COOL-LAMPS II. Characterizing the Size and Star Formation History of a Bright Strongly Lensed Early-Type Galaxy at Redshift 1.02

We present COOL J1323+0343, an early-type galaxy at $z = 1.0153 \pm 0.0006$, strongly lensed by a cluster of galaxies at z = $z = 0.353 \pm 0.001$. This object was originally imaged by DECaLS and noted as a gravitational lens by COOL-LAMPS, a collaboration initiated to find strong-lensing systems in recent public optical imaging data, and confirmed with follow-up data. With ground-based grzH imaging and optical spectroscopy from the Las Campanas Observatory and the Nordic Optical Telescope, we derive a stellar mass, metallicity, and star-formation history from stellar-population synthesis modeling. The lens modeling implies a total magnification of $\mu \sim $113. The median remnant stellar mass in the source plane is M$_* \sim 10.63$ $M_\odot$ and the median star-formation rate in the source plane is SFR $\sim 1.55 \times 10^{-3}$ M$_\odot$ yr$^{-1}$ (log sSFR = -13.4 yr$^{-1}$) in the youngest two age bins (0-100 Myr), closest to the epoch of observation. Our measurements place COOL J1323+0343 below the characteristic mass of the stellar mass function, making it an especially compelling target that could help clarify how intermediate mass quiescent galaxies evolve. We reconstruct COOL J1323+0343 in the source plane and fit its light profile. This object is below the expected size-evolution of early-type galaxy at this mass with an effective radius r$_e \sim$ 0.5 kpc. This extraordinarily magnified and bright lensed early-type galaxy offers an exciting opportunity to study the morphology and star formation history of an intermediate mass early-type galaxy in detail at $z \sim $1 .


INTRODUCTION
Untangling the mechanisms that fuel the evolution of early-type galaxies (ETGs) is a key component to understanding how the diverse population of galaxies in the local Universe formed. The discovery that ETGs at z > 1 are much Author for correspondence: sukay@uchicago.edu more compact than those in the local Universe, with radii between 3-5 times smaller without much change in mass, is a challenge to our understanding of galaxy evolution (Daddi et al. 2005, Trujillo et al. 2006, Trujillo et al. 2007, Newman et al. 2010). There are two theories proposed to explain this growth. First, mergers with small, gas-poor satellite galaxies-known as minor mergers. Second, internal mechanisms like adiabatic expansion. Minor dry mergers would increase the radii of ETGs with-out requiring the addition of a proportional amount of mass (Bezanson et al. 2009, Naab et al. 2009) and some studies found that they are consistent with observations of ETGs at z < 1.6 (Belli et al. 2014). Minor mergers may explain the evolution seen at low redshifts, but they are insufficient to explain the rapid evolution and the scatter in radii at higher redshifts (Fan et al. 2010, Newman et al. 2012, Nipoti et al. 2012. Furthermore, surveys have struggled to find the number of companion satellites required (Newman et al. 2012). Adiabatic expansion triggered by AGN feedback, in combination with dry mergers, might resolve these inconsistencies (Fan et al. 2010).
Another potential explanation is progenitor bias-the idea that the processes that quench ETGs at lower redshifts are different from those at z > 2 and, as a result, latecomers to the ETG population have larger radii. Number density studies found strong evidence that progenitor bias is not sufficient to explain the growth of massive ETGs (Newman et al. 2010, Belli et al. 2014, Belli et al. 2015. For intermediate-mass ETGs (those with 10.5 < logM * < 11), some studies found that progenitor bias explains the majority of the observed evolution after z = 1 (Carollo et al. 2013, Fagioli et al. 2016). However, many other studies found evidence that individual growth is needed, at least in part, to explain the growth of intermediate-mass ETGs (Cassata et al. 2011, Newman et al. 2012, Whitaker et al. 2012, Belli et al. 2014, van der Wel et al. 2014, Belli et al. 2015. The mechanisms that fuel the evolution of ETGs after they quench affects the morphology of evolving objects. Dry mergers with low mass objects result in central regions that have similar densities to very compact ETGs at z ∼ 2, with an envelope of low-density material, high Sérsic indices (n ≥ 5), and negative metallicity gradients (Hopkins et al. 2009, Hilz et al. 2013. Adiabatic processes caused by quasar feedback would make the central regions of ETGs less dense in the local Universe than they are at z ∼ 2 (Fan et al. 2010). Indepth morphological studies will also provide clues to how ETGs quench in the first place. Simulations suggest galaxies that quenched "inside-out" through a central starburst have younger central stars than those on the edge (Wellons et al. 2015). However, galaxies that quenched "outside-in" through cold gas accretion have stellar ages that are the same throughout or older central stars (Feldmann et al. 2016).
