A Snapshot Survey of Nearby Supernovae with the Hubble Space Telescope

Over recent decades, robotic (or highly automated) searches for supernovae (SNe) have discovered several thousand events, many of them in quite nearby galaxies (distances < 30 Mpc). Most of these SNe, including some of the best-studied events to date, were found before maximum brightness and have associated with them extensive follow-up photometry and spectroscopy. Some of these discoveries are so-called “SN impostors,” thought to be superoutbursts of luminous blue variable stars, although possibly a new, weak class of massive-star explosions. We conducted a Snapshot program with the Hubble Space Telescope (HST) and obtained images of the sites of 31 SNe and four impostors, to acquire late-time photometry through two filters. The primary aim of this project was to reveal the origin of any lingering energy for each event, whether it is the result of radioactive decay or, in some cases, ongoing late-time interaction of the SN shock with preexisting circumstellar matter, or the presence of a light echo. Alternatively, lingering faint light at the SN position may arise from an underlying stellar population (e.g., a host star cluster, companion star, or a chance alignment). The results from this study complement and extend those from Snapshot programs by various investigators in previous HST cycles.

1. INTRODUCTION Supernovae (SNe) represent the final, explosive stage in the evolution of certain varieties of stars (e.g., Woosley & Weaver 1986;Wheeler & Harkness 1990;Filippenko 1997;Gal-Yam 2017).Studies of SNe, both observational and theoretical, are closely tied with the physics of stellar evolution, explosion mechanisms and nucleosynthesis, the chemical evolution of galaxies and the Universe, the formation of neutron stars and black holes, and gamma-ray bursts.SNe Ia are also exceedingly useful cosmological tools, revealing the accelerating expansion of the Universe.
Despite the well-sampled early-time light curves of relatively nearby SNe, observations are quite sparse at late times (t ≳ 6 months), primarily because the SNe are extremely faint or their ground-based photometry is contaminated by neighboring stars within the seeing disk.Thus, high-spatial-resolution observations, such as with the Hubble Space Telescope (HST), are required to obtain accurate photometry.Following the multifilter light-curve shapes of these SNe over their long evolution provides important information on their progenitor systems and on the underlying physics leading to the lingering light, and can reveal "SN impostors" -events which are not genuine SNe involving a terminal explosion, but instead are powerful stellar outbursts which occasionally approach the peak luminosity of some kinds of true SNe.
There have been a number of HST Snapshot surveys of the sites of SNe.Various studies have conducted detailed analyses of the late-time emission of these SNe and of their immediate environments (e.g., Li et al. 2002;Fransson et al. 2002;Van Dyk et al. 2003;Sun et al. 2023).In each case it has been shown that HST can effectively resolve the faint SNe at late times from their immediate environments.In some cases, more than one epoch of HST observations was obtained, enabling the measurements of late-time decline rates and providing important information on the nebular evolution of SNe.Observed differences in late-time decline rates, particularly for those significantly diverging from power due to 56 Co radioactive decay, motivate the need for larger samples of light curves to be collected.
Additional sources of late-time luminosity can originate via contribution from light echoes or from interaction of the SN shock with circumstellar matter (CSM).For instance, light echoes observed with HST have been spatially resolved around four nearby SNe Ia: SN 1991T (Sparks et al. 1999), SN 1998bu (Cappellaro et al. 2001), SN 2006X (Wang et al. 2008), and SN 2014J (Crotts 2015;Yang et al. 2017).The presence of light echoes is not limited to SNe Ia -resolved echoes have been detected around a number of core-collapse SNe as well, such as SN 1987A (e.g., Bond et al. 1990), SN 1993J (Sugerman & Crotts 2002;Liu et al. 2003), SN 2003gd (Sugerman 2005;Van Dyk et al. 2006), SN 2008bk (Van Dyk 2013), SN 2012aw (Van Dyk et al. 2015), and SN 2016adj (Stritzinger et al. 2022).Recently, a light echo has also been detected around SN 1987A as detailed by Ding et al. (2021) and Cikota et al. (2023).Snapshot programs, in particular, can provide statistics on the frequency of light echoes around various types of SNe.
The sustained late-time luminosity of some SNe II can be explained by interaction of the SN shock with large amounts of CSM set up by the pre-SN wind (e.g., Fox et al. 2013;Smith 2014Smith , 2017;;Smith et al. 2017).The sustained optical emission in this case likely arises from a radiatively-cooled shell.HST Snapshot programs can also help reveal the nature of the SN impostors, events similar to Type IIn SNe (with relatively narrow H emis-sion lines in their spectra), but which are subluminous compared with core-collapse SNe (M V ≈ −14 mag) near maximum brightness (Smith et al. 2011a;Van Dyk & Matheson 2012).
In this paper, we present the results of HST Snapshot program GO-16179 (PI A. Filippenko), along with some data from previous Snapshot programs such as GO-14668 and GO-15166 (PI A. Filippenko).The primary goal of this study is fairly simple: to determine whether the SNe at late times are essentially following the exponential light-curve decline, as a result of reprocessing of γ-rays and positrons from radioactive 56 Co decay, or whether an additional power source is at work.In Section 2 we provide the details of the HST observations, and Section 3 describes our analysis, including the data reduction and results.The results as they pertain to all of the individual objects are discussed in Section 4. Section 5 provides our summary and conclusions.

OBSERVATIONS
The program was executed during HST Cycle 28 from 2020 November 11 through September 24 2021 (UTC dates are used throughout), with the Wide Field Camera 3 (WFC3) UVIS.The original observing request was for 55 visits, consisting of 9 SNe Ia, 27 SNe Ib/c, 12 SNe II, and 7 SN impostors (for reviews of SN spectral classification, see Filippenko (1997), ,Gal-Yam (2017)), of which 38 visits (∼ 70%) were actually executed.However, one visit, of the target SN IIn 2005ip, failed outright (all data lost), and two other visits, of the SN Ia 2018hfp and SN Ia 2019cth, experienced loss of guiding and were rendered useless.The final observed sample of 35 targets (3 SNe Ia, 9 SNe Ib/c, 19 SNe II, and 4 SN impostors) is summarized in Table 1.All of the data presented in this article were obtained from the Mikulski Archive for Space Telescopes (MAST) at the Space Telescope Science Institute.The specific observations analyzed can be accessed via the associated DOI.
The observing scheme for the program was to obtain all of the executed visits within an optimum "Snapshot" orbit of ∼ 38-40 min, which was intended to increase the likelihood that a visit would be scheduled.To accomplish this, each visit orbit consisted of observations in two bands, with a total exposure time of 710 s in one band and 780 s in the other.Typically, we observed in the F555W band with the shorter exposures and in F814W with the longer ones, but varied the actual filter combination used depending on the specific science goal for each visit.The observations in each band per visit were split into two exposures of equal duration, employing a line dither between exposures to mitigate as best as possible against cosmic-ray hits and detector cosmetic defects.
All of the data from the program had no exclusive access period and were publicly available from the Mikulski Archive for Space Telescopes (MAST)1 as soon as they were processed through the Space Telescope Science Institute (STScI) HST standard pipeline.We note that at least one study not focused on SNe or related transients, but on the optical counterparts of extragalactic ultraluminous X-ray sources (Allak et al. 2022), has already made use of our data.

