Reaction barriers at metal surfaces computed using the random phase approximation: can we beat DFT in the generalized gradient approximation?