JWST Spectroscopy of SN H0pe: Classification and Time Delays of a Triply-imaged Type Ia Supernova at z = 1.78

TL;DR

This paper reports JWST spectroscopy of a triply-imaged Type Ia supernova at z=1.78, classifying it and measuring precise time delays between its images, which aids in gravitational lensing and cosmology studies.

Contribution

First spectroscopic classification and time delay measurement of a high-redshift triply-imaged SN Ia using JWST data, demonstrating the potential for cosmological applications.

Findings

SN H0pe is a Type Ia supernova at z=1.78.

Measured time delays between images with high precision.

Spectroscopic classification confirms SN type and phase.

Abstract

SN H0pe is a triply imaged supernova (SN) at redshift $z=1.78$ discovered using the James Webb Space Telescope (JWST). In order to classify the SN spectroscopically and measure the relative time delays of its three images (designated A, B, and C), we acquired NIRSpec follow-up spectroscopy spanning 0.6 to 5 microns. From the high signal-to-noise spectra of the two bright images B and C, we first classify the SN, whose spectra most closely match those of SN 1994D and SN 2013dy, as a Type Ia SN. We identify prominent blueshifted absorption features corresponding to Si II $\lambda6355$ and Ca II H $\lambda3970$ and K $\lambda3935$. We next measure the absolute phases of the three images from our spectra, which allows us to constrain their relative time delays. The absolute phases of the three images, determined by fitting the three spectra to Hsiao07 SN templates, are $6.5_{-1.8}^{+2.4}$d, $24.3_{-3.9}^{+3.9}$d, and $50.6_{-15.3}^{+16.1}$d for the brightest to faintest images. These correspond to relative time delays between Image A and Image B and between Image B and Image C of $-122.3_{-43.8}^{+43.7}$d and $49.3_{-14.7}^{+12.2}$d, respectively. The SALT3-NIR model yields phases and time delays consistent with these values. After unblinding, we additionally explored the effect of using Hsiao07 template spectra for simulations through eighty instead of sixty days past maximum, and found a small (11.5 and 1.0 days, respectively) yet statistically insignificant ($\sim$0.25$\sigma$ and $\sim$0.1$\sigma$) effect on the inferred image delays.