Is cosmic speed-up due to new gravitational physics (original) (raw)

Modified Gravity as Dark Energy

"We study the effects of introducing modiﬁcations to general relativity (“GR”) at large scales as an alternative to exotic forms of matter required to replicate the observed cosmic acceleration. We survey the effects on cosmology and solar-system tests of Dvali–Gabadadze–Porrati (“DGP”) gravity, f(R) gravity and Modiﬁed-Source Gravity (“MSG”). We ﬁnd that, in addition to the changes to the background expansion history of the universe, these modiﬁcations have substantial impact on structure formation and its observable predictions. For DGP, we develop a scaling approximation for the behaviour of perturbations off the brane, for which the predicted integrated Sachs-Wolf (“ISW”) effect is much stronger than observed, requiring new physics at around horizon scale to bring it into agreement with data. We develop a test based on cross-correlating galaxies and the ISW eﬀect which is independent of the initial power spectrum for perturbations and is a smoking-gun test for DGP gravity. For f(R) models, we ﬁnd that, for the expansion history to resemble that of LCDM, it is required that f,RR> 0. This also ensures that GR-like high-curvature solutions are stable. We then ﬁnd the conditions on f(R) which allow this subset of models to pass solar-system tests. Provided that gravity behave like GR in the galaxy, these constraints are weak. However, for a model to allow large deviations from GR in the cosmology, the galactic halo must differ signiﬁcantly from that predicted by structure evolution in GR. We then discuss the effect that these models have on structure formation, and ﬁnd that even in the most conservative of models, percent-level deviations in the matter power spectrum will exist and should be detectable in the future. Finally, for MSG, we investigate the cosmology of a theory of gravity with a modiﬁed constraint structure. The acceleration era can easily be replicated in these models; however, linear perturbations become unstable as the universe begins to accelerate. Once the perturbations become non-linear, the model reverts to GR, regaining stability. This should leave a signiﬁcant imprint on the structure-formation probes, but one which we cannot calculate."