Search

We study linear and nonlinear structure formation in cosmologies where the accelerated expansion is driven by modifications to general relativity (GR). We focus on Galileon and Nonlocal gravity, which are two classes of models that have been attracting much attention. We derive the linearly perturbed model equations and solve them with suitably modified versions of Einstein-Boltzmann codes. We also derive the perturbed equations keeping the relevant nonlinear terms for small scale structure formation, which we solve using N- body codes and semi-analytical techniques that were developed for these models. Using CMB, SNIa and BAO data we find strong evidence for nonzero active neutrino masses (Σmν ≈ 0.6 eV) in all three main branches of covariant Galileon cosmologies, known as the Cubic, Quartic and Quintic models. However, in all branches, the lensing potential does not decay at late times on sub-horizon scales, which contradicts the measured positive sign of the ISW effect, thereby ruling out the Galileon model. The Nonlocal model we study should be able to fit the CMB with similar parameter values as ΛCDM. The N-body simulation results show that the covariant Galileon model admits realistic halo occupation distributions of luminous red galaxies, even for model parameters whose linear growth is noticieably enhanced (σ8 ≈ 1) relative to ΛCDM. In the Cubic Galileon model the screening mechanism is very efficient on scales smaller than 1Mpc, but in the Quartic and Quintic sectors, as well as in the Nonlocal model, we identify potential tensions with Solar System bounds. We illustrate that, despite the direct modifications to the lensing potential in the Cubic Galileon and Nonlocal models, cluster masses estimated from lensing remain the same as in GR. The lensing effects produced by cosmic voids found in the simulations of the Cubic Galileon are significanly boosted (≈ 100%) compared to GR, which strongly motivates using voids in tests of gravity. The combination of linear and nonlinear theory results presented here for Galileon and Nonlocal gravity is an example of what it could be done for any serious alternative models to ΛCDM, which will be tested by future experiments.