Lensed Type Ia Supernova “Encore” at z = 2 : The First Instance of Two Multiply-Imaged Supernovae in the Same Host Galaxy

A bright ( m F150W , AB = 24 mag), z = 1 . 95 supernova (SN) candidate was discovered in JWST /NIRCam imaging acquired on 2023 November 17. The SN is quintuply-imaged as a result of strong gravitational lensing by a foreground galaxy cluster, detected in three locations, and remarkably is the second lensed SN found in the same host galaxy. The previous lensed SN was called “Requiem”, and therefore the new SN is named “Encore”. This makes the MACS J0138.0 − 2155 cluster the first known system to produce more than one multiply-imaged SN. Moreover, both SN Requiem and SN Encore are Type Ia SNe (SNe Ia), making this the most distant case of a galaxy hosting two SNe Ia. Using parametric host fitting, we determine the probability of detecting two SNe Ia in this host galaxy over a ∼ 10 year window to be ≈ 3% . These observations have the potential to yield a Hubble Constant ( H 0 ) measurement with ∼ 10% precision, only the third lensed SN capable of such a result, using the three visible images of the SN. Both SN Requiem and SN Encore have a fourth image that is expected to appear within a few years of ∼ 2030 , providing unprecedented baseline for time-delay cosmography.


INTRODUCTION
Strong gravitational lensing can cause multiple images of a background source to appear, as light propagating along different paths are focused by the gravity of a foreground galaxy or galaxy cluster (called the "lens"; e.g., Narayan & Bartelmann 1997).Such a phenomenon requires chance alignment between the observer, the background source, and the lens.If the multiply-imaged source has variable brightness then depending on the relative geometrical and gravitational potential differences of each path, the source images will typically appear delayed by hours to weeks (for galaxy-scale lenses, M ≲ 10 12 M ⊙ ) or months to years (for cluster-scale lenses, M ≳ 10 13 M ⊙ ; e.g., Oguri 2019).
Precise measurements of this "time delay" yield a direct distance measurement to the lens system that constrains the Hubble constant (H 0 ) in a sin-  1).
The filters used are F105W+F115W+F125W (blue), F150W+F160W+F200W (green), and F277W+F356W+F444W (red).The images were drizzled to 0. ′′ 02/pix, and the image scale and orientation are as shown.The three detected image positions of SN Encore are circled, but it is not visible at this scale, and all five images of the host galaxy are labeled (A-E).For a zoom in of each of the detected image positions, see Figure 3 (Image Credit: STScI, A. Koekemoer, T. Li).
"SN Requiem" (Rodney et al. 2021) was the first cluster-scale photometrically classified multiplyimaged SN Ia, appearing at z = 1.95 in the MRG-M0138 galaxy that is lensed by the MACS J0138.0−2155cluster (Figure 1).Unfortunately the SN was found archivally several years after it had faded, precluding an H 0 measurement with the visible images.However, there is a fourth image of SN Requiem expected in ∼ 2035, an astounding ∼ 20 year time delay (Rodney et al. 2021).In a truly remarkable turn of events, a  The observed light curve in the JWST/NIRCam short-wavelength filters for image A, the brightest and last to arrive (see Figure 3).The longwavelength filters (F277W, F356W, F444W) were observed but a more in-depth host galaxy modeling effort (or a template image) will be required to measure accurate photometry.
second lensed SN Ia (dubbed SN Encore) has been found by the James Webb Space Telescope (JWST) in the same host galaxy as SN Requiem at z = 1.95.The discovery visit was on 2023 November 17 (JWST-GO-2345, PI Newman), which included JWST NIRCam imaging and NIRSpec IFU observations.Previous observations of MACS J0138.0−2155 and MRG-M0138 are described in Newman et al. (2018, hereafter N18), and SN Requiem is described in Rodney et al. (2021).
This work presents the discovery and observations of SN Encore, and is the start of a series of papers analyzing the system.The outline of this paper is as follows: Section 2 summarizes the Hubble Space Telescope (HST) and JWST observations taken to follow SN Encore.Section 3 leverages the spectrum from Section 2 to identify SN Encore as an SN Ia, while Section 4 discusses our expectations surrounding the discovery of a second lensed SN Ia in the same host galaxy.We conclude with a discussion of the implications for SN Encore in Section 5. SN Encore was discovered by the JWST-GO-2345 team in F150W NIRCam imaging taken 2023 November 17 (MJD 60265) by comparing the data with an archival HST WFC3/IR F160W image (2016 July 18, MJD 57587; N18).These two filters are extremely well-matched in wavelength and transmission, and the source was very bright (≲ 24 mag AB in both images A and B, see Figure 3) in F150W but absent in F160W.However, the separation of SN Encore from its host galaxy MRG-M0138 is only ∼ 0 ′′ .1, which is roughly the pixel-scale of the HST WFC3/IR imaging.We therefore also compared the point-source positions and relative brightnesses to lens model predictions (Ertl et al., Suyu et al. in preparation), and confirmed with forced photometry that there was an increase in flux corresponding to the apparent brightness of SN Encore between the F160W and F150W imaging.All tests were in agreement that this was indeed a multiply-imaged SN, with (at least) two images of the SN detected (images A and B, see Figure 3).Further efforts to remove host galaxy light near image C revealed that a third image of SN Encore was also detected.There two additional images of the host galaxy located in the northeast of the cluster (Figure 1), with the next image of SN Encore expected from our lens modeling to arrive in about a decade (Ertl et al. in preparation).The sharp radial profile of images D and E for MRG-M0138 will provide relatively tight constraints on these long time delays, which will enable a targeted follow-up campaign similar to SN Refsdal (Kelly et al. 2016).N18 reported a spectroscopic redshift for MRG-M0138 of z = 1.95, and the host properties alone suggested that SN Encore was likely an SN Ia at z = 1.95 (Rodney et al. 2021).

