2020. Two main telescopes were used for follow-up observations: (i) the fully robotic 0.76-m Katzman Automatic Imaging Telescope (KAIT; Filippenko et al. 2001), and (ii) the 1-m Anna Nickel telescope, both located at Lick Observatory on Mount Hamilton, near San Jose, CA, USA. Most SESNe in our sample were observed in multiple optical passbands (B, V, R, I), and some have additional Clear-band (unfiltered) data.¹ For a large fraction of SESNe in our sample, spectra were also obtained by our group using multiple facilities. A detailed analysis and release of the LOSS spectra of SESNe was published by Shivvers et al. (2019), so the present paper focuses exclusively on our photometric observations.

Table 1 provides the basic information for each SN in our sample, including its spectroscopic classification, host-galaxy name, distance, recession velocity, and Galactic extinction (Schlafly & Finkbeiner 2011). The listed classifications are adopted from the Transient Name Server² (TNS), but where they conflict with those given by Shivvers et al. (2019), we adopt the latter owing to the more sophisticated and systematic approach that resulted in their determinations. The single exception to this paradigm is SN 2008fz, which was classified as an SN Ic by Shivvers et al. (2019) but for which we adopt the classification of superluminous SN IIn (i.e. SLSN-IIn) from Drake et al. (2010). In summary, our SESN sample consists of 19 spectroscopically normal SNe Ib, 2 peculiar SNe Ib, six SNe Ibn, 14 normal SNe Ic, 1 peculiar SN Ic, 10 SNe Ic-BL, 15 SNe IIb, 1 ambiguous SN IIb/Ib/c, and 2 SLSNe.

The host-galaxy distances reported in Table 1 were obtained from the NASA/IPAC Extragalactic Database of redshift-independent distances (NED-D³) when available, and otherwise calculated using a standard cosmological model with H₀  = 70 km s⁻¹ Mpc⁻¹, Ω_M = 0.30, and Ω_Λ = 0.70. Fig. 1 shows the redshift distribution of the 65 SESNe in our sample which have reliable host-galaxy heliocentric velocity measurements. The redshifts range from 0.0015 (SN 2011dh) to 0.0350 (SN 2009er), with an average value of 0.0136 and a standard deviation of 0.0082. In total, 40 SESNe have z > 0.01. For the remaining five SESNe without host heliocentric velocity measurements, we adopt redshifts from various literature sources (see the footnote of Table 1 for more details).

Figure 1.

The redshift distribution of the 65 (out of 70 in total) SESNe in our sample which have reliable heliocentric velocity measurements from their host galaxies. The average redshift value is 0.0136 with a standard deviation of 0.0082, and 40 SESNe have z > 0.01.

Open in new tabDownload slide

3 DATA REDUCTION

All photometric data published herein were obtained with the Lick KAIT and 1-m Nickel telescopes. Over the past two decades, the KAIT CCD and filter have been updated several times, resulting in four different CCD/filter combinations, which we refer as KAIT1-4 (see Ganeshalingam et al. 2010; Stahl et al. 2019 for more details). In its current configuration (KAIT4), KAIT is equipped with a Finger Lakes Instrument camera with 512 × 512 pixels covering a 6.7 × 6.7 arcmin² field of view. The 1 -m Nickel is a human-operated telescope, but it can be remotely operated from the UC Berkeley campus. It is equipped with a thinned, Loral, 2048 × 2048 pixel CCD (binned by a factor of two along both axes to reduce the readout time), and has a |$6{.}3 \times 6{.}3{\,\rm arcmin}^{2}$| field of view. The filter set on the Nickel telescope was updated once, in 2009 March; we refer to the configuration before and after this change as Nickel1 and Nickel2, respectively (again, see Ganeshalingam et al. 2010; Stahl et al. 2019 for more details).

