Mimicking Lambda with a spin-two ghost condensate

TL;DR

The paper addresses the cosmological constant problem by proposing a braneworld gravity theory in which late-time acceleration emerges without a bulk or brane cosmological constant. It introduces a spin-two ghost condensate in the five-dimensional bulk, with a Gauss–Bonnet–type boundary term that stabilizes linear perturbations and yields a massless graviton zero mode, avoiding the ghost instabilities seen in DGP self-accelerating branches. The authors construct a concrete accelerating background and analyze tensor and brane-bending perturbations, showing that the brane-bending mode decouples and gravity remains effectively four-dimensional at subhorizon scales for large enough $\gamma$. They discuss cosmological viability, fine-tunings, and implications for a low five-dimensional Planck scale, along with potential extensions and observational constraints. The work provides a concrete, ghost-free alternative to DGP-like self-acceleration with testable predictions for subhorizon gravity and cosmology.

Abstract

We propose a simple higher-derivative braneworld gravity model which contains a stable accelerating branch, in the absence of cosmological constant or potential, that can be used to describe the late time cosmic acceleration. This model has similar qualitative features to that of Dvali-Gabadadze-Porrati, such as the recovery of four-dimensional gravity at subhorizon scales, but unlike that case, the graviton zero mode is massless and there are no linearized instabilities. The acceleration rather is driven by bulk gravity in the form of a spin-two ghost condensate. We show that this model can be consistent with cosmological bounds and tests of gravity.