JWST Data
2.1.1.Imaging An HST program was triggered soon after the discovery of SN Encore, which is described in the following section.However, JWST was still re-quired for two reasons: 1) the small host separation described above meant that even with difference imaging, it would be very difficult to detect SN Encore in any but the brightest filter (F160W for HST), and 2) based on the discovery epoch SN Encore was between peak brightness and the second infrared (IR) maximum for SNe Ia (Hsiao et al. 2007;Krisciunas et al. 2017;Pierel et al. 2022).At z = 1.95 HST is only able to cover rest-frame uBV filters, meaning JWST was needed to reach the rest-frame near-IR and leverage the second infrared maximum for time-delay measurements.
A disruptive JWST director's discretionary time (DDT) proposal was subsequently approved (JWST-GO-6549, PI Pierel), which provided three additional epochs of NIRCam imaging in six filters (Table 1).The first NIRCam epoch occurred on 2023 December 5 (MJD 60283), 18 (∼ 6 restframe) days after discovery.The next two epochs took place with a cadence of ∼ 17 observer-frame (∼ 6 rest-frame) days, on 2023 December 23 (MJD 60301) and 2024 January 8 (MJD 60317).This meant the JWST light curve for each image contains four observations, each separated by ∼ 6 restframe days.Although the image C of SN Encore is too faint to see clearly by eye in Figure 3, it is ∼ 27 AB mag in F150W and detected at ∼ 5σ with careful subtraction of the host galaxy light profile in JWST imaging.The full JWST dataset therefore effectively produces a SN Ia light curve with 12 epochs and rest-frame UV-NIR wavelength coverage, which is sufficient to accurately measure phases of each image and therefore the relative time delays (e.g., Pierel & Rodney 2019).The summary of observations is given in Table 1.
All the NIRCam data were retrieved from the STScI MAST1 archive, and were processed and calibrated using the JWST Pipeline2 (Bushouse et al. 2022) version 1.12.5, with CRDS reference Figure 3. JWST/NIRCam color cutouts with F115W (blue), F150W (green) and F200W (red) centered on image A (top), B (middle), and C (bottom) of SN Encore.Note the discovery program only detected SN Encore in F150W, so the first column is a grey-scale image in F150W, while the remainder are 3-color images.The images were drizzled to 0 ′′ .02/pix,with the image scale, orientation, and modified julian date (MJD; relative to discovery) are shown.SN Encore is brightest in the discovery epoch and is on the decline in the subsequent epochs, but the evolution is relatively small at this redshift and is therefore difficult to see by eye (see Figure 2).Image C was first to arrive followed by image B, and although image C is not clearly visible by eye, it was detected at ∼ 5σ (See Section 2.1.1).
files defined in 1180.pmap, with additional improvements described here as follows.As part of the initial processing with the Stage 1 portion of the pipeline, additional corrections were applied to remove 1/f noise as well as low-level background variations across the detectors, including the removal of wisps and other low-level artifacts, with these techniques all described in more detail in Windhorst et al. (2023).The Stage 2 portion of the pipeline was then run to apply photometric calibration to physical units, using the latest photometric calibrations (Boyer et al. 2023).The images were then all astrometrically aligned directly to the Gaia-DR33 reference frame and finally combined into mosaics for each filter with all distortion removed, using the Stage 3 portion of the pipeline, where these mosaics were constructed for each of the individual epochs, as well as full-depth versions combining all the epochs.

