Crossing the phantom divide: theoretical implications and observational status

If the dark energy equation of state parameter w( z) crosses the phantom divide line w = - 1 ( or equivalently if the expression d(H-2(z))/dz-3 Omega H-0m(0)2 ( 1 + z)(2) changes sign) at recent redshifts, then there are two possible cosmological implications. Either the dark energy consists of multiple components with at least one non-canonical phantom component or general relativity needs to be extended to a more general theory on cosmological scales. The former possibility requires the existence of a phantom component which has been shown to suffer from serious theoretical problems and instabilities. Therefore, the latter possibility is the simplest realistic theoretical framework in which such a crossing can be realized. After providing a pedagogical description of various dark energy observational probes, we use a set of such probes ( including the Gold SnIa ( supernovae type Ia) sample, the first-year SNLS ( Supernova Legacy Survey) data-set, the three-year WMAP CMB ( Wilkinson Microwave Anisotropy Probe cosmic microwave background) shift parameter, the SDSS (Sloan Digital Sky Survey) baryon acoustic oscillations peak (BAO), the x-ray gas mass fraction in clusters and the linear growth rate of perturbations at z = 0.15 as obtained from the 2dF (Two-Degree Field) Galaxy Redshift Survey) to investigate the priors required for cosmological observations to favour crossing of the phantom divide. We find that a low Omega(0m) prior ( 0.2 < Omega(0m) < 0.25) leads, for most observational probes ( except of the SNLS data), to an increased probability ( mild trend) for phantom divide crossing. An interesting degeneracy of the integrated Sachs-Wolfe effect in the CMB perturbation spectrum is also pointed out.