All this tells me is, that the actual black&white developer you used after the prebath is an inferior product. Whatever this prebath achieves, should and can be achieved by a well formulated black&white developer.
I don't have TEA at home so I can't check, but is there a chance, that this "amine smell" could also come from an amine known to be in C-41 CD? Can you confirm, that this amine smell is also present in the alkaline concentrate part, i.e. the concentrate not containing the HAS and also not the CD-4?
Metol is not "regenerated" by sulfite, instead sulfite forms a sulfonate with oxidized Metol, and this Metol Sulfonate lets go of the silver halide grain that the oxidized Metol would have stuck to. If we assume, that less than 1g of Metol is used up during development of a roll of film, then 3 g Sulfite should be plenty. Little wonder, that FX 2 is very active with its rather high pH.
Maybe the developer could be formulated in some special way to bring out this speed increase.. but I've yet to figure out any method of doing that. Developers I've used with the pre-bath include Dektol 1+20, DD-X 1+3, D-76 1+1, and others. I'm not using some crazy weird developer for most of this, though the original test that I did the A-B test in was a special D-23 formulation (~1/4 strength, with less sulfite) designed for low contrast. The results are always night and day. With the prebath, I can typically get some shadow detail even at ISO 25, while without it, I'm lucky to get shadow detail at ISO 3.
TEA to me has minimal smell unless you're heating it and hold your nose over the beaker. DEA is where it gets fishy really quick. Also most homebrew and patent formulas for color developers I've seen include amines but often not triethanolamine. They tend to be DEA or MEA or an ethylamine derivative. All of these I'd assume to have a stronger odor.
Yea, I forgot about that aspect. Metol is just removed by sulfite rather than regenerated. 3g of sulfite is enough to remove it, but it's also a matter of time scale. I know from lith developer formulation that 2g of sulfite can still be fairly slow to actually remove oxidized developing agents, especially if the developer is rapid in action and produces a lot of oxidation productions. Also FX-2 is somewhat high pH due to using carbonate, but 2g of carbonate is an extremely weak buffer in relative terms.
Searching ebay or alibaba "polyethylene glycol" there are many results, quite cheap
My guess is that the right accelerator may contribute to need less the solvent. At the end any solvent action will decrease DMax potential.
The accelerator is very responsible for the infectious development, this is developing those non exposed crystals that are close to exposed crystals, the more the infectious development do that the lower DMin, so less solvent action required.
In my opinion we the home reversal practitioners do not know all the tricks that proprietary Pro labs may use, those densities stated by DR5 (for example) play in another division. I guess home reversal may be improved a lot.
Also I don't think PEG alone is the magic bullet, but perhaps it would be interesting to see what effect it has.
Unsure how related it is, but ethylene glycol is very often an MSDS listed ingredient in commercial lith developers (though I've never seen a homebrew formula containing it). I think it would be reasonable to assume they could contain polyethylenes as well if they are non-toxic enough to not need listed on the MSDS
I highly doubt that infectious development is a desirable property of any reversal developer, and if so it would require the developer to use a minimal amount of sulfite (often less than 2g per liter, but more can sometimes be used).