Cosmological constraints on DGP braneworld gravity with brane tension

We perform a Markov chain Monte Carlo analysis of the self-accelerating and normal branch of Dvali-Gabadadze-Porrati braneworld gravity. By adopting a parametrized post-Friedmann description of gravity, we utilize all of the cosmic microwave background data, including the largest scales, and its correlation with galaxies in addition to the geometrical constraints from supernovae distances and the Hubble constant. We find that on both branches brane tension or a cosmological constant is required at high significance with no evidence for the unique Dvali-Gabadadze-Porrati modifications. The crossover scale must therefore be substantially greater than the Hubble scale $H_0{r}_c>3$ and 3.5 at the 95% C.L. with and without uncertainties from spatial curvature. With spatial curvature, the limit from the normal branch is substantially assisted by the galaxy cross correlation which highlights its importance in constraining infrared modifications to gravity.

Figure 1

Best-fit CMB temperature anisotropy power spectrum for $\Lambda \mathrm{CDM}$ and sDGP (left panel). Examples of nDGP models (right panel) illustrate the degeneracy between ${\Omega }_{\Lambda }$ and ${\Omega }_K$ corresponding to models in Table .

Figure 2

Best-fit distance modulus for sDGP, as well as the overlapping predictions for nDGP-B and nDGP-F with respect to $\Lambda \mathrm{CDM}$.

Figure 3

Best-fit $\Lambda \mathrm{CDM}$ and sDGP galaxy-ISW cross correlations for the different galaxy samples, roughly ordered in increasing effective, bias-weighted, redshift. Note the distinct predictions for the previously degenerate nDGP-B and nDGP-F models.

Figure 4

Marginalized likelihood for ${\Omega }_{r_c}$ in the nonflat nDGP and sDGP models. The vertical lines indicate 68%, 95%, and 99% C.L.

Figure 5

Contours of 2D marginalized 68%, 95%, and 99% confidence boundaries using WMAP5, ACBAR, CBI, VSA, SNLS, and SHOES (left panel), including gISW (right panel) for nDGP.