One trick to treat them all: SuperEasy linear response for any hot dark matter in $N$-body simulations

TL;DR

The paper extends the SuperEasy linear response method to any hot dark matter species, enabling accurate, parameter-free predictions of matter power spectra in cosmological simulations with diverse HDM components.

Contribution

It introduces a generalized, analytical, and fitting-free linear response approach for simulating any hot dark matter species in cosmology, improving accuracy and efficiency.

Findings

Achieves sub-percent accuracy in matter power spectrum predictions.

Validates the method against particle HDM simulations.

Shows non-standard HDM models have similar non-linear signatures to neutrinos.

Abstract

We generalise the SuperEasy linear response method, originally developed to describe massive neutrinos in cosmological $N$-body simulations, to any hot dark matter (HDM) species with arbitrary momentum distributions. The method uses analytical solutions of the HDM phase space perturbations in various limits and constructs from them a modification factor to the gravitational potential that tricks the cold particles into trajectories as if HDM particles were present in the simulation box. The modification factor is algebraic in the cosmological parameters and requires no fitting. Implementing the method in a Particle-Mesh simulation code and testing it on HDM cosmologies up to the equivalent effect of $\sum m_\nu = 0.315$ eV-mass neutrinos, we find that the generalised SuperEasy approach is able to predict the total matter and cold matter power spectra to $\lesssim 0.1\%$ relative to other linear response methods and to $\lesssim 0.25\%$ relative to particle HDM simulations. Applying the method to cosmologies with mixed neutrinos\Plus{}thermal QCD axions and neutrinos\Plus{}generic thermal bosons, we find that non-standard HDM cosmologies have no intrinsically different non-linear signature in the total matter power spectrum from standard neutrino cosmologies. However, because they predict different time dependencies even at the linear level and the differences are augmented by non-linear evolution, it remains a possibility that observations at multiple redshifts may help distinguish between them.