Superfluidity in e-h bilayers in graphene and GaAs has been predicted many times but not observed. A key problem is how to treat the screening of the Coulomb interaction for pairing. Different mean-field theories give dramatically different conclusio
ns, and we test them against diffusion Monte-Carlo calculations. We get excellent agreement with the mean-field theory that uses screening in the superfluid state, but large discrepancies with the others. The theory predicts no superfluidity in existing devices and gives pointers for new devices to generate superfluidity.