Spatially resolved imaging and spectroscopy targeting ETGs from 0.5 < z < 2 should allow us to understand how they quench and what processes drive their structural evolution after star formation ceases. It is difficult to spatially resolve the most compact systems that may be littlemodified descendants of compact high-redshift ETGs (e.g., Stockton et al. 2014). Strong gravitational lensing enables study of more representative quiescent galaxies with better spatial resolution and signal-to-noise (e.g., Oldham et al. 2017, Akhshik et al. 2020, Akhshik et al. 2021, Man et al. 2021. Taking advantage of lensing magnification, Akhshik et al. (2020Akhshik et al. ( , 2021 were able to measure the age gradients and SFHs across seven spatial bins in a massive quiescent galaxy at z = 1.88. More examples of lensed ETGs, particularly with large magnifications that enable detailed studies, is key to further progress.
We discuss here the discovery and initial characterization of COOL J1323+0343: a compact intermediate mass ETG with old stellar populations which would be near impossible to study in-depth without strong gravitational lensing. COOL J1323+0343 extraordinary magnification of µ > 100 offers the opportunity to study a representative ETG at z ∼ 1, and makes it an especially compelling target for more detailed follow-up imaging and spectroscopy.
This paper is structured as follows. Section 2 briefly describes the discovery of COOL J1323+0343.
Section 3 describes the follow-up imaging and spectroscopy of COOL J1323+0343. In Section 4, we report the results of stellar-population synthesis modeling. The lens modeling and source plane reconstruction are described in Section 5. Section 6 puts COOL J1323+0343 in context with other ETGs and discusses what we expect to learn from detailed follow-up.
All reported magnitudes are calibrated to the AB system. The fiducial cosmology model used assumes a standard flat cold dark-matter model with a cosmological constant (Λ CDM), corresponding to WMAP9 observations (Hinshaw et al. 2013). For inferred parameters with uncertainties, we report 16th, 50th and 84th percentile values, unless otherwise specified.
2. DISCOVERY COOL J1323+0343 (hereafter CJ1323) (13h23m04.12s 03 • 43'19.4") was discovered in a search for strong lenses in the Dark Energy Camera Legacy Survey (DECaLS) Data Release 8 (DR8, Dey et al. 2019) by COOL-LAMPS -ChicagO Optically-selected strong Lenses -Located At the Margins of Public Surveys. Details about this search and results will be described in a future publication, so we only provide brief details here.
We found CJ1323 as part of a comprehensive visual search of the northern galactic cap portion of the southern DECaLS dataset. Specifically, CJ1323 was found in a search that examined lines of sight centered on luminous red galaxies, out to a photometric redshift of z ∼ 0.7. Both the foreground lensing cluster and the lensed source were immediately apparent in the DECaLS images; six co-authors examined this particular field yielding an aggregate score of 2.4 on a scale of 0 to 3, where a score of 3 signifies a definite lens.
The foreground lens was first noted as a cluster by Hao et al. (2010) and was included in several other more re-cent cluster catalogs (e.g., Rykoff et al. 2014, Hilton et al. 2021. CJ1323 was independently discovered as a strong lensing candidate by Huang et al. (2021), and noted as DESI-200.7678+03.7216. They first searched DR7 with a residual neural network (Lanusse et al. 2018) using a training sample consisting of only observed images, but did not find this lens (Huang et al. 2020). Using an improved "shielded" model with a larger training set on the complete DR8 dataset, the neural net presented in Huang et al. (2021) gave CJ1323 a probability of 0.34, well above their threshold of 0.1. Through visual inspection, it was assigned a final grade of 4 on a scale of 1 to 4, which indicates it is among the systems showing the clearest evidence of strong lensing.

Imaging
Near-infrared imaging of CJ1323 in the H-band was obtained using the FourStar Infrared Camera (FOURSTAR; Persson et al. 2008) on the Magellan/Baade telescope, Chile, on February 22nd, 2020. The total integration time was 183s; the apparent brightness despite the briefness of this total time is a testament to the brightness of the lensed source images. We reduced the data to final astrometrically-and photometrically-calibrated stacked images using a custom pipeline built via IRAF (Tody 1986, Tody 1993) and PHOT-PIPE (Rest et al. 2005, Garg et al. 2007, Miknaitis et al. 2007). We show a color image combining the H-band image with the DECaLS grz data in Figure 1.