Data Reduction
We ran the suite of STScI Drizzlepac routines (STSCI Development Team 2012) on the data from each visit, to construct a drizzled image mosaic in each band.To locate the general sites of SNe in each of the image mosaics, coordinates from the Transient Name Server (TNS)2 were first used.To more precisely isolate the SN site, either we directly compared the new data with previous HST images containing the SN from previous epochs (in many of the cases, from previous Snapshot programs) when the SN was brighter, or, if no prior HST data were available, astrometrically aligned the HST Snapshot mosaics with ground-based images of the SN (in many cases, obtained by us with the 0.76 m Katzman Automatic Imaging Telescope (KAIT ;Filippenko 2003) or at the Nickel 1 m telescope, both at Lick Observatory).In the former cases, we could simply blink the Snapshot images with the previous HST images and visually identify the SN in the new data.In the latter cases, stars were found in common between the HST and ground-based images, and astrometric registration between the two datasets was performed.The SN position in the ground-based data was then transformed to the HST reference frame.The typical astrometric uncertainty was in the range of ∼ 0.1 ′′ , and in no case where we performed this alignment were there any other sources in the error circle.In most cases, it was then readily apparent which object in the Snapshot data was the SN in question.
To obtain photometry from the data in both bands for each visit, we ran the individual frames for the entire visit through the Dolphot package (Dolphin 2016), using one of the mosaics as the reference image for source detection.Generally, the recommended WFC3/UVIS parameters from Dolphin (2016) were used.

Results
Once an SN location was isolated, its image coordinates were matched to the output from Dolphot to retrieve photometric information for the SN, along with any potential error/quality flags.The photometric results for all of the SNe are given in Table 2.The brightness measurements for the SNe are in Vega magnitudes for the HST flight system bands, as indicated in the table.In a number of cases, nothing was detected by Dolphot at the SN position, and subsequently we estimated upper limits to detection and provide these in the table.The upper limits were based on the formal estimations of signal-to-noise ratio (S/N ) from Dolphot, and we have set the significance at S/N = 5.The major caveat is that the uncertainties, particularly at low flux levels and in crowded environments, are underestimated by Dolphot (Williams et al. 2014), so the significance levels of the nondetections are likely overestimates; see Van Dyk et al. (2023) for a discussion of this issue.
For each event we also indicate in Table 2 our assessment of whether the late-time light curves appear to be powered by radioactive decay (the 56 Co decay rate is shown for comparison in the light-curve figures; see Section 4), with either a "yes" or "no."In some cases we were unable to make this assessment, since the event is either an SN impostor (and therefore still likely undergoing a superoutburst) or early-time photometry did not exist; these are listed as "N/A," for "not applicable."In a number of cases we were unable to confidently ascribe the decline to radioactive decay, primarily because we could only place an upper-limit constraint on the latetime emission, and these are indicated by a question mark.
In the next section we discuss each of the individual objects separately.

SN 1988Z
SN 1988Z was recognized early as an unusual SN II (Stathakis & Sadler 1991).From the luminous radio (Van Dyk et al. 1993;Williams et al. 2002) and X-ray (Fabian & Terlevich 1996;Schlegel & Petre 2006) emission detected from the SN, along with the characteristics of its optical photometric and spectroscopic emission (Turatto et al. 1993;Aretxaga et al. 1999), it was posited that long-lived interaction of the SN shock with a pre-existing dense CSM was the likely source of this Note-SN positions and discovery dates are adopted from the TNS.Foreground Milky Way visual extinction, AV (MW), is adopted in each case from the NASA/IPAC Extragalactic Database (NED; Schlafly & Finkbeiner 2011).Distances and heliocentric velocities (v hel ) are also obtained from NED. radiation.In fact, recent spectra show that SN 1988Z is still strongly interacting with dense CSM even three decades after explosion (Smith et al. 2017).SN 1988Z is generally considered to be a Type IIn SN, even a prototype of this subclass.The Snapshot observations were obtained on 2021 February 19 in F625W (∼ R) and F814W (∼ I).As can be seen in Figure 1, amazingly the SN was still detectable in the HST images 11,758 d (32.2 yr) after discovery, at m F625W = 24.83± 0.03 and m F814W = 24.98±0.10mag.(We had intentionally used the F625W band, sensitive to any remaining Hα line emission, and the F814W band, potentially sensitive to hot dust, to increase the probability of detection at such late times.)We pinpointed the location of the faint SN in these images by employing previously unpublished imaging in 2013 February from our Snapshot program in Cycle 20 (GO-13029, PI A. Filippenko), when the SN was brighter (m F625W = 24.08 ± 0.04 and m F814W = 24.49± 0.07 mag).
We have included the two sets of Snapshot data together with the ground-based, earlier-time R light curves from Aretxaga et al. (1999) and Turatto et al. (1993).The light curve clearly does not follow the trend for radioactive-decay power, and the very latetime points from HST appear to follow the break in the light curve that began at ∼ 2000 d.

SN 1993J
SN 1993J is one of the best-studied and historically prominent SNe ever discovered.It remains a benchmark SN IIb to which more recent discoveries are often compared, with a rich array of multiwavelength observations collected over the last 30 yr.The proximity of its host galaxy facilitated detection and characterization of its progenitor as a K-type supergiant, even from the ground (Filippenko et al. 1993;Aldering et al. 1994;Cohen et al. 1995;Van Dyk et al. 2002), and excess flux in the blue and near-ultraviolet (UV) bands suggested the presence of a binary companion, consistent with models of supergiant mass loss onto the secondary (e.g., Podsiadlowski et al. 1993;Maund et al. 2004;Fox et al. 2014).The optical light curve of the SN has been powered by ongoing interaction with the CSM in a relatively slow decline.The SN had faded enough by 2004 that it was evident that the supergiant progenitor had vanished (Maund & Smartt 2009).The SN had remained too luminous, however, for a binary companion to be isolated via imaging, until Fox et al. (2014)  We located the SN site in our Snapshot data in F336W (∼ U ) and F814W from 2020 December 14, when the SN was at 10,123 d (27.7 yr), by consulting Fox et al. (2014, their Figure 1) and also comparing with the data from 2012 February from program GO-12531 (PI A. Filippenko), when the SN was brighter in F336W and F814W (22.33 ± 0.02 and 20.87 ± 0.01 mag, respectively).We also compared to previously-unpublished data from 2015 March at F336W from GO-13648 (PI O. Fox; the SN was at 22.62 ± 0.05 mag); see Figure 2. SN 1993J clearly is not following the radioactive-decay trend, which has been the case for most of its late-time (≳ 500 d) history.However, the SN appears to have faded more rapidly, compared to the more gradual decline up to (and possibly beyond) 2015.This could be indicating that the SN shock was encountering a less dense circumstellar environment than previously, consistent with the results of modeling of the declines in both the radio and X-ray emission (Kundu et al. 2019).
We can now clearly detect in F814W (see Figure 2) a star immediately to the northwest of the SN, the presence of which was just hinted at by Fox et al. (2014).Following the labeling scheme from Fox et al. (2014), this star "O" has m F814W = 22.88 mag.We have reprocessed the 2011 data from Fox et al. (2014), adopting the same Dolphot parameters that we used here (which differ somewhat from those used in that previous study); our results are presented in Table 3.We also include our measurements for F336W data (640 s) from GO-13648 (PI O. Fox).Furthermore, we present our results for these same stars from our Snapshot data.On average, the stars have essentially the same measured brightnesses in F814W, with the Snapshot values being slightly fainter (by ∼ 0.06 mag), whereas the F336W Snapshot measurements appear to differ by substantially more, ∼ 0.46 mag fainter, than our remeasurements of the Fox et al. (2014) data.We can potentially account for this large difference in that the 2011 total exposures (3000 s) in F336W were a factor of ∼ 4.2 deeper than the 710 s total Snapshot exposure, so the S/N was substantially higher for the former than the latter.

