# How can an accretion disk form around the central star of a planetary nebula?

I was reading the 2012 paper A tidally destructed massive planet as the progenitor of the two light planets around the sdB star KIC 05807616 about the possible origin of two exoplanets that orbit a subdwarf B star. The researchers suggested that they could have been formed by the fragmentation of the metal core of a gas giant that was orbiting the star while it was still in the main sequence. Just at the moment when it became a red giant and absorbed the planet, which was fragmented by tidal forces when it reached the core.

Here is an excerpt from the summary:

We suggest the following alternative scenario for the formation of the two planets (section 2). A single massive planet of mass $$m_P \gtrsim 5M_J$$ went through the CE evolution inside the RGB envelope. It spiralled all the way to the center. The released gravitational energy is behind the removal of the stellar envelope, as is commonly the case with CE evolution. The massive planet reached the tidal destruction radius (eq. 1). The gaseous mass of the planet was lost and part of it formed a temporary accretion disk around the core, that now is the EHB star. The metallic core of the massive planet was tidally destructed into several Earth-like bodies immediately after the gaseous envelope was removed. Different bodies had different energy per unit mass. Some of them spiral-in and were further destructed by the core, while other survived at orbital separations of $$\gtrsim 1R_⊙$$ within the gaseous disk. The bodies interacted with the disk and among themselves and migrated, as planets around young stars do. Two of the bodies survived and reached an almost resonance. These are the observed Earth-like planets.

I infer that at the time the gaseous accretion disk was formed around the core, the planetary nebula was still in formation, since the high temperatures of the core must have ionized the surrounding gas, at least for a few thousand years.

I understand that the gas that is expelled from the interior of a dying star in its red giant phase does so at several kilometers per second. My question is: why the gaseous accretion disk that formed around the core was not immediately ejected into space as was the case with the gas from the nascent planetary nebula? Why was it gravitationally tied to the core while the surrounding gas was escaping from gravity?