The H-band image was calibrated to 2MASS stars (Skrutskie et al. 2006) within the field of view, with the calibration derived automatically by PHOTPIPE routines. Uncertainty on the zeropoint relative to 2MASS is ∼ 0.02 magnitudes. We used the provided zeropoints for the DECaLs data; the uncertainties on these values were insignificant compared to other measurement uncertainties.
While the photometric analysis in this paper was completed primarily with the DECaLs optical imaging and FOURSTAR near-IR imaging, late in the preparation of this manuscript we obtained grz-band imaging. The images were taken with the Low Dispersion Survey Spectrograph (LDSS-3 1 ) on the Magellan/Baade telescope, Chile, on January 16th, 2021, for a total integration time of 360s per filter, in subarcsecond seeing conditions. These higher resolution images were used to confirm the lensing configuration implied by the earlier data, as described in Section 5.

Spectroscopy
Spectroscopic observations of CJ1323 were obtained on April 19th, 2020, using the Alhambra Faint Object Spec-1 http://www.lco.cl/?epkb_post_type_1=ldss-3-user-manual Figure 1. gr(z + H) image of COOL J1323+0343. This image was constructed using grz-band imaging from DECaLS and H-band imaging from FOURSTAR on Magellan. The lensed galaxy is seen as a bright red multiply-imaged source, with three images visible in this figure. The foreground cluster has two similarly bright central galaxies. The bright blue source between the two brightest cluster galaxies is a foreground star. trograph and Camera (ALFOSC) at the 2.56m Nordic Optical Telescope (NOT). Two 2400s exposures were obtained using Grism #20 (R = 590, λ = 5650 − 10150Å), using a 1. 3 longslit. Halogen lamps were used for spectroscopic flat fielding, and wavelength solutions were calibrated using HeNe and ThAr arc lamps. Spectrophotometric calibration was performed using observations of the standard star SP1045+378. The longslit was placed to sample the most northern and most southern apparent images of the lensed source, and a dither along the slit was made between the two exposures sufficient to place both spectra of the source on slit regions that were clear of other objects.
Reduction to one dimensional wavelength-and fluxcalibrated spectra was accomplished using standard routines in IRAF. Sky-subtraction was ultimately accomplished using A-B subtraction of the two dithered spectra; while we explored sky-subtraction using adjacent sky apertures, we found they gave poor results. Light from the star and lens galaxy in between the two source images (see Figure 1) contaminates regions that might otherwise be used as sky apertures and limits the accuracy with which the sky can be subtracted in this case.

Redshifts
CJ1323 is lensed by GMBCG J2007+03722 (Hao et al. 2010), a galaxy cluster indicated by the abundance of redsequence galaxies (Gladders & Yee 2000) easily visible in Figure 1 as elliptical galaxies with a similar orange hue. The two brightest of these galaxies near the cluster center have redshifts reported in the SDSS Legacy Survey at z = 0.3535±0.0001 (York et al. 2000;Strauss et al. 2002) and by BOSS at z = 0.35256 ± 0.00006 (Eisenstein et al. 2011;Dawson et al. 2013) for the east and west galaxies, respectively. We took the cluster redshift as the average of these two values.
The redshift of the lensed galaxy CJ1323 is z = 1.0153 ± 0.0006, based on Ca H & K, H-δ, and G-band features clearly visible in the NOT-ALFOSC spectrum shown in the bottom panel of Figure 2.

Model Photometry with GALFIT
We used the parametric fitting code GALFIT (Peng et al. 2002(Peng et al. , 2010 to fit light profiles to lensing cluster galaxies, other line-of-sight contaminants, and the targets of interest. For the H-band, we followed the process for making pointspread functions (PSFs) and utilizing GALFIT as described in Khullar et al. (2021). We used one or more 2D Sérsic components to model the light of the components of the arc, galaxy cluster, and other objects. The foreground star between the cluster galaxies was fit with a PSF and an additional Sérsic component to account for residual differences between this star and nearby isolated stars that were used to construct a reference PSF. Statistical magnitude uncertainties were measured as described in Khullar et al. (2021), with the final best fit model image injected into blank regions of the image, and then refit, with the distributions of results from these inject-and-recover tests providing the uncertainties. Additionally, models were built independently by three co-authors, and we found that the systematic uncertainties induced by the decision process inherent in this type of iterative model building was insignificant for the H-band measurements.
