A New Perspective on DGP Gravity

TL;DR

This work reframes DGP gravity as five-dimensional Einstein gravity with brane-localized terms and shows that extending the bulk to Schwarzschild reveals severe pathologies, including pressure singularities and self-accelerating branch instabilities. Using a Schwarzschild bulk, the authors derive a Friedmann-like constraint for brane trajectories, uncovering how the bulk mass μ and branch choice (SA vs N) shape brane cosmology and stability. They demonstrate that SA branes can carry arbitrarily negative five-dimensional energy, and that positive bulk mass can induce brane-induced pressure singularities, signaling nonperturbative instabilities. A Euclidean analysis indicates unsuppressed nucleation of SA branes in empty 5D space, suggesting strong quantum mixing with the Minkowski vacuum. Collectively, the results expose fundamental inconsistencies in the DGP framework unless a UV completion or targeted modifications are found to regulate these nonlinear and nonperturbative pathologies, and they urge caution in using SA solutions for observational tests.

Abstract

We examine brane induced gravity on codimension-1 branes, a.k.a DGP gravity, as a theory of five-dimensional gravity containing a certain class four-dimensional branes. From this perspective, the model suffers from a number of pathologies which went unnoticed before. By generalizing the 5D geometry from Minkowski to Schwarzschild, we find that when the bulk mass is large enough, the brane hits a pressure singularity at finite radius. Further, on the self-accelerating branch, the five-dimensional energy is unbounded from below, implying that the self-accelerating backgrounds are unstable. Even in an empty Minkowski bulk, standard Euclidean techniques suggest that the spontaneous nucleation of self-accelerating branes is unsuppressed. If so, quantum effects will strongly modify any classical intuition about the theory. We also note that unless considered as Z_2-orbifold boundaries, self-accelerating branes correspond to `wormhole' configurations, which introduces the usual problematic issues associated with wormholes. Altogether these pathologies present a serious challenge that any proposed UV completion of the DGP model must overcome.