Bose-Einstein condensation and Silver Blaze property from the two-loop Φ-derivable approximation

We extend our previous investigation of the two-loop $\Phi$-derivable approximation to finite chemical potential $\mu$ and discuss Bose-Einstein condensation (BEC) in the case of a charged scalar field with $O(2)$ symmetry. We show that the approximation is renormalizable by means of counterterms which are independent of both the temperature and the chemical potential. We point out the presence of an additional skew contribution to the propagator as compared to the $\mu=0$ case, which comes with its own gap equation (except at Hartree level). We solve this equation together with the field equation, and the usual longitudinal and transversal gap equations to find that the transition is second order, in agreement with recent lattice results to which we compare. We also discuss a general criterion an approximation should obey for the so-called Silver Blaze property to hold, and we show that any $\Phi$-derivable approximation at finite temperature and density obeys this criterion if one chooses a UV regularization that does not cut off the Matsubara sums.