A novel automated photometry pipeline⁴ was developed by Stahl et al. (2019) to process the significant volume of SN observations produced by LOSS consistently and accurately, while at the same time requiring minimal human intervention. Although we defer the details to Stahl et al. (2019), we briefly summarise the main procedures here. All images are first treated to remove the bias level, and are then flat-fielded before being astrometrically calibrated using code provided by astrometry.net⁵ (Lang et al. 2010). Where necessary, image subtraction is applied so as to remove host-galaxy contamination, with the template images being sourced on dark nights using the Nickel telescope after the SNe have faded beyond detection (generally >6 months after discovery). Point-spread-function (PSF) photometry is obtained using DAOPHOT (Stetson 1987) from the IDL Astronomy User’s Library.⁶ Nearby stars are chosen from the Pan-STARRS1⁷ catalogue for calibration. Their magnitudes are first transformed into the Landolt system (Landolt 1983, 1992) using the empirical prescription (eq. 6) presented by Tonry et al. (2012), and then transformed to the appropriate KAIT/Nickel natural system (i.e. KAIT1-4 or Nickel1-2 as appropriate based on the equipment configuration on the date of observation). All apparent magnitudes are measured in the natural system, and the final results are then transformed to the standard system (see equation 1 a–d of Stahl et al. 2019) using local calibrators and the appropriate colour terms as given by Ganeshalingam et al. (2010) and Stahl et al. (2019). Note that when transforming from the natural system back to the standard system, there are additional errors associated with the transformation that are not accounted for owing to differences between the spectral energy distributions of SNe and the reference stars (e.g. Stritzinger et al. 2005).

4 RESULTS

4.1 Photometry data release

We provide our final photometry in Tables 2 (standard system; see full Table S4 in the supplementary material) and 3 (natural system; see full Table S5 in the supplementary material) for all 70 SESNe in our sample. On average, each SN has 81 observations (median of 58) and 22 photometric points (median of 13 points) per filter, at a cadence of 3.6 d (median of 1.2 d). SN 2016coi has the best coverage (434 points), followed by SN 2015ap (351). The total number of distinct photometric observations published in this work is 5682. Of the 70 SNe in our sample, 38 have pre-maximum coverage and thus have their peak brightness measured in at least one band. Note that although several SNe in our sample have had their photometry previously published in individual papers, the magnitudes presented herein supersede these earlier measurements because (i) for some fields requiring image subtraction we have obtained new, higher quality templates, and (ii) better calibration sources are now available that were not used previously. Moreover, our processing is now significantly more systematic and self-consistent owing to our use of the battle-tested LOSSPhotPypeline (de Jaeger et al. 2019; Stahl et al. 2019; Stahl et al. 2020)

Note that in this release, we do not include the systematic uncertainty of 0.03 mag in BVRI that was determined and discussed in detail by Stahl et al. (2019). This amount of systematic uncertainty was estimated by investigating many factors that may contribute to the error, including evolution of colour terms, evolution of atmospheric terms, configurations between different telescopes, and galaxy-subtraction procedures. None of these factors contributed uncertainty over 0.03 mag, consistent with the estimate of Ganeshalingam et al. (2010). Though not included in our photometry tables (Table 2 and 3) or light-curve figures (e.g. Fig. 2 and Fig. S1 in the supplementary material, this uncertainty must be accounted for when combining our data set with others. Alongside the recent LOSS photometry release for SNe Ia (Stahl et al. 2019) and SNe II (de Jaeger et al. 2019), we aim for our SESN photometry to be used and further analysed by the astronomical community.

Open in new tabDownload slide

Open in new tabDownload slide

Open in new tabDownload slide

Open in new tabDownload slide

Figure 2.

Apparent-magnitude light curves of the SESNe in our sample in the standard system (except for SN 2016P, which has data only in the natural system as shown in Fig. S1 in the supplementary material, without any corrections for extinction. Blue triangles are magnitudes in B, green diamonds are V, red squares are R, black circles are Clear, and dark-red inverted triangles are I. All dates have been shifted relative to the time of maximum V-band brightness if determined, and relative to the time of the first epoch otherwise. In each panel, the IAU name and the type are given.