However, a similar initial analysis of the coarser DECaLS imaging indicated a significant issue with modeling systematics. Due to the overlapping mosaic of sampling from individual integrations that comprise the DECaLS images, there was no nearby isolated bright star to use as a reference, and hence in these data we were forced to use the star in the center of the field as a PSF reference. This was accomplished by first fitting the central region with a GALFIT model, absent any PSF convolution, with the star itself described by a twocomponent Moffat profile. The best fit model of the star was then extracted and used as a PSF reference for a complete GALFIT model of field. This limitation was possibly the source of the observed systematic differences between completed models. To investigate this further, we used GALFIT to fit nearby galaxies and stars with the targeted lensed images masked so that, after modeling, only the arc remained in the residual image. The photometry was then measured from this residual image using complex arc-like apertures at various scales, in order to measure the photometric curve of growth and a total lensed image magnitude (e.g., Wuyts et al. 2010). This second approach produced results consistent with other methods. In the following analysis, we used the resulting photometry from this method for the grz filters.
We accounted for Galactic extinction by adjusting our photometry using the values reported in Schlafly & Finkbeiner (2011) as implemented by the NASA/IPAC Extragalactic Database's extinction calculator. 2 Finally, zeropoint uncertainties and the statistical and systematic uncertainties estimated as above were combined in quadrature to compute total uncertainties for each measurement of each physical object. Table 1 shows total magnitudes for the sum of the three lensed images visible in Figure 1.

STELLAR POPULATIONS IN CJ1323
We derived the properties of stellar populations in CJ1323 by jointly fitting the spectra and photometry in the image plane with Prospector, an MCMC-based stellar population synthesis and parameter inference framework (Conroy & Gunn 2010;Foreman-Mackey et al. 2013;Leja et al. 2017). We assumed a nonparametric star formation history (SFH). We used seven age bins: 0 − 50, 50 − 100, 100 − 500, 500 − 1000, 1000 − 3000, 3000 − 5000, and 5000 − 5800 Myr in lookback time, with 5800 Myr being the age of the Universe at z = 1.015. The age bins were represented by the parameters log(SFR ratios n), referring to the ratio of total star formation in adjacent time bins. These ratios were fit with a continuity prior (see Leja et al. 2019 for details). We ran our model with 1024 walkers, 1024 iterations, and a burn in = [8192,4096,2048,1024,512].
We show corner plots constructed via pyGTC (Bocquet & Carter 2016) for two different attenuation curves in Figure  3 Kriek & Conroy (2013) dust attenuation curve (blue) and other fits from the posterior distribution (gray) calculated via Prospector using grzH photometry (orange) and optical/NIR Nordic Optical Telescope/ALFOSC spectroscopy (green), with the spectra's 1σ error shown in light green. Best-fit photometry is shown as black squares. (Bottom) A zoom in on the black box in the middle plot to show the spectra in more detail, with H-delta emission line, Calcium H and K absorption lines, and the center of the G-band in the rest-frame labeled. All three plots are in units of maggies. is red. We show the results of both models to give an idea of the systemic uncertainty introduced by how we model the dust attenuation. For both models, in addition to the star formation ratios, we simultaneously fit the following free parameters: total mass formed in the galaxy (M tot , in units of M ), stellar metallicity log(Z/Z ) (where Z = 0.0142), velocity smoothing in units of kms −1 , and spectrum normalization (the ratio of flux between the spectrum and photometry) to account for uncertainties in flux and spectral response calibration. For each chain in the MCMC analysis, we calculated the remnant stellar mass M * ,remnant . This accounted for mass loss from stars that have moved off the main-sequence, as well as stellar mass locked in stellar remnants, and is used as the mass throughout this paper. For the model using the C+00 attenuation curve, dust2 was a free parameter, which sets the overall normalization regardless of the age of stars, in units of opacity at 5500Å. For the model using the K&C13 attenuation curve, dust1 and dust2 were free parameters, which describe the attenuation of stellar light younger and older than 10 7 years, respectively. Finally, for the K&C13 model, dust_index (dust slope) was a free parameter, which corresponds to the strength of the 2175 Å UV bump. The stellar mass and metallicity distributions of both models are similar. However, the C+00 model's SFH suggests a slightly slower quenching than the K&C13 model (see Figure 3). Furthermore, for the C+00 model, star formation declines up to the epoch of observation, while the K&C13 model shows a slight rejuvenation in the most recent bin.
