Infrared Modification of Gravity

TL;DR

Can General Relativity be consistently modified at very large distances? A DGP-like brane-world model with a 4D Einstein term on the brane and a 5D bulk term yields an IR modification with crossover at $r_c$ and extra graviton polarizations. The linear theory shows a strong-coupling scale $q_s = (M_P/ r_c^2)^{1/3}$, but non-linear resummation gives regular brane solutions with an Einsteinian limit; the cosmology is governed by $H^2 \pm H/r_c = (8π G_N/3) ρ$, admitting a self-accelerated phase with $H = 1/r_c$. The results imply testable signatures in precision cosmology and short-distance gravity, and UV completions appear to shield brane physics from high-energy details, preserving near-GR behavior.

Abstract

In this lecture I address the issue of possible large distance modification of gravity and its observational consequences. Although, for the illustrative purposes we focus on a particular simple generally-covariant example, our conclusions are rather general and apply to large class of theories in which, already at the Newtonian level, gravity changes the regime at a certain very large crossover distance $r_c$. In such theories the cosmological evolution gets dramatically modified at the crossover scale, usually exhibiting a "self-accelerated" expansion, which can be differentiated from more conventional "dark energy" scenarios by precision cosmology. However, unlike the latter scenarios, theories of modified-gravity are extremely constrained (and potentially testable) by the precision gravitational measurements at much shorter scales. Despite the presence of extra polarizations of graviton, the theory is compatible with observations, since the naive perturbative expansion in Newton's constant breaks down at a certain intermediate scale. This happens because the extra polarizations have couplings singular in $1/r_c$. However, the correctly resummed non-linear solutions are regular and exhibit continuous Einsteinian limit. Contrary to the naive expectation, explicit examples indicate that the resummed solutions remain valid after the ultraviolet completion of the theory, with the loop corrections taken into account.