Phase Spaces for asymptotically de Sitter Cosmologies
William R. Kelly, Donald Marolf
TL;DR
The authors construct two phase spaces of asymptotically de Sitter gravity, Γ(dS_k=0) and Γ(dS_k=-1), in any d≥3, each with a finite, conserved symplectic structure and well-defined charges that generate their asymptotic diffeomorphisms. In general, the ASG matches the isometries of the corresponding dS patch, but in d=3, k=−1 they uncover a Virasoro structure and an extended two-Virasoro algebra with imaginary central charges, connected to dS/CFT expectations. The d=3, k=−1 sector hosts nontrivial wormhole spacetimes that yield a BTZ-like mass gap despite the absence of local degrees of freedom. The charges agree with Strominger–Balasubramanian-type constructions on appropriate frames, clarifying how dS/CFT charges act on a physical phase space and revealing rich global structures in de Sitter cosmologies.
Abstract
We construct two types of phase spaces for asymptotically de Sitter Einstein-Hilbert gravity in each spacetime dimension $d \ge 3$. One type contains solutions asymptotic to the expanding spatially-flat ($k=0$) cosmological patch of de Sitter space while the other is asymptotic to the expanding hyperbolic $(k=-1)$ patch. Each phase space has a non-trivial asymptotic symmetry group (ASG) which includes the isometry group of the corresponding de Sitter patch. For $d=3$ and $k=-1$ our ASG also contains additional generators and leads to a Virasoro algebra with vanishing central charge. Furthermore, we identify an interesting algebra (even larger than the ASG) containing two Virasoro algebras related by a reality condition and having imaginary central charges $\pm i \frac{3\ell}{2G}$. Our charges agree with those obtained previously using dS/CFT methods for the same asymptotic Killing fields showing that (at least some of) the dS/CFT charges act on a well-defined phase space. Along the way we show that, despite the lack of local degrees of freedom, the $d=3, k=-1$ phase space is non-trivial even in pure $Λ> 0$ Einstein-Hilbert gravity due to the existence of a family of `wormhole' solutions labeled by their angular momentum, a mass-like parameter $θ_0$, the topology of future infinity ($I^+$), and perhaps additional internal moduli. These solutions are $Λ> 0$ analogues of BTZ black holes and exhibit a corresponding mass gap relative to empty de Sitter.