SN 2000ch
SN 2000ch, discovered with KAIT at magnitude 17.4, was suspected early on to be an unusual and very luminous variable in NGC 3432.Wagner et al. (2004) initially described its erratic behavior, and Pastorello  Our Snapshot observations were obtained in F555W (∼ V ) and F814W on 2020 December 13.As one can see in Figure 3, the object is easily detectable in the HST images, and its light curve appears very much unlike that of a typical SN.Aghakhanloo et al. (2022a) have recently analyzed the continued photometric evolution of SN 2000ch, finding periodicity to the cycle of repeating outbursts, which suggests a binary nature for the transient.The object, since it is a likely SN impostor, is not powered at late times by radioactive decay.

SN 2010jl
SN 2010jl was classified as a luminous Type IIn SN, and Stoll et al. (2011) presented early-time light curves and spectra.Smith et al. (2011b) identified a luminous blue (M F300W ≈ −12.0 mag) point source at the SN location that they identified as a candidate progenitor, although Fox et al. (2017) have since demonstrated that this is less likely.Smith et al. (2011b) noted that even if the blue source was a nearby star cluster, its young age suggested a high initial mass of > 30 M ⊙ for the progenitor of SN 2010jl.Further optical and near-infrared (IR) monitoring of SN 2010jl has been presented by Zhang et al. (2012), Ofek et al. (2014), Borish et al. (2015), Jencson et al. (2016), and others.As an SN IIn, similar to the case of SN 1988Z (Section 4.1), there has Uncertainties (1σ) are in parentheses.
1" long been multiwavelength evidence for strong circumstellar interaction (e.g., Smith et al. 2011bSmith et al. , 2012;;Fransson et al. 2014;Chandra et al. 2015).Most notable is the observational and analytical indications for the presence of dust associated with the SN (e.g., Andrews et al. 2011;Smith et al. 2012;Gall et al. 2014;Sarangi et al. 2018;Bevan et al. 2020).and October 2016 by GO-14668 (PI A. Filippenko), at m F336W = 22.08 ± 0.03 and m F814W = 22.67 ± 0.03 mag (see Fox et al. 2017).Analysis of a set of HST images from February 7 2018(nearly 3 yr prior to our data), previously-unpublished data from GO-15166 (PI A. Filippenko), shows that SN 2010jl was still detectable at F814W, with m F814W = 23.11± 0.04 mag, but had already become undetectable in F336W (with an upper limit of 23.5 mag).Given the late-time brightness of the SN, radioactive decay could not have been the object's primary source of power.We conclude from analysis of our Snapshot images that the SN has now vanished, with upper limits of 24.8 mag in F336W and 25.6 mag in F814W.

SN 2010mc
Ofek (2012) discovered SN 2010mc during the course of the PTF survey.Howell & Murray (2012) classified it subsequently as an SN IIn at redshift z = 0.035.Looking back in the PTF data, Ofek et al. (2013) discovered an astounding outburst event ∼ 40 d prior to the apparent SN.Smith et al. (2013) pointed out that SN 2010mc was a near twin of the remarkable event SN 2009ip, and Smith et al. (2014) proposed that both events were the terminal SN IIn explosions arising from eruptive blue supergiant progenitors.A terminal SN explosion has since been confirmed for SN 2009ip (Smith et al. 2022).The last published spectrum of SN 2010mc was from day 1024 (Smith et al. 2014), which at that time showed strong shock-broadened Hα emission indicative of ongoing CSM interaction.
We detected SN 2010mc in both our F555W and F814W Snapshots from 2021 September 24, 4053 d (11.1 yr) after discovery; see Figure 5.We had isolated the site of the SN using HST data from 2017 March 26 obtained by program GO-14668 (PI A. Filippenko), when the SN was at m F555W = 24.26± 0.04 and m F814W = 25.35 ± 0.10 mag.One will note that, both in 2017 and 2021, the SN is significantly brighter in F555W than in F814W, which we speculate must be the result of sustained luminous Hα emission from the SN within the F555W bandpass, with much less luminous continuum emission in F814W.SN 2010mc has diverged from radioactive-decay power since day ∼ 400 and continues to do so, most likely as a result of sustained CSM interaction.However, it is possible that some of the light is contributed by a star cluster coincident with the SN, as was the case for SN 2009ip; deeper and higher-resolution observations are needed to identify such a cluster.

SN 2011dh
The nearby SN 2011dh in M51 has become, along with SN 1993J, one of the best-studied SNe IIb, if not one of the best-studied SNe of any type thus far.Extensive UV, optical, and near-IR follow-up observations were carried out not long after discovery by Arcavi et al. (2011), Sahu et al. (2013), Shivvers et al. (2013), Ergon et al. (2014, 2015), Mauerhan et al. (2015), Marion et al. (2014), and others.Multiwavelength observations, including X-ray and radio, were indicative of circumstellar interaction; see, for example, Martí-Vidal et al. (2011), Krauss et al. (2012), Horesh et al. (2013), Maeda et al. (2014), de Witt et al. (2016), and Kundu et al. (2019).Both Maund et al. (2011) and Van Dyk et al. (2011) independently identified the SN's progenitor.Soderberg et al. (2012), followed by Van Dyk et al. (2011), argued that the progenitor was compact, while Maund et al. (2011) pointed to the detected yellow supergiant as the star that exploded, and this was supported by the modeling by Bersten et al. (2012) and subsequently confirmed by the supergiant progenitor's disappearance (Van Dyk et al. 2013b).Furthermore, detailed theoretical modeling of the progenitor by Benvenuto et al. (2013) supported the binary origin for the SN.Folatelli et al. (2014) claimed detection of a possible blue companion to the progenitor, although Maund et al. (2015) and Maund (2019) cast some doubt on that possibility.Maund (2019), analyzing a veritable treasure trove of HST imaging serendipitously covering the SN site, argued that a light echo originating from dust with a preferred disk geometry could be responsible for the observed extended late-time emission.We pinpointed the location of the SN in our Snapshot images in F555W and F814W from 2020 December 10 (3480 d ≈ 9.5 yr after discovery), using a number of these prior HST images for comparison, and as can be seen, the emission is significantly fainter than from the analysis by Maund (2019); see Figure 6.Since that study, previously-unpublished HST observations also have been obtained of the SN site in 2019 on November 24 with the Advanced Camera for Surveys (ACS) Wide-Field Channel (WFC) in F814W by GO-15645 (PI D. Sand; m F814W = 23.89± 0.02 mag) and in 2020 January 29 with WFC3/UVIS in F555W by GO-16024 (PI A. Filippenko; m F555W = 25.02 ± 0.05 mag).(We note that data were also obtained by GO-16024 in F225W, not shown; however, the SN was not detected, to a limit of 24.0 mag; cf.Maund et al. 2015.)Taken together, these late-time HST data indicate a slow, steady fading of the SN emission -but clearly the SN at very late times has not followed a radioactive decay-powered decline.We obtained our Snapshot observations in F606W ("Wide" ∼ V ) and F814W on 2021 February 16, 3329 d (9.1 yr) after discovery.We had purposely selected F606W in this case, rather than F555W, since the for-mer is the preferred bandpass in which to acquire data, together with F814W, for use in tip-of-the-red-giantbranch (TRGB) distance estimates (Anand et al. 2021).Our intention was to obtain data that can better constrain the distance to SN 2012A, although that TRGB estimation is beyond the scope of this paper.(Our focus for the Snapshots was on the SN itself, which is within the main body of the host galaxy, and not the galaxy halo in which the TRGB would likely be more apparent.)We astrometrically aligned KAIT images from de Jaeger et al. ( 2019) with our Snapshot images, in order to isolate the SN site, and concluded from that analysis that the SN was no longer detectable, to 26.9 and 25.8 mag in F606W and F814W, respectively; see We pinpointed the location of the SN in our Snapshot data in F555W and F814W from 2021 February 17, using HST data obtained on 2016 October 24 by GO-14668 (PI A. Filippenko), as well as the pre-explosion WFPC2 images in which the progenitor had been identified; see Figure 8.What is most strikingly apparent is the continued presence of the light echo in both bands around the SN site.Whereas the SN was obscured by the echo at earlier times (Van Dyk et al. 2015), the SN had become recoverable in both the HST F814W and F555W images 3261 d (8.9 yr) after discovery.
The echo itself is seen almost as a perfect ring, although asymmetric relative to the SN position, with the SN offset by ∼ 2.6 pixels (∼ 0. ′′ 103 at the UVIS pixel scale) southeast from the ring center.The surface brightness of the echo is still highest to the east and southeast, as reported by Van Dyk et al. (2015).We estimate that the radius of the echo is ∼ 7.4 pixels (∼ 0. ′′ 293).This is nearly double the radius, at ∼ 2492 d later than when first discovered by Van Dyk et al. (2015) in 2014.A detailed analysis of the echo and its evolution is beyond the scope of this paper.

