Although I won't make the case that they're anywhere near the quality of the original plates, I'd say I'm about 95% of the way to being able to make proper autochromes. I can capture and reproduce colors, though for now my only major successes have been with creating color negatives and digitally inverting them.
I've found that the screens from bad plates can be reused, which saves a considerable amount of effort when experimenting with the process. The gelatin is allowed to soak in warm or hot water for some time, before being scraped off the screen. My screens are somewhat soft when warm, so I prefer to do this under cold running water. I haven't really found anything that works better than my fingernail, but I can't say I've put any considerable amount of thought into finding a better tool.
As I touched on in my last post, I theorize that my emulsion is currently too thick, leaving a blanket
of unexposed AgX in the "back" of the plate, so that reversal processing leaves the photograph more or less opaque. A thin chunk section from a botched pouring job seemed to indicate this was correct - though still dark, a transmitted color positive could be seen. This is further expounded upon in the Lavendrine/Gandolfo autochrome book, "The average of all of the analyses carried out on many photographs shows that the photographic image - positive or negative - uses no more than 25% of the silver bromide contained in the emulsion, and that the other 75% is normally eliminated in the fixing step. This means that if the same quantity of silver salts were to be used in the Autochrome's light sensitive layer as in other types of photographic plate, the reversal step would result in leaving 50% of the unexposed silver bromide in the emulsion, which would cause a tremendous degree of darkening in the image. It therefore is essential to closely control the thickness and regularity of the emulsion layer and to produce a very thin emulsion [approximately 4 microns]."
Initially tried a 2mL coating instead of 20mL, though it was difficult to spread evenly before the gelatin began to gel up. It showed slightly better color, though the emulsion was so dense that it had to be backlit by a slide projector to really show off its colors. The second method I tried I called the pour off method - coating the plate, then draining excess emulsion back into the container the way collodion is drained back into its bottle. After the plate stopped dropping, it was allowed to sit on a level surface to dry. This produced some of the best positives I have yet obtained, though all of them are extremely dark.
After a bit of research, I found a method that the holography guys like to use to create nice, even, thin emulsions - 'mold coating'. This involves smashing the emulsion with another piece of glass with a hydrophobic coating (Rain-X in my case) and removing the glass after the emulsion has set up to allow it to dry. I only tried it once, but I was not very successful - the emulsion already seemingly has a weaker affinity for the second varnish than the glass, and preferred to stick to the bare glass regardless of the Rain-X treatment.
I also briefly tried to spin coat the plate using a 78rpm record player, another suggested method from the holography community. I failed to get an even coating on the screen, though the parts that did make it on were fairly even. The gelatin seemed to cool off quickly before it was able to travel to the end of the plate, and set up around the edges in a thick ring. The middle of the plate is thin -- possibly too thin. I think a better method would be to pour off the plate and allow it to dry on the record player, ensuring a fairly even and thin coating.
A few other miscellaneous failures:
Instead of setting a pour off plate back level, I tried just leaving it stand against something for an ultra thin emulsion. Very muddy and low contrast (like the spin coated plate but worse).
This monstrosity, which ultimately caused the clean detachment of the emulsion from the screen:
After some helpful advice from Photo Engineer over at the PHOTRIO forums, I think my problem lies with the fact that there is not enough solvent action in my first developer. And after that was pointed out, it seemed to make sense! I'm only using about 1/3 of the amount of ammonia when compared to the original autochrome developer - this means the first developer does not build enough density when developing. This technically makes a nice, clean negative but it leaves us with a poor, muddy positive as a result. On further advice from PE, I will be trying the Dufaycolor first developer instead, which relies on potassium thiocyanate as a silver solvent instead of ammonia. The ingredients have been ordered. Let's see how that goes!