electropolish procedure - lab scale

I use electropolishing as a surface preparation for laboratory X-ray measurement of residual stress. I've found plenty of suggestions and anecdotal reports of what to use in various situations, but I've not come across any decent theoretical models of the electropolishing process that would allow predictions about the optimum conditions. From my experience with ferritic steel, common acid electrolytes such as the sulphuric/phosphoric/water seem to work well for low carbon content steels, and give a rougher surface with higher carbon steels. I also found the electrolyte sometimes goes purple, and I don't know why. For rail steel (about 0.7wt% C), I've only managed proper surfaces with 8% perchloric acid in acetic acid. This is just speculation, but I think the reason is that when the Fe atoms are 'oxidised' during the electropolishing, this only means oxidation in the sense that they ionised and go into solution. C atoms, on the other hand, must literally react with oxygen and be burnt to C02 (or maybe H2CO3). So for higher carbon content steels, you need a chemical that liberates oxygen. If perchloric cannot be used, other possibilities might be chromium(VI) oxide, hydrogen peroxide or potassium permanganate electrolytes, although I've never heard of the latter being used in electropolishing.
Joe