SN 2013df
SN 2013df is an SN IIb in NGC 4414 and was first studied in detail by Van Dyk et al. (2014), who also identified its yellow supergiant progenitor.Morales-Garoffolo et al. (2014), Maeda et al. (2015), and Szalai et al. (2016) performed additional optical and near-IR follow-up observations.It was established early that SN 2013df strongly resembled SN 1993J (Section 4.2), both in SN and progenitor properties.The radio and X-ray emission from the SN (Kamble et al. 2016), together with the late-time (∼ 670 d) optical spectral characteristics (Maeda et al. 2015), were indicative of circumstellar interaction.
We pinpointed the exact location of SN 2013df by comparing directly with HST observations from 2013 July 15 (GO-12888; PI S. Van Dyk), when the SN had m F555W = 16.15±0.01mag.As one can see from Figure 9, the SN was clearly detected in our Snapshot images from 2021 February 15, both in F336W and F555W, 2811 d (7.7 yr) after discovery.That the SN is still relatively bright in F336W indicates that the interaction was still ongoing at the time of our observations.The brightness in F555W is likely dominated within the bandpass by continued luminous Hα emission, as well as less prominent He i/Na i emission, as seen in the latetime spectra (Maeda et al. 2015).In the figure we have overlaid the F336W light curve of SN 1993J (see Section 4.2) on the light curve of SN 2013df in this same band.As one can see, the two agree amazingly well, which would imply that, based on the photometric evolution over more than 2800 d (7.7 yr), SN 2013df is essentially a twin of SN 1993J, as the early-time data, including the progenitor identification, tended to indicate as well.2023), based on the brightness of the SN in these Snapshot data, concluded that the progenitor identified by Fraser et al. (2014) had vanished (confirming an earlier inference made by Mauerhan et al. 2017).
Technically, we covered the field again in bands F438W (∼ B) and F625W when we observed AT 2019krl (see Section 4.32); however, unfortunately the SN 2013ej site fell within the chip gap for both of those bands.

SN 2014C
SN 2014C in NGC 7331 is a fascinating event, having been classified soon after explosion as an SN Ib (without H) and, after ∼ 1 yr, exhibited distinct and strong signs of circumstellar interaction, similar to an SN IIn, with strong Hα emission (e.g., Milisavljevic et al. 2015).From radio and X-ray monitoring Margutti et al. (2017) inferred that the progenitor star had ejected a massive (∼ 1 M ⊙ ) H shell decades to centuries before explosion, and that possibly as many as ∼ 10% of all SN Ib progenitors might experience a similar history.Brethauer et al. (2022) have since interpreted that the shell, with as much as ∼ 2 M ⊙ , has a radius of ∼ 2 × 10 16 to ∼ 10 17 cm.Milisavljevic et al. (2015) identified in preexplosion F658N HST imaging a luminous Hα source at the SN's position, which they inferred was a stellar cluster that was home to the progenitor.Sun et al. (2020) performed a detailed analysis of this cluster and estimated an age for it of ∼ 20 Myr, which they found consistent with a ∼ 11 M ⊙ star stripped partially of its envelope via mass transfer with a companion in a relatively wide binary system, followed by an eruptive ejection of the remaining H prior to explosion.
We located the SN in our Snapshot images obtained in F336W and F625W on 2021 August 20 (2786 d ≈ 7.6 yr after discovery), using HST data obtained in 2016 October for program GO-14668 (PI A. Filippenko); see Figure 11.We intentionally selected the F336W band, to sample any late-time UV emission from the SN, and F625W, to sample Hα emission, both being indicators of ongoing interaction.Zheng et al. (2022) undertook early-time optical monitoring with KAIT, during the "Ib" phase of the SN, prior to the onset of strong interaction, and we combine that photometry here with our Snapshots and other available late-time HST data.What can be seen in the figure is the contribution to the light curves from the circumstellar interaction and that the UV emission, in particular, has declined somewhat in strength since day ∼ 1500.This behavior would be consistent with the decline in the observed X-ray luminosity after day ∼ 1000 (Brethauer et al. 2022).

SN 2015cp
SN 2015cp (also known as PS15dpq), in a host at z ≈ 0.04, was originally classified as a "91T-like" overluminous SN Ia at ∼ 40 d post-peak, but was shown to be experiencing circumstellar interaction at 681 d after explosion, based on a near-UV (NUV) HST detection (Graham et al. 2019 and references therein).Harris et al. (2018), based on radio and X-ray follow-up observations of this "SN Ia-CSM," constrained the total circumstellar mass at < 0.5 M ⊙ .Graham et al. (2019) estimated constraints on the inner radius of the CSM of R CSM > 10 16 and < 10 17 cm.There was little earlytime optical follow-up photometry of the SN, beyond a minimal iPTF light curve in the g and R bands.
Our Snapshot observations were obtained on 2020 November 30 (1798 d ≈ 4.9 yr after discovery) in

SN 2016adj
SN 2016adj in NGC 5128 (Centaurus A) is certainly one of the nearest SNe (at 3.42 Mpc) of any type in recent years.The object was classified as a corecollapse SN, potentially with a stripped-envelope progenitor, with a C-rich SN Ic classification inevitably proposed (Stritzinger et al. 2023).It became readily apparent at early stages that SN 2016adj was heavily reddened (A V ≈ 2-4 mag) by internal host dust (Stritzinger et al. 2016).In time it also became obvious that there was, at first, one (Sugerman & Lawrence 2016) and then several prominent light echoes apparent around the SN (Stritzinger et al. 2022).We observed the SN as part of our Snapshot program on 2021 July 28, 1998 d (5.5 yr) after discovery, in F438W and F555W.The former band was chosen intentionally to capture the blue light from the echoes, whereas the latter was used to establish the echo colors.We found the precise position of the SN in our HST data comparing with images obtained on 2016 February 22 for GO-14115 (PI S. Van Dyk) in F438W and F814W; see Figure 14.The SN was not detected in the current Snapshots, at 27.4 and 26.5 mag in F438W and F555W, respectively.The echoes, however, are quite prominent (see also Stritzinger et al. 2022, who used our Snapshot data as well in their work).A detailed analysis of the light echoes is beyond the scope of this paper.