4.2 Light curves

Fig. 2 shows the apparent-magnitude light curves of all SESNe from our sample in the standard Landolt system without any extinction corrections applied. Note that we also include Clear-band light curves where available. Although unfiltered and thus non-standard, it is most similar to the R band (Li et al. 2003). The temporal axes are all in the observer frame and shifted such that times are measured relative to the times of maximum V-band brightness as determined by fitting the near-maximum data with low-order Legendre polynomials. In the 37 cases, where no maximum could be found via this method, the temporal axes are shifted relative to the time of the first observation. Such fitting was also applied to other bands if the data – after being supplemented with corresponding observations from Drout et al. (2011), Bianco et al. (2014), or Stritzinger et al. (2018a) – had sufficient near-maximum coverage.

4.3 Colour evolution and host extinction

Fig. 3 shows the (B − V) (top left-hand panel) and (V − R) (top right-hand panel) colour evolution of the SESNe in our sample which have the requisite observations in both bands, after correcting for Milky Way (MW) Galactic extinction (Schlafly & Finkbeiner 2011) but not host-galaxy extinction. Individually, we find that the colours trend blue at very early times (e.g. ∼10 d before V-band peak), and then become progressively more red until ∼20 d after V-band maximum. As a whole, however, the colours we observe span a wide range at nearly every epoch, in part because we have not (yet) performed any corrections for host extinction. As SESNe often reside in dusty star-forming regions (e.g. Van Dyk, Hamuy & Filippenko 1996; Kelly, Kirshner & Pahre 2008), it is likely that the extinction due to the galactic hosts of our SNe is generally the dominant component of the total line-of-sight extinction.

Figure 3.

Left-hand panel: (B − V) colour evolution of the SESNe in our sample having simultaneous observations in both bands, after correction for Galactic extinction (top panel), and with further correction for host extinction (bottom panel) using the empirical colour method (see the text for more details). Right-hand panel: similar to the left-hand panel, but for the (V − R) colour.

Open in new tabDownload slide

One useful way to estimate host extinction is to exploit its relationship with the equivalent width of Na I D absorption as measured from high-resolution spectra (e.g. Poznanski, Prochaska & Bloom 2012; Stritzinger et al. 2018b). Unfortunately, such data are difficult to obtain, and low-resolution spectra, though easier to procure, are usually not of sufficient quality for such measurements (Poznanski et al. 2011).

Distinct from spectral proxies using Na I D absorption, Drout et al. (2011) found that the (V − R) colour evolution of SESNe has a small dispersion at ∼10 d after V-band peak, and thus can be exploited as a useful diagnostic for estimating host-galaxy extinction (Drout et al. 2011; Taddia et al. 2015; Stritzinger et al. 2018b). We therefore follow the approach of Drout et al. (2011) to estimate the extinction induced by the host galaxies of the SNe in our sample. To do this, we measure the (V − R) colour at 10 d after V-band maximum, and apply extra extinction (assumed to be from the host) so that the (V − R) colour reaches 0.26 mag – the mean (V − R) colour found by Drout et al. (2011) after all corrections. We successfully applied this method to 31 SESNe in our sample with available data in both V and R at 10 d after V-band maximum. We note that a few SESNe in our measured sample have negative implied host-extinction values by doing this, but such values are sufficiently small to be consistent with no host extinction.

Fig. 4 shows a histogram of the estimated E(B − V) values for the 31 SESNe in our sample (also listed in Table 4) assuming a Cardelli, Clayton & Mathis (1989) reddening law with R_V = 3.1. It clearly shows that SESNe usually suffer moderately-high extinction from their host galaxies, with a mean E(B − V) value of 0.32 mag and standard deviation of 0.19 mag – consistent with the value found by other works using different SESN samples (Drout et al. 2011; Taddia et al. 2015; Stritzinger et al. 2018b). This mean host extinction is higher than recently studied samples of SNe Ia with mean E(B − V) ≈ 0.11 mag (Burns et al. 2011; Stahl et al. 2019) and SNe II with mean A_V ≈ 0.9 mag (E(B − V) ≈ 0.29; Smartt et al. 2009).

Figure 4.

Distribution of host-galaxy extinction values for the 31 SESNe in our sample (solid red), revealing a mean E(B − V) value of 0.32 mag with a standard deviation of 0.19 mag. The corresponding MW extinction is shown in dashed black.

Open in new tabDownload slide