Spectroscopy
The JWST-GO-2345 program also included NIRSpec integral field unit (IFU) spectroscopy for image A of MRG-M0138 using both the G140M and G235M gratings.There was a failure in the G140M observation causing it to be scheduled 39 days later on 2023 December 27, but the G235M observation was successful.Here we extract the SN spectrum for the G235M grating, aligned to the G140M spectrum using their overlapping wavelength range, with the G140M spectrum anchored to the F150W NIRCam image to reveal the SN location.The SN spaxels have unique spectral fea-tures in the G235M observation, distinct from the galaxy spectra, making it possible to disentangle.We establish contour levels for each IFU datacube by calculating the median flux across the entire wavelength range, which allows us to subtract the host galaxy spectra from the spaxels containing SN flux.Finally, we isolate the SN spectrum by subtracting off the median spectrum of host galaxy from what is observed in the SN spaxels.We leave discussion of the G140M spectrum and in-depth analysis of the G235M spectrum to the upcoming SN spectroscopic analysis (Dhawan et al. in preparation), host galaxy analysis (Newman et al. in preparation), and investigations into relationships between the host galaxy and SN Ia properties (Larison et al. in preparation).The G235M spectrum is used briefly here to simply confirm the spectroscopic SN type for SN Encore in Section 3, and is shown in Figure 4.

HST Data
LensWatch4 is a collaboration with the goal of finding gravitationally lensed SNe, both by monitoring active transient surveys (e.g., Fremling et al. 2020;Jones et al. 2021) and by way of targeted surveys (Craig et al. 2021).The collaboration maintained a Cycle 28 HST program (HST-GO-16264) given long-term (three-cycle) target-ofopportunity (ToO) status.The program included three ToO triggers (two non-disruptive, one disruptive), and was designed to provide high-resolution follow-up imaging for a ground-based lensed SN discovery, which is critical for galaxy-scale multiply imaged SNe due to their small image separations (e.g., Goobar et al. 2017).LensWatch triggered HST-GO program 16264 on 2023 November 20 (MJD 60268, three days after discovery) to obtain follow-up imaging of SN Encore, including 14 orbits in three filters spread over two epochs.The trigger was disruptive but there were small delays related to HST being in safe mode, which led to the first epoch being 26 (∼ 9 rest-frame) days later on 2023 December 16 (MJD 60294) and the second epoch 45 days after the first on 2024 January 30 (MJD 60339).These observations are separated by 10-22 observer-frame (3-7 rest-frame) days from the JWST observations described in the previous section, and were designed to be complementary to JWST.All observations are summarized in Table 1.
All the HST data were retrieved from the STScI MAST archive, and the calibrated exposures were subsequently processed with additional improvements described here as follows.The default astrometry from the archive was corrected by realigning all the HST exposures to one another, as well as to the JWST images and to Gaia-DR3, using an updated version of the HST "mosaicdrizzle" pipeline first described in Koekemoer et al. (2011), which also removes residual low-level background variations across the detectors.This pipeline was then used to combine all the HST data into mosaics for each filter with all distortion removed, producing mosaics for the separate epochs as well as fulldepth versions combining all the epochs.