SN 2016bkv
SN 2016bkv is an exceptional example of a lowluminosity SN II-P, with an extraordinarily long plateau phase (≳ 140 d) and very low expansion velocities, in addition to a strong initial bump in the light curve, as well as "flash-ionization" features, all signs of short-lived, early-time circumstellar interaction (Nakaoka et al. 2018;Hosseinzadeh et al. 2018).Nakaoka et al. (2018) concluded that the progenitor mass-loss rate within a few years of explosion was quite high, ∼ 1.7 × 10 −2 M ⊙ yr −1 (although see Deckers et al. 2021), possibly indicating that the star had experienced a violent outburst.Hosseinzadeh et al. (2018) further suggested that SN 2016bkv is an example of an electron-capture (EC) SN.Through radiative-transfer modeling of the spectra, Deckers et al. (2021) inferred an odd surface composition for the progenitor, implying that it was more likely a binary rather than a single star, with the primary either accreting unprocessed material from its companion or undergoing a merger before explosion.2022).Also shown are ("Other") B and V (c) light-curve data (Stritzinger et al. 2023), together with our upper limits.We see a diffraction spike going straight through the F555W image, but it does not affect the SN site.
Our Snapshot data were obtained on 2020 December 13, 1721 d (4.7 yr) after discovery, in F555W and F814W.The location of the SN was pinpointed by referring to early-time HST F555W data from 2016 April 14 (GO-14115, PI S. Van Dyk), when the SN was at m F555W = 16.06 ± 0.01 mag.The SN was still clearly detected in our Snapshots in both bands; see Figure 15.

AT 2016blu
AT 2016blu, also known as NGC 4559-OT1, PSN J12355230+2755559, and Gaia16ada, was actually discovered by the Lick Observatory Supernova Search earlier, in 2012 (Kandrashoff et al. 2012), and classified as a luminous blue variable (LBV) or SN impostor (see also Sheehan et al. 2014).The object is highly variable and has been "rediscovered" a number of times over the years thereafter (e.g., Vinokurov et al. 2021) -hence, the multiple identifiers for the same object.Not long after discovery, Van Dyk et al. (2012a) identified a possible precursor in HST images from 2005 and, based on preliminary photometry, estimated that the star had M V = −9.4mag with intrinsic colors B − V = 0.10 and V − I = 0.36 mag, consistent with an early-F spectral type.Aghakhanloo et al. (2022b) recently conducted an analysis of the recurring outbursts from the transient.They found a periodicity to the outbursts, and proposed that AT 2016blu is probably an LBV in an eccentric interacting binary very much like SN 2000ch.
Our HST program obtained observations on 2021 February 17 (3320 d since discovery) in F606W and F814W.The F606W band was used rather than F555W, expressly in order to probe the TRGB for estimation of the distance to the host galaxy (NGC 4559;however, McQuinn et al. 2017 had already performed a TRGB analysis, with different HST data, and found a distance of 8.91 Mpc).
We pinpointed the AT's location by astrometrically aligning with KAIT ground-based data from 2016 April, as well as precursor HST images from 2005 March (GO-10214, PI R. Soria).AT 2016blu is still strongly detected in our Snapshot images in both bands (see Figure 16).
Our Snapshot observations of the SN were executed on 2020 December 6, 1655 d (4.5 yr) after discovery, in F336W and F814W.In order to pinpoint the SN location in the Snapshots, we compared with HST data from 2016 October 4 for our previous Snapshot program GO-14668 (PI A. Filippenko), when the younger SN was at m F555W = 16.79 ± 0.01 and m F814W = 16.18 ± 0.01 mag.The SN was not detected in our late-time Snapshots, to limits of 26.1 and 26.0 mag in F336W and F814W, respectively; see Figure 17 1993J (Section 4.2 andSN 2011dh (Section 4.6;Tartaglia et al. 2017b).Both Arcavi et al. (2017) and Piro et al. (2017) modeled the first cooling peak of the SN to infer properties of the progenitor.Tartaglia et al. (2017b) and Kilpatrick et al. (2017) independently identified a progenitor candidate in pre-explosion HST images.Bersten et al. (2018) also identified and characterized the pro-genitor, as well as the SN itself (which included KAIT photometry, enhanced with further data by Zheng et al. 2022).
The HST Snapshots were obtained on 2021 August 19, 1795 d (4.9 yr) after discovery, in F438W and F606W.The SN location was found using HST data taken 2016 October 10 for GO-14116 (PI S. Van Dyk), when the SN was young and bright, at m F555W = 15.11± 0.01 mag.Kilpatrick et al. (2022) revisited the SN and, using our Snapshots found that m F606W = 25.10 ± 0.07 and m F438W = 26.61± 0.27 mag.Our results differ from these, with m F606W = 24.95± 0.04 and m F438W > 26.1 mag.We can potentially ascribe the discrepancy in F606W as due to differences in assumed Dolphot input parameters; however, as can be seen in Figure 19 of the late-time light curve in F606W implies that CSM interaction may be a source of additional power.

AT 2016jbu
AT 2016jbu (Gaia16cfr) in NGC 2442 has been considered since early after discovery to be an SN impostor, although it has been argued that it should actually be considered a pre-explosion LBV (Kilpatrick et al. 2018b) or simply an interacting, SN 2009ip-like transient (Brennan et al. 2022b).Both Kilpatrick et al. (2018b) and Brennan et al. (2022c) independently identified in preoutburst HST images and subsequently characterized the precursor: a massive (∼ 22-30 M ⊙ ) yellow supergiant enshrouded by a dusty circumstellar shell.
The object was detected in our Snapshots in F555W and F814W on 2021 August 21, 1725 d (4.7 yr) after discovery.We isolated the location of the object using HST data obtained on 2019 March 21 for our previ-ous Snapshot program GO-15166 (PI A. Filippenko), together with finder charts from the ensemble of HST observations presented by Brennan et al. (2022c); see Figure 20.Brennan et al. (2022a) also obtained multiband HST observations of AT 2016jbu on 2021 December 6, only 107 d later, and discovered that the observed brightness of the transient was less than the precursor levels.Those authors further found that it was difficult to explain this dimming in terms of increasing dust obscuration and therefore concluded that the precursor had likely vanished -thus, AT 2016jbu may have actually been a terminal explosion.Our Snapshot data also confirm the dimming and precursor disappearance.2023) concluded that the RSG candidate was indeed the progenitor, and also confirmed the late-time CSM interaction, manifested as a UV excess.

