Nothing to see here: Failed supernovae are faint or rare

TL;DR

This study searches for failed supernovae using a decade of galaxy observations but finds no conclusive evidence, setting upper limits on their occurrence rate and predicting future detection prospects with upcoming observatories.

Contribution

The paper develops a specialized pipeline to detect faint transients, applies it to extensive data, and constrains the rate of failed supernovae, providing valuable limits and future detection estimates.

Findings

No confirmed failed supernovae detected in 10 years of data.

Established upper limits on failed supernova rates based on non-detections.

Predicted 1.7 to 3.7 failed supernovae per year detectable by the Vera C. Rubin Observatory.

Abstract

The absence of Type IIP core-collapse supernovae arising from progenitors above 17 solar masses suggests the existence of another evolutionary path by which massive stars end their lives. The direct collapse of a stellar core to a black hole without the production of a bright, explosive transient is expected to produce a long-lived, dim, red transient known as a failed supernova. Despite the detection of a number of candidates for disappearing massive stars in recent years, conclusive observational evidence for failed supernovae remains elusive. A custom-built pipeline designed for the detection of faint transients is used to re-analyse 10 years of observations of 231 nearby galaxies from the PTF/ZTF surveys. This analysis recovers known supernovae, and yields a number of interesting transients. However, none of these are consistent with a failed supernova. Through Monte Carlo tests the recovery efficiency of our pipeline is quantified. By assuming failed supernovae occur as a Poissonian process with zero detections in the data set, 95 per cent upper limits to the rate of failed supernovae are calculated as a function of failed supernova absolute magnitude. We estimate failed supernovae to be less than 0.61, 0.33, 0.26, or 0.23 of the core-collapse SN rate for absolute magnitudes of $-11$, $-12$, $-13$, and $-14$ respectively. Finally, we show that if they exist, the Vera C. Rubin Observatory will find 1.7 - 3.7 failed SNe per year for an absolute bolometric luminosity of $\sim 6 \times 10^{39} \textrm{ erg s}^{-1}$ out to distances of 33 - 43 Mpc, depending on their assumed spectral energy distribution.