CLASSIFYING SN ENCORE
Lensed SNe Ia offer smaller uncertainties for H 0 than lensed core-collapse SNe (CC SNe) (e.g., Goldstein et al. 2018;Pierel & Rodney 2019;Birrer et al. 2022a) and unique leverage on lens modeling systematics (Pascale et al. 2024), making them the most valuable type of lensed SN.We infer the type of SN Encore using the supernova identification (SNID; Blondin & Tonry 2007) and Next-Generation SuperFit (NGSF)5 packages with the observed G235M spectrum (Table 1, Figure 4).Both packages classify SN Encore as Type Ia with > 90% confidence (9 of the top 10 bestfit reference spectra are from SNe Ia and the last is a SN Ia-91T, which is a subclass known to be spectroscopically similar to normal SNe Ia at late times), with the best matches being SN 2004eo (SNID) and SN 2007sr (NGSF).Both of these SNe decline faster than the mean decline rate (SALT3 x 1 ≲ 0), which is in agreement with light curve fitting for SN Encore (x 1 ∼ −1.3), but all three SNe exhibit light curve shapes that are well-within traditional cosmology cuts (e.g., −3 < x 1 < 3 Scolnic et al. 2018).NGSF also returns an estimate of the host galaxy contamination in the SN spectrum, which is < 1%.Finally, the phase is constrained by the classification codes to 26.5 ± 5.7 rest-frame days, which is in 1.1σ agreement with the results of our light curve fitting (20.1 ± 3 rest-frame days; Figure 2).Overall the result of this analysis is therefore that SN Encore is a SN Ia, and moreover is likely a relatively fast-declining SN that should nevertheless be standardizable.A full analysis of both the G235M and G140M spectrum, a comparison to low-z SN Ia spectra and theoretical models for SN Ia progenitors, will be presented in Dhawan et al. (in preparation).
While the most precise time-delay measurements come from lensed SNe observed before peak brightness, a lensed SN Ia observed roughly in the phase of images A and B in multiple filters should still provide a time-delay measurement uncertainty of ∼ 5 observer-frame days (Pierel & Rodney 2019).Image C was observed well after peak brightness in the nebular phase of SNe Ia, where the slow and uncertain evolution will likely result in a lower time-delay precision of ≳ 10 observer frame days (Pierel et al. 2024).Initial expectations from lens models of the MACS J0138.0−2155cluster and SN Encore photometry suggest the delay between images A and B will be ≳ 1 month, while the delay between images A and C will be ∼ 1 year.Assuming a fiducial cluster lens model uncertainty of ∼ 6% (Grillo et al. 2020;Kelly et al. 2023a), SN Encore should produce an H 0 uncertainty of ∼ 10%.A detailed lens modeling analysis will be presented by Ertl et al. and Suyu et al. (in preparation), while measurements of the time-delay and H 0 will be presented by Pierel et al. (in preparation)

TWO LENSED SNE IA FROM A SINGLE HOST GALAXY
While it seems remarkable to find two lensed SNe Ia in the same host galaxy at z = 1.95, we briefly explore the probability of such a discovery.A full analysis of the relationship between the local/global properties of MRG-M0138 and both SN Encore and Requiem will be presented by Larison et al. (in preparation), leveraging the full dataset shown in Table 1.Here, we simply provide estimates assuming the results of N18.
N18 describes the specifics of their data acquisition, and here we use their host properties derived from a near-IR spectrum from the MOSFIRE spectrograph instrument on the Keck I telescope.
N18 have already provided best-fit SFHs by fitting the MRG-M0138 spectrum with a parametric SFH model.The parametric model they use is of the form, SFR ∝ exp(−t/τ ).Table 5 in N18 provides the τ value for the SFH based on two images of MRG-M0138 -we adopt an average value of 180 Myr which is consistent with fits from both images.
There are one of two approaches we investigate here.In the first approach, we assume the SN Ia delay-time distribution (DTD) to be an exponential function from Strolger et al. (2020), and convolve it with our SFHs derived above to obtain an SN Ia rate for MRG-M0138.The SN Ia rate calculation also requires assumptions to be made about the form of the initial mass function (IMF) and the fraction of stars within 3-8 M ⊙ that are ultimately successful SN Ia explosions (ϵ).Following Strolger et al. (2020), we assume an IMF with a powerlaw index of α ≈ −2.3 (Salpeter 1955;Kroupa 2001) and ϵ=0.06.
MRG-M0138 is quiescent and not exceptionally star forming in the observed epoch, at a rate between ≈ 1.8 and 2.6 (±1.7)M ⊙ yr −1 , depending on which image is used.Using the method above implies a SN Ia rate of approximately 90 events (1000 yr) −1 , which is consistent with the average rate of SNe Ia per galaxy, regardless of mass, from Li et al. (2011).Assuming effectively a 10-year monitoring of the host, i.e., in which no detectable SN Ia event was missed in the last 10 years (3.39 years in the rest frame of the host), it would be reasonable to expect the probability of having one SN Ia in that time frame to be low, but not zero, with Pois(k = 1) = 0.23.Similarly, the probability of finding two such events is also low, with Pois(k = 2) = 0.03.It is interesting to note that assuming a power-law delay time distribution (e.g., Maoz et al. 2012;Freundlich & Maoz 2021) would make the rate at the observed epoch even lower, 5.3 events (1000 yr) −1 , and the likelihood of observing two supernovae in MRG-M0138 in 10 years, Pois(k = 2) = 0.0001.However, another approach could be to estimate the expected yield from the mass-weighted rates as is done in Smith et al. ( 2020) and Scolnic et al. (2020).This has the convenience of estimating the rate from the galaxy's current properties, but loses the detailed analysis provided by the Strolger et al. (2020) approach. Li et al. (2011) show an average mass-weighted SN Ia rate per galaxy of 0.54±0.12events per century per 10 10 M ⊙ .N18 determine log M * /M ⊙ = 11.7 ± 0.2 (for both images of the host), resulting in a expected rest-frame yield of about 260 (1000 yr) −1 , or 0.88 events in the 10year observation window.The probability of having one SN Ia in that time-frame would therefore be a bit higher, with Pois(k = 1) = 0.37, and the probability of finding two such events seemingly more feasible, at ≈ 16%.Both approaches are broadly consistent with the rate of discovery of SN Ia siblings from Kelsey (2024) and Scolnic et al. (2020).