SN 2017gax
SN 2017gax (DLT17ch) was discovered in NGC 1672 by Tartaglia et al. (2017a) on 2017 August 14.The SN was spectroscopically classified, within a day of discovery, as an SN Ic by Jha et al. (2017).The SN location was established in our F336W and F814W Snapshots from 2020 November 27, 1202 d (3.5 yr) after discovery, using HST data from 2017 October 19 (GO-14645, PI S. Van Dyk), when the SN was at m F555W = 15.94 ± 0.01 mag.The SN was not detected in the Snapshot data in either band; see Figure 23.Unfortunately, no published photometry exists, beyond the report by Maguire et al. (2017) of the SN at V = 16.1 ± 0.1 mag on 2017 November 9; thus, we are unable to show a light curve for this SN.

SN2017gkk
SN 2017gkk was discovered in NGC 2748 on 2017 August 19 at 15.6 mag by Balanutsa et al. (2017), and then later rediscovered at 14.7 mag (both unfiltered) by Itagaki (2017).The classification spectrum obtained just days after discovery by Onori (2017) showed it was an SN IIb.The SN was detected in our Snapshot images on 2021 September 24, 1485 d (4.1 yr) after discovery, in both F555W and F814W; see Figure 24.The location of the SN was established using data obtained on 2019 February 22 for our previous Snapshot program GO-15166 (PI A. Filippenko), when the SN was at m F555W = 23.55 ± 0.02 and m F814W = 23.10 ± 0.04 mag.Limited early-time unfiltered ("clear"; ∼ R) photometry was obtained with KAIT.The light-curve behavior at late times in both Snapshot bands may imply that CSM interaction is contributing to the SN luminosity.

SN 2017ixv
SN 2017ixv was discovered in NGC 6796 on 2017 December 17 by Cortini (2017).It was classified shortly thereafter as an SN Ic-BL by Leadbeater (2017).Unfortunately, we are not aware of any published followup photometry.SN 2017ixv was not detectable in our F555W and F814W Snapshots from 2021 January 11, 1122 d (3.1 yr) after discovery.We note that the SN site is in an edge-on spiral galaxy, and therefore the exact location is difficult to confirm without any earlier imaging data, owing to the crowded environment.In order to locate the SN, we used the absolute position (Cortini 2017), assuming a 0. ′′ 2 uncertainty.We further added this in quadrature with a quoted uncertainty of 0. ′′ 03 in the Gaia-based HST astrometric grid, and the total uncertainty is reflected in the radius of the dashed circle in Figure 25.Based on this position, it appears that the SN may have been in or near a patch of nebulosity in the host galaxy.

SN 2018gj
SN 2018gj was discovered in NGC 6217 by Wiggins (2018) on 2018 January 1.It was classified as an SN IIb (and possible II-P) by Bertrand (2018) and as an SN II by Kilpatrick et al. (2018a).Teja et al. (2023) conducted extensive photometric and spectroscopic monitoring of

SN 2018zd
SN 2018zd in NGC 2146 was monitored and analyzed independently by Zhang et al. (2020) and Hiramatsu et al. (2021).Both studies considered the event to be a low-luminosity SN II; however, the latter considered this to be the best example so far for an electoncapture SN, whereas the former found it to have properties more consistent with a normal SN II-P.Both studies found it likely that the progenitor candidate, identified in pre-explosion HST images, could have been a superasymptotic-giant-branch star.

AT 2018cow
AT 2018cow, in CGCG 137−068 at z = 0.0141, is a particularly intriguing object.It very rapidly became exceedingly luminous (∼ −22 mag absolute) and blue, and is considered a prototypical "fast blue optical transient," or FBOT."The Cow," as it has been dubbed, stimulated intense interest in the community, leading to several multiwavelength monitoring campaigns and theoretical analyses, including those of Perley et al. (2019), Margutti et al. (2019), andXiang et al. (2021).Despite all of the focused effort, the nature of AT 2018cow and its precursor is still not settled.For instance, Fox & Smith (2019) surmised, based on similarities with various interacting SNe, that CSM interaction in a relatively H-depleted system could explain some its observed properties.Chen et al. (2023) concluded that a fading transient UV source persists, which may be from ejecta-CSM interaction or from a central engine, more specifically a precessing accretion disk.Additionally, Inkenhaag et al. (2023) studied the late-time brightness of AT 2018cow and estimated the potential black hole's mass using our Snapshot data.
Our F555W and F814W Snapshots F555W and F814W were obtained on 2021 July 25, 1134 d (3.1 yr) after discovery.The location of the object was con-

SN 2018ivc
From high-cadence follow-up spectroscopic and photometric observations since discovery, Bostroem et al. (2020) concluded that SN 2018ivc in NGC 1068 is an un-usual SN II.That study placed limits on the properties of the progenitor, based on available pre-explosion HST of this famous Seyfert galaxy.Maeda et al. (2023b) considered SN 2018ivc as a possible variant of SN II-L, with transitional characteristics between II-P and IIb.Interaction of the SN shock with pre-existing wind matter appears to be playing a strong role in the SN's emission (Maeda et al. 2023b), with origins in a progenitor experiencing an extreme form of binary evolution (Maeda et al. 2023a).The SN was detected in our HST

SN 2020dpw
Wiggins (2020) discovered SN 2020dpw in NGC 6951 (based on the discovery position, the host must have been incorrectly reported as NGC 6952) and Kawabata (2020) classified it as an SN II-P.Unfortunately, we have no knowledge of existing early-time photometry, although the SN was easily detected in our HST F555W and F814W Snapshots from 2020 December 13, near the reported discovery location 292 d (0.8 yr) after discovery; see Figure 33 2023) also identified the progenitor through preexplosion images.

