Axion minivoids and implications for direct detection

TL;DR

This paper investigates how axion minivoids, formed from early universe inhomogeneities, affect local dark matter density measurements, revealing a significant suppression but not complete absence of axions in direct detection experiments.

Contribution

The study uses simulations to quantify the reduction in local axion density due to minivoids, providing new insights into the expected signals in axion haloscope experiments.

Findings

Typical axion density in experiments is about 10% of galactic dark matter density.

Density measurements can vary by 20-30% over year-long timescales.

Minivoids are not completely devoid of axions, allowing for potential detection.

Abstract

In the scenario in which QCD axion dark matter is produced after inflation, the Universe is populated by large inhomogeneities on very small scales. Eventually, these fluctuations will collapse gravitationally to form dense axion miniclusters that trap up to $\sim$75% of the dark matter within asteroid-mass clumps. Axion miniclusters are physically tiny however, so haloscope experiments searching for axions directly on Earth are much more likely to be probing ``minivoids'' -- the space in between miniclusters. This scenario seems like it ought to spell doom for haloscopes, but while these minivoids might be underdense, they are not totally devoid of axions. Using Schr\"odinger-Poisson and N-body simulations to evolve from realistic initial field configurations, we quantify the extent to which the local ambient dark matter density is suppressed in the post-inflationary scenario. We find that a typical experimental measurement will sample an axion density that is only around 10% of the expected galactic dark matter density. Our results also have implications for experimental campaigns lasting longer than a few years, as well as broadband haloscopes that have sensitivity to transient signatures. We show that for a $\mathcal{O}$(year)-long integration times, the measured dark matter density should be expected to vary by 20--30%.