Uncovering a population of gravitational lens galaxies with magnified standard candle SN Zwicky

Abstract

Detecting gravitationally lensed supernovae is among the biggest challenges in astronomy. It involves a combination of two very rare phenomena: catching the transient signal of a stellar explosion in a distant galaxy and observing it through a nearly perfectly aligned foreground galaxy that deflects light towards the observer. Here we describe how high-cadence optical observations with the Zwicky Transient Facility, with its unparalleled large field of view, led to the detection of a multiply imaged type Ia supernova, SN Zwicky, also known as SN 2022qmx. Magnified nearly 25-fold, the system was found thanks to the standard candle nature of type Ia supernovae. High-spatial-resolution imaging with the Keck telescope resolved four images of the supernova with very small angular separation, corresponding to an Einstein radius of only θ_E = 0.167″ and almost identical arrival times. The small θ_E and faintness of the lensing galaxy are very unusual, highlighting the importance of supernovae to fully characterize the properties of galaxy-scale gravitational lenses, including the impact of galaxy substructures.

Keywords: Cosmology; General relativity and gravity; Time-domain astronomy.

Fig. 1. Spectroscopic identification of SN Zwicky as an SN Ia and redshift measurements of the SN host galaxy and the intervening lensing galaxy.

The SN spectra obtained with the P60, Nordic Optical Telescope (NOT) and Keck telescopes (black lines) are best fitted by a normal SN Ia spectral template. The green line shows a comparison with the nearby type Ia SN 2012cg at a similar restframe phase, redshifted to z = 0.354. The SN flux peaked on 2022 August 17. The bottom panels show a zoomed-in view of a VLT/MUSE spectrum from 2022 September 30, displaying narrow absorption and emission lines, from which precise redshifts of the lens (z = 0.2262) and host (z = 0.3544) galaxies were determined. [O ii], Ca ii H and K, Na i D, Hα, [N ii] and [S ii] lines can be seen at the restframe of the lens (blue lines) and host (red lines) galaxies. ALFOSC, Alhambra Faint Object Spectrograph and Camera.

Fig. 2. Image of the field of SN Zwicky using pre-explosion g- and r-band images from ZTF.

Top left: a 2′ × 2′ section of the ZTF g- and r-band pre-SN images (FWHM 2.3″), centred on the location of SN Zwicky. Top right: a zoomed-in composite image of SN Zwicky using adaptive optics (AO)-enhanced VLT MUSE (g/r-band) and HAWK-I (J-band) observations on September 2 and 4 (FWHM ~0.4″). Bottom: a 2″ × 2″ portion of the Keck/NIRC2 LGSAO J-band image (FWHM 0.09″), resolving the four multiple images of SN Zwicky (labelled A, B, C, D). The blue dashed ellipse shows the critical line of the lens, corresponding to the inferred θ_E = 0.167″ (0.6 kpc at z = 0.2262), enclosing the lens M = (7.82 ± 0.06) × 10⁹ M_⊙. The host galaxy nucleus is located 1.4″ to the northeast of the lens, implying that SN Zwicky exploded at a projected distance of 7 kpc from the centre of its host galaxy.

Fig. 3. Multicolour lightcurve of SN Zwicky showing that the supernova is 3.5 mag brighter than an unlensed SN at the same redshift.

The magnitudes are measured with respect to time of maximum light (modified Julian Date 59808.67, corresponding to 2022 August 17) in ZTF g and r and Liverpool Telescope (LT) g, r, i, z filters. The solid lines show the SALT2 model with the best fit to the spatially unresolved data. The blue dashed lines indicate the expected lightcurves at z = 0.354 (without lensing), where the bands represent the s.d. of the brightness distribution for type Ia supernovae. To fit the observed lightcurves, a brightness increase corresponding to 3.5 mag is required. Also shown, as dotted lines, are the modelled individual lightcurves for the four SN images A–D. The flux ratios were measured from the NIRC2 data in Fig. 2. From these lightcurves we extract the time delays between the images, all in units of days, Δt_AB = −0.4 ± 2.9, Δt_AC = −0.1 ± 2.3 and Δt_AD = −0.1 ± 2.7, as described in Supplementary Information. The shaded areas in the lower panels indicate the regions with data outside the phase range where the SALT2 model is defined, and therefore excluded from the lightcurve fit analysis. The error bars correspond to 1 s.d.

Fig. 4. Stellar mass versus Einstein radius for lens galaxies discovered in galaxy surveys, demonstrating that SN Zwicky, iPTF16geu and the unresolved lensed supernova PS1-10afx point to a poorly known population of small-image-separation lensing systems.

Strongly lensed galaxy systems are represented by yellow diamonds for the BOSS Emission-Line Lens Survey (BELLS) sample, blue triangles for Sloan Lens ACS (SLACS) lenses and purple squares for SL2S lenses. The green circles correspond to lensed quasars, of which the filled circles have been detected optically and the open circles through radio emission. The shaded grey contours show the 90% and 68% confidence levels for the full sample of 155 lensed galaxies and 45 lensed quasars. The lensed supernova data are presented as median values ± 1 s.d. For the unresolved lensed supernova PS1-10afx, only an upper limit of the Einstein radius is available. The stellar mass derivation for SN Zwicky is detailed in Methods.