DISCUSSION AND CONCLUSIONS
We have conducted an analysis of images that we obtained from an HST Snapshot survey during Cycle 28 of nearby SNe at late times.We were ultimately able to observe successfully the targeted sites of 31 SNe of various types and 4 SN impostors.The goal of the program was to reveal the possible origins of their late-time emission or lack thereof.Only 2 of the 31 SNe (SN 2020hvp and SN 2020jfo) listed in Table 2 convincingly exhibited lingering emission most likely ascribed to radioactive decay of 56 Co.For 12 of the remaining SNe (indicated by "Yes?" in Table 2) we could not determine the source of the late-time emission, since these events were no longer detectable, and upper limits to their lu- minosities were not sufficiently constraining.All three of the observed SNe Ia fall in this category.For three of the SNe in the observed sample (SN 2017gax, SN 2017ixv, and SN 2020dpw), no early-time photometry was available, and the former two SNe were no longer detectable, so it was not possible to determine whether radioactive decay was powering the light at late times.A remaining 15 SNe were detected; however, it was clear from their extended light curves that the emission was in excess of what we would expect for radioactive decay.We can infer in these cases that the emission may arise, at least in part, from sustained CSM interaction or a light echo, or both.SN 2010jl had exhibited previous indications of CSM interaction, but was no longer detectable in our Snapshot data.It is also worth mentioning the possibility that the sustained late-time luminosity could at least partially be due to radioactive decay of elements with longer lifetimes.
We have also detected the known resolved light echoes around SN 2012aw and SN 2016adj, and we note that their geometries have evolved since they were first discovered (Van Dyk et al. 2015;Stritzinger et al. 2022).
Of the four events that we consider to be SN impostors, all are still detectable in our Snapshots, implying that their eruptive behavior is persisting even at late times.Note that we have considered SN 2016jbu as an SN impostor, although Brennan et al. (2022a) concluded that the event may have actually been a terminal explosion and that the precursor has vanished.
Whereas we can likely infer from the few observed SNe Ia that their late-time emission was consistent with radioactive decay, for a significant number of corecollapse SNe II, CSM interaction may contribute to the luminosity even as late as ∼ 10, 000 d.This is not entirely surprising for the SNe IIn in our sample, and also to some extent for the SNe IIb, which have largely shown signs of CSM interaction at early times.However, for otherwise-normal SNe II-P, such as SN 2016bkv and SN 2017eaw, the SN shock unexpectedly continues to interact at ≳ 1000 d with the pre-existing CSM lost by the progenitor prior to explosion.The presence of such interaction provides important information about the extent of the CSM and the duration and nature of the mass loss(which can be further constrained through information gathered from the SN spectrum), with further implications for the evolution of the massive progenitor.
Snapshot surveys, such as ours, can efficiently provide a broad overview of the late-time properties of SNe and SN impostors and represent a reasonable use of valuable HST observing time.Approximately 70% of our originally proposed program was actually completed.The only wrinkle is that one has no control over which targets actually get executed, yet developing a relatively comprehensive sample is important in order to obtain a set of statistically significant results.Here we chose to target a large range of object types, to obtain knowledge of the late-time luminosity across a range of events from different astrophysical conditions.However, one could limit the sample to a large number of one particular SN type, nominally arising from a distinct progenitor population.Such is the case for HST programs pointedly targeting samples of SNe Ia (Graham et al. 2019;in Cycle 24) and SNe II (PI C. Kilpatrick in Cycle 30; PI W. Jacobson-Galán in Cycle 31).HST Snapshot programs have been executed specifically to detect light echoes at late times around SNe (PI P. Garnavich in Cycle 10, in this case those around SNe Ia).
A number of investigators have already exploited our publicly-available Snapshot data, and we have cited those studies in this paper, including our own spin-off study on disappearing progenitors (Van Dyk et al. 2023).We anticipate that other scientists will find this dataset valuable for their own use in the future, further proving that such surveys possess an archival legacy.To that end, we have examined our data for the possible detection of SNe other than the ones we had originally targeted, the sites of which are also serendipitously covered by our Snapshots.We provide a summary of those results in Section A in the Appendix.

A.4. SN 2006X
We managed to catch the light echo around the SN Ia 2006X (Wang et al. 2008;Crotts & Yourdon 2008) in our SN 2019ehk Snapshots (Section 4.31).The SN itself has disappeared, with upper limits of 26.2 and 25.8 mag in F438W and F625W, respectively; see Figure 37.An analysis of the evolution of the echo is beyond the scope of this paper.

A.5. SN 2006ov
The SN II-P 2006ov (e.g., Spiro et al. 2014) was serendipitously captured in our SN 2020jfo Snapshots in F555W and F814W (Section 4.35); see Figure 38.We ascertained the continued presence of the SN using HST ACS images from program GO-10877 (PI W. Li) obtained around the time of discovery.

A.6. SN 2012bv
The SN II 2012bv was serendipitously observed in our observations of SN 2017ixv (Section 4.25).There was no prior HST or optical ground-based imaging of the SN, and thus the absolute position was used to locate the site of the SN.The SN was no longer detectable in either band; see Figure 38.

Figure 1 .
Figure 1.A portion of the WFC3 image mosaic containing SN 1988Z, from observations on 2021 February 19, in (a) F625W and (b) F814W.Here, and in all other figures showing HST images in this paper, north is up and east is to the left; also, whenever the SN is visible, it is indicated by tick marks.Also shown are the ("Other") R (c) light curves from Aretxaga et al. (1999) and Turatto et al. (1993), together with the Snapshot detections from programs GO-13029 and GO-16179.

Figure 2 .
Figure 2. A portion of the WFC3 image mosaic containing SN 1993J, from observations on 2020 December 14, in (a) F336W and (b) F814W.Also shown are the ("Other") U and I (c) light curves from Richmond et al. (1996), together with prior HST data from Van Dyk et al. (2002), Fox et al. (2014), and previously-unpublished data from GO-13648 (PI O. Fox), as well as our Snapshots.et al. (2010) later detailed the occurrence of multiple outbursts from the star.Smith et al. (2011a) compared SN 2000ch to a host of other objects considered to be luminous blue variables or SN impostors, which may survive their eruptive outbursts.SN 2000ch has continued to experience brief, regularly recurring outbursts (Aghakhanloo et al. 2022a), and can fool transient hunters as being a new event (e.g., Van Dyk et al. 2013a).Our Snapshot observations were obtained in F555W (∼ V ) and F814W on 2020 December 13.As one can see in Figure3, the object is easily detectable in the HST images, and its light curve appears very much unlike that of a typical SN.Aghakhanloo et al. (2022a) have recently analyzed the continued photometric evolution of SN 2000ch, finding periodicity to the cycle of repeating outbursts, which suggests a binary nature for the

Figure 3 .
Figure 3.A portion of the WFC3 image mosaic containing SN 2000ch, from observations on 2020 December 13, in (a) F555W and (b) F814W.Also shown are the ("Other") V and I (c) light curves from Pastorello et al. (2010) and Aghakhanloo et al. (2022a), together with the Snapshot detections.

Fox
et al. (2017) detected SN 2010jl in HST images up to 1618 d (4.4 yr) after discovery.Our Snapshots in F336W and F814W from 2020 December 29 are when the SN is significantly older, at 4118 d (11.3 yr); see Figure 4. We located the SN 2010jl site via comparison with prior HST images obtained in 2015 October by GO-14149 (PI A. Filippenko), when the SN was at m F336W = 21.77± 0.04 and m F814W = 22.23 ± 0.02 mag,

Figure 4 .Figure 5 .
Figure 4.A portion of the WFC3 image mosaic containing SN 2010jl, from observations on 2020 December 29, in (a) F336W and (b) F814W.As the SN is not detected in either band, its location is indicated by the dashed circle.Also shown are the ("Other") U and I (c) light curves from Ofek et al. (2014) and Jencson et al. (2016), and F336W and F814W measurements from Fox et al. (2017), together with HST points from GO-15166 (PI A. Filippenko) and the Snapshot detections.
4.10.SN 2013ej SN 2013ej in M74 has been considered an atypical SN II, possibly an SN II-P/II-L hybrid (Mauerhan et al. 2017).A number of investigators have observed and analyzed the SN; see Van Dyk et al. (2023) and references therein.Early-time monitoring was also undertaken by KAIT; see de Jaeger et al. (2019).Our HST Snapshots were obtained in F555W and F814W on 2021 August 19, 2948 d (8.1 yr) after discovery.The location of the SN in our data was established based on Mauerhan et al. (2017, their Figure 11); see Figure 10.The SN 2013ej light curves in both bands have flattened out significantly at late times, showing essentially no decline in brightness (< 1 mag decline in both F555W and F814W) over more than 2000 d (5.5 yr).Van Dyk et al. (

Figure 8 .Figure 9 .
Figure 8.A portion of the WFC3 image mosaic containing SN 2012aw, from observations on 2021 February 17, in (a) F555W and (b) F814W.One can also clearly see the light echo surrounding the SN.Also shown are the Lick (de Jaeger et al. 2019) V and I (c) light curves, along with ("Other") data in these bands from Spogli et al. (2020), Bose et al. (2013), and Dall'Ora et al. (2014), together with the Snapshot detections.

