In a new paper, Damien Gagnier looks at what happens during late stages of common envelope after the companion plunges in and the two cores orbit in a shared envelope. In the simulations, we mimick the plunge-in by depositing angular momentum in the envelope in a controlled way and study transport of angular momentum by advection, turbulence, gravitational torques, and viscosity. We find that the binary orbit decays on a timescale of 10^3 to 10^5 orbits of the central binary suggesting that the orbit will continue shrinking for a long time. This might help resolving the discrepancy between simulations and observed final binary separations. Although too slow to directly study with multi-d hydro, the evolution is often much faster than the envelope thermal timescale – perhaps there is no need for any thermal “self-regulation” of the late inspiral. This also lends support to the usual energy-conserving alpha formalism. Interestingly, much of the hydrodynamics is very similar to circumbinary disks. For example, the “lump” is also seen, but it is more of a 3D phenomenon rather than a well-localized object. 3D simulations were done with Athena++ in spherical coordinates with AMR. By excising a sphere around the binary, the simulations can be run for 100s of orbits of the central binary. The cost was ~4.6 million CPU hours at IT4Innovations. The paper is at https://arxiv.org/abs/2302.00691