Core radii and common-envelope evolution

Many classes of objects and events are thought to form in binary star systems after a phase in which a core and companion spiral to smaller separation inside a common envelope (CE). Such a phase can end with the merging of the two stars or with the ejection of the envelope to leave a surviving binary system. The outcome is usually predicted by calculating the separation to which the stars must spiral to eject the envelope, assuming that the ratio of the core-envelope binding energy to the change in orbital energy is equal to a constant efficiency factor a. If either object would overfill its Roche lobe at this end-of-CE separation, then the stars are assumed to merge. It is unclear what critical radius should be compared to the end-of-CE Roche lobe for stars which have developed cores before the start of a CE phase. After improving the core radius formulae in the widely used BSE rapid evolution code, we compare the properties of populations in which the critical radius is chosen to be the pre-CE core radius or the post-CE stripped remnant radius. Our improvements to the core radius formulae and the uncertainty in the critical radius significantly affect the rates of merging in CE phases of most types. We find the types of systems for which these changes are most important.