Figure 12 .Figure 13 .
Figure 12.A portion of the WFC3 image mosaic containing SN 2015cp, from observations on 2020 November 30, in (a) F275W and (b) F625W.The SN was not detected in either band; the site is indicated by the dashed circle.Also shown is the iPTF R (c) light curve (adjusted from AB mag to Vegamag) and our rereduction of the HST F275W detection on 2017 September 12 from Graham et al. (2019), together with our Snapshot upper limits.

Figure 14 .
Figure 14.A portion of the WFC3 image mosaic containing SN 2016adj, from observations on 2021 July 28, in (a) F438W and (b) F555W.The SN was not detected in either band; the site is indicated by the dashed circle.What is most obvious in both bands are the light echoes around the SN site; see also Stritzinger et al. (2022).Also shown are ("Other") B and V (c) light-curve data(Stritzinger et al. 2023), together with our upper limits.We see a diffraction spike going straight through the F555W image, but it does not affect the SN site. .

Figure 15 .Figure 16 .
Figure 15.A portion of the WFC3 image mosaic containing SN 2016bkv, from observations on 2020 December 13, in (a) F555W and (b) F814W.Also shown are previously unpublished Lick V and I (c) light curves, along with ("Other") data including upper limits from Nakaoka et al. (2018) and Hosseinzadeh et al. (2018), together with the Snapshot detections.

Figure 17 .Figure 18 .
Figure 17.A portion of the WFC3 image mosaic containing SN 2016coi, from observations on 2020 December 6, in (a) F336W and (b) F814W.The SN was not detected in either band; the site is indicated by the dashed circle.Also shown is the Lick (Zheng et al. 2022) I (c) light curve, along with ("Other") data in I, as well as U , from Kumar et al. (2018), Terreran et al. (2019), Prentice et al. (2018), and Tsvetkov et al. (2020), together with the Snapshot upper limits.
4.21.SN 2017cfdSN 2017cfd in IC 511 was discovered on 2017 March 16 with KAIT and classified as a normal SN Ia(Han et al.

Figure 19 .Figure 20 .Figure 21 .
Figure 19.A portion of the WFC3 image mosaic containing SN 2016gkg, from observations on 2021 August 19, in (a) F438W and (b) F606W.The SN is detected in F606W, as indicated by tick marks, but not detected in F438W; the site in that band is indicated by the dashed circle.Also shown are the Lick (Bersten et al. 2018; Zheng et al. 2022) V and I (c) light curves, along with ("Other") data from Tartaglia et al. (2017b) and additional previous HST and B data from Kilpatrick et al. (2022), together with the Snapshot detections.

Figure 26 .
Figure 26.A portion of the WFC3 image mosaic containing SN 2018gj, from observations on 2021 January 27, in (a) F555W and (b) F814W.Also shown are unpublished Lick V and I (c) light curves, along with ("Other") data from Teja et al. (2023), together with the Snapshot detections.

Figure 27 .
Figure 27.A portion of the WFC3 image mosaic containing SN 2018zd, from observations on 2021 February 7, in (a) F606W and (b) F814W.The SN was not detected in either band; the site is indicated by the dashed circle.Also shown are V and I (c) light curves based on ("Other") data from Zhang et al. (2020), Hiramatsu et al. (2021), and Callis et al. (2021), together with F555W and F814W data from Hiramatsu et al. (2021) and the Snapshot upper limits.firmed using prior HST data taken on 2018 August 6 for GO-15600 (PI R. Foley), as well as from finding charts in prior literature on the object, such as Perley et al. (2019) and Margutti et al. (2019); see Figure 29.Chen et al. (2023) used our Snapshots as part of a larger work, looking at a variety of late-time observations (from 50 to 1423 d post-discovery) to better understand the object.They concluded that the nature of a putative black hole at the center of the accretion disk is still up for debate, given the various intriguing properties of the late-time emission.(Both Sun et al. 2022 and Sun et al. 2023 also made use of our Snapshot data.)The final identity of the precursor object therefore remains unknown.

Figure 28 .Figure 29 .
Figure 28.A portion of the WFC3 image mosaic containing SN 2018aoq, from observations on 2020 December 5, in (a) F555W and (b) F814W.The SN was not detected in either band; the site is indicated by the dashed circle.Also shown are previously unpublished Lick V and I (c) light curves, along with ("Other") data from O'Neill et al. (2019) and Tsvetkov et al. (2019), together with the Snapshot upper limits. .

Figure 30 .Figure 31 .
Figure 30.A portion of the WFC3 image mosaic containing SN2018ivc, from observations on 2020 November 27, in (a) F555W and (b) F814W.Also shown are previously unpublished Lick V and I (c) light curves, along with ("Other") data from Bostroem et al. (2020; which include F555W and F814W data from GO-15151, PI S. Van Dyk), together with the Snapshot detections.

Figure 32 .Figure 33 .
Figure 32.A portion of the WFC3 image mosaic containing AT 2019krl, from observations on 2021 February 15, in (a) F438W and (b) F625W.Also shown are B and R ("Other") (c) light curves from Andrews et al. (2021), together with the Snapshot detections.

Figure 34 .Figure 35 .
Figure 34.A portion of the WFC3 image mosaic containing SN 2020hvp, from observations on 2021 May 22, in (a) F555W (b) F814W.Also shown are previously unpublished Lick V and I (c) light curves, together with the Snapshot detections.

Figure 36 .
Figure 36.Left two panels: A portion of the WFC3 image mosaic containing SN 1999el, caught serendipitously in observations of SN 2020dpw (Section 4.33), in (a) F555W and (b) F814W.The SN is no longer detectable in either band; the site is indicated by the dashed circle.Right two panels: A portion of the WFC3 image mosaic containing SN 1999gi (e.g., Leonard et al. 2002), caught serendipitously in observations of SN 2016bkv (Section 4.15), in (a) F555W and (b) F814W.The SN is no longer detectable; the site is indicated by the dashed circle.

Figure 37 .
Figure 37. Left two panels: A portion of the WFC3 image mosaic containing SN 2000E (Valentini et al. 2003), caught serendipitously in observations of SN 2020dpw (Section 4.33), in (a) F555W and (b) F814W.The SN is no longer detectable; the site is indicated by the dashed circle.Right two panels: A portion of the WFC3 image mosaic containing SN 2006X, caught serendipitously in observations on 2021 February 21 of SN 2019ehk (Section 4.31), in (a) F438W and (b) F625W.Whereas the SN is no longer detectable (the site is indicated by tick marks), the light echo around it is still quite apparent.

Figure 41 .
Figure 41.Left two panels: A portion of the WFC3 image mosaic containing SN 2021J, caught serendipitously in observations on 2021 February 15 of SN 2013df, in (a) F336W and (b) F555W.The SN, which was at V = 13.63 and I = 12.70 mag (Gallego-Cano et al. 2022), is hopelessly saturated in both bands; hence, the Snapshots do not provide any additional photometric information.Right two panels: A portion of the WFC3 image mosaic containing the location of SN 2021sjt, caught serendipitously in observations of SN 2020dpw, in (a) F555W and (b) F814W.A progenitor candidate is not detectable, as indicated by the dashed circle.

Figure 42 .
Figure 42.A portion of the WFC3 image mosaic containing the location of SN 2022aau, caught serendipitously in observations of SN 2017gax, in (a) F336W and (b) F814W.The progenitor is not detectable, as indicated by the dashed circle.

Table 1 .
Properties of the Targeted Events and Their Hosts

Table 2 .
Photometry of Seven Supernovae with Unpublished Data (mag) 1