Stability analysis and future singularity of the $m^2 R \Box^{-2} R$ model of non-local gravity

TL;DR

This paper investigates the classical stability of a non-local gravity model with a specific $m^2 R ox^{-2} R$ term, finding that flat space-time is unstable at cosmological scales but the model remains viable due to damping effects in the late universe.

Contribution

It provides a detailed stability analysis of the Maggiore-Mancarella non-local gravity model, clarifying the role of ghost-like modes and demonstrating the model's phenomenological viability.

Findings

Flat space-time is unstable under scalar perturbations at cosmological scales.

Late-time solutions lead to a big rip singularity.

Hubble friction damps scalar ghost perturbations, ensuring stability of the FLRW solution.

Abstract

We analyse the classical stability of the model proposed by Maggiore and Mancarella, where gravity is modified by a term $\sim m^2 R \Box^{-2} R$ to produce the late-time acceleration of the expansion of the universe. Our study takes into account all excitations of the metric that can potentially drive an instability. There are some subtleties in identifying these modes, as a non-local field theory contains dynamical fields which yet do not correspond to degrees of freedom. Since some of them are ghost-like, we clarify the impact of such modes on the stability of the solutions of interest that are the flat space-time and cosmological solutions. We then find that flat space-time is unstable under scalar perturbations, but the instability manifests itself only at cosmological scales, i.e. out of the region of validity of this solution. It is therefore the stability of the FLRW solution which is relevant there, in which case the scalar perturbations are known to be well-behaved by numerical studies. By finding the analytic solution for the late-time behaviour of the scale factor, which leads to a big rip singularity, we argue that the linear perturbations are bounded in the future because of the domination of Hubble friction. In particular, this effect damps the scalar ghost perturbations which were responsible for destabilizing Minkowski space-time. Thus, the model remains phenomenologically viable.