Dark energy might interact with cold dark matter in a direct, nongravitational way. However, the usual interacting dark energy models (with constant $w$) suffer from some catastrophic difficulties. For example, the $Q\propto {\rho }_{\mathrm{c}}$ model leads to an early-time large-scale instability, and the $Q\propto {\rho }_{\text{de}}$ model gives rise to the future unphysical result for cold dark matter density (in the case of a positive coupling). In order to overcome these fatal flaws, we propose in this paper an interacting dark energy model (with constant $w$) in which the interaction term is carefully designed to realize that $Q\propto {\rho }_{\text{de}}$ at the early times and $Q\propto {\rho }_{\mathrm{c}}$ in the future, simultaneously solving the early-time superhorizon instability and future unphysical ${\rho }_{\mathrm{c}}$ problems. The concrete form of the interaction term in this model is $Q=3\beta H\frac{{\rho }_{\mathrm{de}}{\rho }_{\mathrm{c}}}{{\rho }_{\mathrm{de}}+{\rho }_{\mathrm{c}}}$, where $\beta$ is the dimensionless coupling constant. We show that this model is actually equivalent to the decomposed new generalized Chaplygin gas (NGCG) model, with the relation $\beta =-\alpha w$. We calculate the cosmological perturbations in this model in a gauge-invariant way and show that the cosmological perturbations are stable during the whole expansion history provided that $\beta >0$. Furthermore, we use the Planck data in conjunction with other astrophysical data to place stringent constraints on this model (with eight parameters), and we find that indeed $\beta >0$ is supported by the joint constraint at more than $1\sigma$ level. The excellent theoretical features and the support from observations all indicate that the decomposed NGCG model deserves more attention and further investigation.

• Received 25 December 2013

DOI:https://doi.org/10.1103/PhysRevD.89.083009