Core-Collapse Supernova Rates in the Dark Energy Survey

Abstract

Over its five-year campaign (2013-2017), the Dark Energy Survey (DES) supernova (SN) survey recorded 31,636 light curves. Using a custom-built, two-stage random forest classifier, we identified 2,445 likely core-collapse SNe (CC SNe) within the survey rate analysis footprint. With this sample, we measure the volumetric CC SN rate from redshift z = 0 to 1.3 in 11 redshift bins, representing the most detailed single-survey coverage to date across nine billion years of lookback time. Over this interval, the CC SN rates are well fit by a power law R_CC(z) = R_0 (1 + z)^α, with R_0 = 1.58 \pm 0.17 \times 10^{-4} \text{ yr}^{-1} \text{ Mpc}^{-3} and α= 2.85 \pm 0.25. The rate turnover, expected at cosmic noon, is not yet apparent. Because CC SNe trace recent star formation, we compare our measurements with recent cosmic star-formation history (CSFH) models. Our measurements are ~70 per cent higher than CSFH models using a Salpeter initial mass function (IMF) to predict CC SN rates. Replacing the Salpeter IMF with a Baldry and Glazebrook IMF and a CC SN progenitor mass range of 8-40 M_\odot bridges the gap. Alternatively, CSFHs may miss a hitherto unidentified population of low-luminosity galaxies. We further find that the local ratio of stripped-envelope (SE) SNe to all CC SNe is 31 \pm 5 per cent, remaining constant from z = 0 to 1.3. Combined with an IMF analysis accounting for the "missing progenitor problem" of hydrogen-rich SNe, our results favour a mixture of single massive stars and binaries as SE SN progenitors, with the redshift-invariant ratio favouring binary channels.