DISCUSSION
Here we presented the discovery and observations of SN Encore, a SN lensed by the MACS J0138.0−2155cluster and hosted by the galaxy MRG-M0138 at z = 1.95.We spectroscopically confirm that SN Encore is of Type Ia, making it just the second lensed SN Ia and third overall lensed SN useful for measuring H 0 with ≲ 10% uncertainty.Remarkably, SN Encore is the second lensed SN Ia hosted by MRG-M0138 in the last ∼ 7 years, making it the first known case of two lensed SNe from the same host galaxy.While the appearance of two SN Ia in the same multiply imaged galaxy is remarkable, MRG-M0138 was targeted because it is by far the brightest and most massive (M * = 5 × 10 11 M ⊙ ) among the rare class of magnified quiescent galaxies at z ≳ 2 (Toft et al. 2017;Akhshik et al. 2023).MRG-M0138's extreme stellar mass, old stellar population, and lack of strong dust attenuation lead to a low but non-negligible derived probability of ≲ 3% for detecting two SNe Ia in the last ∼decade.SN Requiem and SN Encore will both reappear in the next ∼5-10 years, providing an exceedingly rare opportunity to monitor MACS J0138.0−2156 for two SNe Ia with known reappearance location and time.The resulting baseline will be unprecedented for time-delay cosmography and can provide unique insights into SN Ia physics at z = 1.95.
While SN Encore will not break the H 0 tension (e.g., Riess et al. 2022) on its own, it still plays a critical role as the third lensed SN (and second lensed SN Ia) to provide a precise measurement of H 0 .With a sample of three the combined H 0 uncertainty could plausibly be ∼ 5%, and we will start to have a clearer picture of the relative agreement across lensed SN measurements of H 0 .Upcoming surveys such as the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST;Ivezic et al. 2019;Arendse et al. 2023) and the Nancy Grace Roman Space Telescope High Latitude Time Domain Survey (HLTDS; Pierel et al. 2021;Rose et al. 2021) are expected to deliver dozens of useful galaxy-and cluster-scale multiplyimaged SNe, and although more complicated, clusters provide longer time-delays, wider image separations, and extra lens modeling constraints on the mass-sheet degeneracy (Falco et al. 1985;Grillo et al. 2020Grillo et al. , 2024;;Pascale et al. 2024).As a result, cluster-lensed SNe are likely to continue as an extremely valuable subset of the future cosmological sample.
tional Science Foundation (NSF); by the Ministry of Science & Technology, Israel; and by the Israel Science Foundation Grant No. 864/23.

Figure 1 .
Figure 1.Left: HST WFC3/IR two-color image in the region of MACS J0138.0−2155 from 2016, using F105W (blue) and F160W (orange).SN Requiem is marked in its three visible image positions by white circles, notably absent in 2023.Right: Combined JWST/NIRCam and HST/WFC3 color image from programs 6549 and 16264 (Table1).The filters used are F105W+F115W+F125W (blue), F150W+F160W+F200W (green), and F277W+F356W+F444W (red).The images were drizzled to 0. ′′ 02/pix, and the image scale and orientation are as shown.The three detected image positions of SN Encore are circled, but it is not visible at this scale, and all five images of the host galaxy are labeled (A-E).For a zoom in of each of the detected image positions, see Figure3(Image Credit: STScI, A. Koekemoer, T. Li).

Table 1 .
Summary of observations taken of SN Encore. .