Autochromes...

[Photo Engineer]

Oh, but this is what I mean: "Mixtures labelled naphtha have been produced from natural gas condensates, petroleum distillates, and the distillation of coal tar and peat. In different industries and regions naphtha may also be crude oil or refined products such as kerosene. Mineral spirits, also historically known as "naphtha", are not the same chemical."

So, we are still at an impasse for our foreign friends.

Sorry.

 

[ThePhotoChemist]

Indeed. Xylene or some other, less ambiguous solvent might work well too. Unfortunately, I've got too many things to try and not enough screens, so just naming the manufacturer and product is the best I can do at the moment.

I'm wondering if you have any thoughts on long processing times I'm experiencing? My gelatin is probably 5 years old at this point, and I've ordered new stuff because I'm nearly out. Increasing the gelatin% of the emulsion didn't seem to help speed things up much. I'm not using chrome alum or any other finals in this particular batch, either. I'm applying 10mL of emulsion per 4x5, and have tried both 2.8% and 6% gelatin content, with similar results. Even after a 45 minute wash under running water I can still see there is still developer in the gelatin (the plate darkened as soon as I turned on the lights).

 

[Photo Engineer]

Permeability comes to mind. It could be that solvent and vehicle is migrating to the surface and inhibiting permeability. IDK at this time. I would need step by step information on each coating layer in great detail, and more information on your bleach.

In that regard, a bleach with low acidity can be low in activity. That is another source of your problem.

PE

 

[ThePhotoChemist]

Alright, I managed to bring the processing time down quite a bit. I ended up diluting what remained of my emulsion to 1L, bringing the gelatin% back down to 2.7%, with less than a quarter of what remained of the silver halide content. Processing time is much more "normal", with a FD/Wash/Bleach/Clear going from about 30-50 minutes each down to about 10 minutes each. The plate is still quite opaque after coating (I was worried that I may have diluted it TOO much).

Yellow staining was confirmed to be caused by insufficient washing after clearing. I give it a solid 15 minutes now before second development and have yet to encounter any yellow discoloration at all.

Attached is a plate I made last night, which is comparable to brightness, contrast and color rendering as the plate I made a few days ago.

So now I think the next problem to tackle is the very high contrast nature of my results. Pushing the first development too far yields a very weak result, especially the shadows (though despite the attached image, color rendition was quite nice). I'm still using the DufayColor FD, which is as follows:

Metol - 0.5g
Hydroquinone - 4g
Sodium Sulphite - 25g
Sodium Carbonate - 17.5g
KBr - 2.5g
Potassium Thiocyanate - 4.5g
H20 - 500mL

Second development I'm just using D-76, undiluted.

Per this post here by Rudeofus:

1. If slide contrast is too low, then you need to extend development in first developer. If contrast is too high, then first development needs to be decreased.
2. If contrast is ok but the highlights are too dark, then either exposure is too low, or first developer is too weak (too dilute, shows exhaustion)
3. If contrast is ok but the shadows are too bright, then either exposure is too high, or first developer is foggy.

This seems counter to my results so far. Longer exposure / shorter FD seems to yield a much higher contrast for me, not lower, though the plate coating was much thicker when I last tried.

If anyone has any suggestions, I'd love to hear them! I'm considering making an emulsion with reduced KI content, but I won't be able to get to that until next week I think...

 

[Nodda Duma]

Nice results!

Not sure how you mix your emulsion or what recipe you're using, but there's a handful of ways to reduce contrast during the make. Longer precipitation, lower mixing rpm's come to mind. Add silver nitrate to the salts under the surface instead of dripping into the salt solution from above. If you are sensitizing with Sulphur, switch to Steigmann's solution since gold salts sensitize with less contrast. Higher contrast = higher effective speed, so you'll want to ripen a few minutes longer to compensate. I wouldn't dork with the KI content because you'll screw up your delicate color balance (it acts as a spectral sensitizer in the blue).

 

[Photo Engineer]

Don't try to control contrast with the FD. It is too chancy as it can cause fog. D76 is a solvent developer. Have you tried something else? Dufay used the FD as the 2nd Dev IIRC.

PE

 

[ThePhotoChemist]

Nice results!

Not sure how you mix your emulsion or what recipe you're using, but there's a handful of ways to reduce contrast during the make. Longer precipitation, lower mixing rpm's come to mind. Add silver nitrate to the salts under the surface instead of dripping into the salt solution from above. If you are sensitizing with Sulphur, switch to Steigmann's solution since gold salts sensitize with less contrast. Higher contrast = higher effective speed, so you'll want to ripen a few minutes longer to compensate. I wouldn't dork with the KI content because you'll screw up your delicate color balance (it acts as a spectral sensitizer in the blue).

