Hi chaps, bit of a lurker to date, but this is relevant to my interests, so to speak! (I've done degrees in Chemistry and Material Science)
Forge scale is iron oxides - with an emphasis on the plural. Whilst there might be subtle differences due to the exact forging environment, it's typically a layer of haematite, magnetite and wustite, with the hematite as the layer between the iron and scale. The rough thickness of each are in a 1:3:100 ratio, with the haematite as the thinest.
In the context of hot rolled steel, the rolling action of the forge ensures that the scale is cracked.
Based on the photo's of the wiping off of the scale, I suspect that the picking solution is preferentially attacking the interfacial layers, and thus the primary mechanism is not attack on the wustite. Either that, or it is the magnetite layer that is attacked, and the remaining haematite layer is so thin to be invisible - but I suspect not in this case. The preferential attack of the interfacial layers is dependant on the cracks in the scale to allow the solution to penetrate, at least that will have a strong effect on the time required.
(By the way, Richard, are you sure the two photos of the steel in the trunk haven't been switched? The second one (of the underside) has a remaining scale pattern that appears to be an exact match of the first one (of the upper side), a situation that seems to have a vanishingly small probability ... )
Chloride ions act as a catalyst for oxidation of iron - to rust or Fougerite (green rust) without oxygen present. Green rust formation is slow, so that strongly suggests that if oxidation of iron is relevant, then it's an oxygen dependant process. This idea would mesh with the observation that the underside cleaned slower than the upper side - and is the only air dependant process I can identify that might be occurring.
This effect of chloride ions is not going to deplete them, therefore we can reasonably assume that the decline in effectiveness is due to the vinegar becoming neutralised (or, possibly, some mechanism sequestering the chlorine somewhere - but chloride salts are usually highly soluble, in particular the iron ones are; so I discount that). Therefore the action of the vinegar is crucial - but even with the chloride catalyst, an acid environment will attack iron faster than water will, so that doesn't rule out this from being a primary effect.
Rob points out about gas production reducing contact with the solution. It is possible to get a small amount of hydrogen produced, but I can't see it being enough to have a major effect - but then, if it's forming in the micro cracks, then might just be enough. The concentrations involved seem too low for significant gas production to me. However, there's a way to tell, and cover that in a moment.
Let me distill that down to a summary, then I'll make some concrete suggestions:
Suggested mechanism: Vinegar attacks the oxide layers, probably preferentially the layers next to the iron, the chloride ions oxidise the iron underneath the scale - this is what loosens the scale. This process would require oxygen from the air.
For the vinegar, look for 'non-brewed condiment' - it's purer and cheaper (and less tasty!) than actual vinegar.
To establish the exact mechanism (which is, frankly, not really needed, but the scientist in me has to document this stuff!):
Place two pieces of scaled steel in a pickling bath, one raised up on blocks to just under the surface, and the other directly beneath it, at the bottom of the tank. This then gives 4 surfaces exposed to pickling, which we can label from the top as A, B, C and D. Then pickle for however long, and check the under side of the top piece (side B) against the top side of the lower piece (side C). If there is a gas dependant mechanism, then if it relates to gas escape, then side C will be better pickled (and clean easier) than side B; whereas if it related to gas reaching it from the air, then side B will be clean easier than side C. Note that the placement of the top piece above the bottom amplifies the effect of any gas production, so aught to give a clear result in that case. If both are the same, then that implies that there is both a gas production, and an air requirement - or neither. Comparison with side A and D can separate these - if A cleans easier than D, then there _is_ a gas requirement, if they are the same, and the same as B and C, then there is neither.
In the event that there is neither effect, then that would suggest that the reason that the underside was more difficult to clean was something to do with the steel itself, and not the pickling.
If there is a gas production effect, that implies that the optimal geometry of the work piece to be picked is to have the main surfaces vertical (rather than the more natural horizontal). Also, if there is a gas production effect, that implies that it should _not_ be used in a sealed container.
Equally, if there is a gas requirement (which is almost certainly oxygen), that also suggests a sealed container would be less efficent.
As I said, I'm not really expecting anyone to actually do that experiment, unless you're _really_ keen, but it's a habit to document this stuff these days!
I can go further into the structure of forge scale if anyone wants - I spend quite a while doing computational modelling of wustite a while back (although that was for the magnetic properties, rather than chemical) ... but I've probably already gone past the 'useful' level of detail. Still, I hope that's some useful stuff in there for people.
Stuart
