"Heat-polish plastic" suggests 3D-printed plastic parts, which get small corrugations from the line-wise additive process, while injected plastic parts are as smooth as the mould.
Hobbyists polish printed parts with local heat and skilful hands, avoiding to soften the whole part. But sinking the part in hot liquid would be easier and more industrial, if indeed buoyancy provides support for the softened part. Very nice.
Though, this needs buoyancy to go on until the part has cooled down enough, which - if I grasp the goal properly - implies that the liquid can't solidify too early. It must stay liquid down to room temperature or at least down to moderate heat.
At least at Aldrich, I see prices like 500€/50g for ionic liquids. Even if the amounts and the supplier differ, I fear that would be expensive for such a use, where 50kg are little. Polyethylene glycol, preferably permethylated, would at least be cheap. It is flammable at heat, yes, similarly to frying oil. Not or little toxic, little polluting. Oily but washable in water. The long-term stability at +240°C isn't broadly documented.
To save heat in a series production, the region of the liquid where the plastic part softens should better stay hot. Stratification would be a solution, with hot liquid on top, colder one at the bottom where the part is carefully introduced. A second column of liquid can communicate by their bottom with the stratified one to introduce the parts - or have some lock instead.
I would be fantastic to sinter plastic parts after 3D-printing using a similar process. Presently, these parts stay more fragile than their injected equivalents as the successive lines of polymer don't aggregate perfectly, be it because the previous ones have cooled down or because of remaining voids.
Softening the polymer by heat and applying isostatic pressure should remedy both, while buoyancy from a liquid with similar density would avoid the most trivial deformations. This needs much pressure but probably less heat. For parts of moulded titanium it's done in silicone, hence far below the melting point, but at many 100 bar.
Raveling doesn't mention this and I didn't check the prior art, so I can't tell whether it's his intention, my idea, or common practice.
Depending on the chosen pressure, the liquid should better be stiff ("bulk modulus"), as for instance silicone oil increases its density by 15% at 3000 bar. Again polyethylene glycol is stiffer than most liquids, nearing solid polymers.
Air bubbles aren't desired when pressure is applied. Vacuum, possibly when the part is submerged, helps this.
For series production, the chamber seal rings are an interesting challenge. Standard parts exist up to 1500 bar at room temperature.
Marc Schaefer, aka Enthalpy