Kriek & Conroy 13 Calzetti+00 Figure 3. Corner plot with posterior distributions and correlations for inferred parameters in the Prospector SED fitting analysis for the lensed source in the image plane, with contours corresponding to 1σ (dark) and 2σ (light). We show results from the models using the Kriek & Conroy (2013) attenuation curve (blue) and the Calzetti et al. (2000) attenuation curve (red). In the model using the C+00 curve, dust_ index is fixed to 0 and dust2 is the only normalization factor. The best fit values for the model using the K&C13 curve are shown with gray dashed lines. This plot clearly shows the results for mass and metallicity are similar regardless of model choice.
We show the best fit SED model using the K&C13 dust attenuation curve in Figure 2 and use the results of this model throughout the paper. The stellar mass and SFH of the arc are found to be robust when fixing stellar metallicity at the best fit value, log(Z/Z ) = -0.19. The median value for the remnant stellar mass (M * ) in the image plane (i.e., the current value) is log(M/M ) = 12.48 +0.12 −0.08 and the best fit SFR in the image plane is SFR = 1.98 M yr −1 , averaged in the two youngest age bins (0-100 Myr). The median SFR = 0.11 +2.28 −0.11 M yr −1 . (The mismatch between the median and best fit is due the SFR having a strong negative skew, as shown in the top right panel of Figure 4). We find that the best fit dust index is negative as seen in other quiescent and low mass galaxies (e.g., Salim et al. 2018, indicating that the attenuation curve is steeper than the Calzetti et al. (2000) curve, where dust_index = 0.

LENS MODELING AND SOURCE PLANE RECONSTRUCTION
We modeled the mass distribution using LENSTOOL, a parametric lens modeling tool (Jullo et al. 2007). All halos were treated as pseudo isothermal ellipsoids (dPIEs, Elías- The specific star formation rate for each age bin. Both models show that CJ1323 was quiescent at the epoch of observation and agree on the rate of star formation in early age bins. The model using the K&C13 curve (blue) quenches faster and has a slight renewal in the most recent bin, while the model using the C+00 curve (red) quenches more gradually, with consistently lower sSFR in each age bin. dóttir et al. 2007) with parameters and uncertainties estimated using MCMC sampling. The overall modeling process closely followed the process described in Sharon et al. (2020). Astrometric constraints on the lens model were derived from the GALFIT model of the H-band image (the best seeing image available) without convolving the model with the PSF reference. This was essentially a parametric component deconvolution of the image. Astrometric reference points were placed on matching isophotes, which allowed us to use not just the centroid of each of the visible lensed images but also several locations along each image as constraints on the lens model. The appropriate configuration for these isophote-matched locations was guided by an initial simple model constructed using a single cluster-scale dPIE and smaller halos on each red-sequence cluster galaxy with positions, ellipticity, and position angles tied to the observed stellar light and the other parameters determined via scaling relations. Using the image labeling shown on the right panel of Figure 6, this initial model strongly suggests two key characteristics for this lensing configuration, namely that : 1) the primary magnification axis of image 3 is nearly orthogonal to images 1 and 2, and 2) that a fourth image should be present near the bright (and confounding) foreground star in the center of the field. While attempts to recover this fourth image from the initial DECaLS imaging were unsuccessful, additional optical imaging taken as described in Section 3.1 does clearly show the expected fourth image, once the stellar im- age was fit and removed using GALFIT (see the right panel of Figure 5). The final lens model used four individually-described dPIE components-two fixed on the two visible bright central galaxies, and two cluster-scale halos allowed to vary about those positions. Other cluster-galaxy mass contributions were added via a scaling relation, as described above. The left panel of Figure 5 shows the caustic and critical curves for this final model, and the input constraints derived from the H-band image that informs the model.
