Testing dark matter with high-redshift supernovae

Abstract

Dark matter in the Universe consisting of macroscopic objects such as primordial black holes may cause gravitational lensing of distant objects. The magnification associated with lensing will lead to additional scatter in the received flux from standard candles, and too small an observed scatter could rule out compact dark matter entirely. In this letter, we show how the scatter in fluxes of distant Type 1a supernovae could be used to distinguish between models with and without lensing by macroscopic dark matter. The proposed SNAP project, with $\sim 2400$ supernovae in the range $0.1\ls z\ls 1.7$, should be able to identify models at 99.9% confidence, if systematic errors are controlled. Note that this test is independent of any evolution of the mean supernova luminosity with redshift. The variances of the current Supernova Cosmology Project sample do not rule out compact lenses as dark matter: formally they favour such a population, but the significance is low, and removal of a single faint supernova from the sample reverses the conclusion.