Late-time dynamics of brane gas cosmology
Antonio Campos
TL;DR
The paper analyzes the late-time dynamics of brane gas cosmologies by explicitly modeling the decay of winding brane modes into non-winding loops and coupling this to the dilaton-driven background. A key finding is that decay into static loops (equation-of-state parameter $eta=0$) prevents the three large spatial dimensions from growing, whereas allowing the brane gas to be non-static (nonzero effective velocity) with $0<eta eq 0$ enables expansion and can resolve the brane problem. A phase-space analysis identifies attractor lines $l=-eta f$ and shows that maintaining a small string coupling requires $eta o 1/3$ for efficient expansion, while loitering phases naturally arise and help restore causal contact. The work also derives scaling relations for the winding network and demonstrates that, for $eta>0$, the horizon scales with cosmic time, whereas $eta=0$ leads to extended loitering; incorporating a non-static brane gas broadens the parameter space in which a realistic, expanding universe emerges. Overall, the results highlight the crucial role of the loop equation of state and dilaton dynamics in determining whether brane gas cosmologies can reproduce our observed 3+1-dimensional expanding universe, and point to non-static branes as a robust mechanism to overcome the brane problem.
Abstract
Brane gas cosmology is a scenario inspired by string theory which proposes a simple resolution to the initial singularity problem and gives a dynamical explanation for the number of spatial dimensions of our universe. In this work we have studied analytically and numerically the late-time behaviour of these type of cosmologies taking a proper care of the annihilation of winding modes. This has help us to clarify and extend several aspects of their dynamics. We have found that the decay of winding states into non-winding states behaving like a gas of ordinary non-relativistic particles precludes the existence of a late expansion phase of the universe and obstructs the growth of three large spatial dimensions as we observe today. We propose a generic solution to this problem by considering the dynamics of a gas of non-static branes. We have also obtained a simple criterion on the initial conditions to ensure the small string coupling approximation along the whole dynamical evolution, and consequently, the consistency of an effective low-energy description. Finally, we have reexamined the general conditions for a loitering period in the evolution of the universe which could serve as a mechanism to resolve the {\sl brane problem} - a problem equivalent to the {\sl domain wall problem} in standard cosmology - and discussed the scaling properties of a self-interacting network of winding modes taking into account the effects of the dilaton dynamics.