With a final lens model in hand, we reconstructed CJ1323 in the source plane with PyLenstool, a Python-based wrapper for LENSTOOL. 3 The reconstructions were made with the un-convolved H-band GALFIT model as described above. We added together the source plane reconstructions, centered on the brightest point as determined by a single Sérsic component GALFIT model. We show the combined image in the left panel of Figure 6. We then fit a Sérsic and sky component to the coadded image using GALFIT and report the effective radius, Sérsic index, and axis ratio in Table  2. To understand the variation in the source plane morphology of CJ1323 introduced by uncertainties in the model, we created combined images using the first ∼300 realizations of the lens model produced using the bayesCleanlens method in LENSTOOL. We report the median and uncertainty from the GALFIT models of these combined images in Table 2. We find that all source plane reconstructions had n < 3 and r e < 1.7, showing that our modeling is robust against degeneracies between n and r e . We projected a contour of the image plane half light area into the source plane using the lens model. The resulting area in the source plane divided by the original image plane area is the magnification. The best fit magnifications are µ 1 = 55, µ 2 = 31, and µ 3 = 28 for image 1, 2, and 3 respectively. The best fit total magnification is µ = 113 and the median is µ = 74 +49 −28 . The magnification uncertainty was found using the same method on ∼ 300 realizations of the lens model.  CJ1323 is a compact, intermediate mass quiescent galaxy at z ∼1-an object difficult to observe without gravitational lensing. Its stellar mass is 0.8 +0.5 −0.3 times the characteristic mass of the stellar mass function reported in Muzzin et al. (2013) for quiescent galaxies at 1 ≤ z < 1.5. Compared with van der Wel et al. (2014)'s expected size evolution for an ETG with M * = 4.27 × 10 10 , CJ1323 is small at 0.3 times the expected radius at its redshift, σ = 1 below the relation. Furthermore, the stellar density of CJ1323 is ρ = 5 +9 −3 × 10 10 M kpc −3 , comparable to compact ETGs at z = 2.3 (e.g., Kriek et al. 2008). This size is consistent with it being an unmodified relic ETG, that is, an early-type galaxy that quenched before z = 2 and has not grown since (e.g., Stockton et al. 2014, Hsu et al. 2014, Trujillo et al. 2014, Ferré-Mateu et al. 2017. The Sérsic index of the combined source plane reconstructions is 2.3, indicating CJ1323 is more disky than a standard n = 4 de Vaucouleurs profile. "Inside-out" growth leads disky ETGs at high redshifts to become elliptical in the local Universe as their outer envelopes build-up (van Dokkum et al. 2010). Most samples of intermediate redshift ETGs found that a significant portion of their objects are disky (e.g., Stockton et al. 2010, Hsu et al. 2014. However, all the source galaxies in Oldham et al. (2017)'s survey of intermediate redshift, 0.4 < z ≤ 0.7, earlytype/early-type lens systems (EELs) had at least one component with n > 4. That is, they have a bulge and are not disky. As Oldham et al. (2017) noted, this could indicate that their targets are more evolved counterparts of objects like CJ1323 and the ETGs in previous studies. We caution that source plane reconstructions based on ground based data are quite uncertain, and the preliminary results on morphology  Oldham et al. (2017). It is clear that, on average, lower redshift galaxies have larger radii, despite there not being as significant a trend for mass. Furthermore, CJ1323 is one of the most compact galaxies.
presented here should not be over-interpreted. Nevertheless, the suggestion from the apparent morphology and size is that CJ1323 is at an early stage of its morphological evolution post-quenching.
The SFH makes it clear that CJ1323 has been quiescent for at least 1 Gyr before the epoch of observation. Man et al. (2021)'s deep spectroscopic study found all three lensed quiescent galaxies had short bursts of star formation and then quenched quickly, with all three lying on the τ = 0.1 Gyr quenching curve from Belli et al. (2019). Artificial U-V and V-J colors (1.67 and 0.88, respectively) from the best fit SED model for CJ1323 also lie on this curve. Deeper spectroscopy will reveal more about the SFH of CJ1323, including quenching time scale and age. While better resolution data is needed to fit a more granular SFH, CJ1323's SFH, Sérsic index, radius, and stellar density suggest that it has not significantly changed since cosmic noon. If CJ1323 is an unmodified relic, it will give us the opportunity to study the initial conditions of ETGs in exquisite detail. If not, future studies could uncover evidence for the post-quenching evolutionary mechanisms that fuel the growth in radii of ETGs.
The results from this preliminary data make it clear that CJ1323 is a compelling target. Its total H-band magnitude of 17 makes it one of the two brightest lensed ETGs known, comparable only to MRG-M0138 (Newman et al. 2018, Jafariyazani et al. 2020. CJ1323 can help probe the processes that cause ETG morphology to change so significantly. Resolved spectra-that are in principle possible because of CJ1323's extreme magnification-would clarify how it quenched by providing a more detailed and better resolved SFH, both spatially and temporally. With sharper imaging, a more constrained lens model could be developed, yielding a more accurate and precise source-plane morphology of COOL J1323+0343, aided by near orthogonal magnifications provided by the different lensed images.
These observations are now being pursued for CJ1323 and several other lensed ETGs discovered by the COOL-LAMPS collaboration.