Okay, thanks for the info! I'll stick with the same amount of KI for now. I used Steigmann's once, a few years ago, but it caused a massive amount of fog. Back then I had more kitchen equipment than lab equipment, so I'm sure I put too much in, ran the emulsion too hot, didn't measure properly, or something else. How much do you recommend adding per mmol AgX? You're correct, up until now I've only been sensitizing with sulfur.

Don't try to control contrast with the FD. It is too chancy as it can cause fog. D76 is a solvent developer. Have you tried something else? Dufay used the FD as the 2nd Dev IIRC.

PE

Good to know about using DufayColor as the second dev too. I don't think I've tried that yet.

Ah, thanks. Maybe this is one huge misstep I've been taking, because I've been trying to control contrast with the FD for quite some time now. I thought the point of an FD was to create a ton of fog, to help build density in the emulsion? It seems like a lot of silver halide at the back of the plate doesn't get enough exposure, and the extra fog helps cut the final density down quite a bit after bleaching. Maybe I'm coming at this at completely the wrong angle.

For other FDs, I've used the formula found here. It uses sodium thiosulfate for the silver solvent.

"The first developer formula for black and white slide
1 Water 50° C 750 ml
2 Metol 2 gr
3 Sodium sulphite anhydrous 100 gr
4 Hydroquinone 5 gr
5 Sodium carbonate monohydrate 60 gr
6 Sodium thiosulfate pentahydrate (Hypo) 16 gr
7 Potassium bromide 4 gr
8 Water Up to 1000 ml
Developing time 10 minutes at 20 °C for concentrated "

Though I don't recall using it much, I don't remember why (this was a year and a half ago, give or take). I think I was still getting better results with the Dufaycolor FD at the time. It'd definitely be worth a second look. I also used that second formula, where they added sodium carbonate, sodium thiosulfate and KBr to D-76, and only tried that once or twice. I don't remember my results with that, either.

For the second developer, I think I've only used D-76. I've got some Rodinal sitting around, I see some references to that being used both as a first and second developer with autochromes. Do you have any other recommendations?

 

[Photo Engineer]

I looked at your reference and finally recognize that the FD is basically D76 with hypo added. Hah. Missed that. However, he is using Dektol as the 2nd developer which is what I almost put in my note above.

As for fog in reversal systems, that does work, but you have to be careful as it can run out of control. In E6 films, the fog is in both the film emulsions and the FD. Of course, it is different in this case.

PE

 

[ThePhotoChemist]

The good news:

I tried out the FD formula that used D-76 with the 3 additional ingredients that I referenced in my previous post. Though the plate was dark after processing, there wasn't a ton of fog like the Dufaycolor FD, and the contrast seemed much more acceptable. I was just winging it, so I think dialing in the exposure and development might yield some nice results. The developer was noticeably discolored after a single use, though.

The bad news:

I managed to turn on the lights 3 times yesterday (I had a migraine and was a bit absent minded), which completely trashed the whole liter of emulsion I had. Today I managed to leave the top of my box off where I stored the rest of the plates, and managed to totally fog them too (no excuse here, I was just being dumb). So my experimenting will need to take a brief break while I make some more. Hopefully I Thursday I can play around a bit more!

 

[Photo Engineer]

If the emulsion was unmelted, it can be salvaged by removing a small amount of the outer edges of the exposed emulsion. There is little light penetration into the emulsion. It is attenuated rapidly.

PE

 

[ThePhotoChemist]

Nope, unfortunately the emulsion was melted and stirring when I turned on the lights -- and it was a full 5 seconds before I could turn them off again. I coated and tried the plates out anyway - after two minutes in the first dev they were completely black, and completely clear after bleaching. Totally unusable. I have all my lighting on voice control, which is *usually* really convenient -- but not so much if you mis-speak!

 

[ThePhotoChemist]

Alright, I've shot about 20 plates over the last few days. I don't have anything to show for it, but I'm really hopeful for two reasons:

1. Most screen's second varnish layer seem to be holding up really well, much better than my original screens. One or two leaked early on, and about half have taken on some odd looking artifacts that don't affect the final image much, but most seem pretty hardy dealing with the heavy abuse I've been giving them. The polyurethane plates are starting to peel a bit at the edges after processing, but nothing has leaked into the dye yet. I have a few plates using previously mentioned Golden's Acrylic Varnish which perform really well.

2. I really like working with the D-76+hypo FD. Like the DufayColor FD, it seems like the sweet spot is really picky. One minute too little and the plate is far too dark, and a minute too much will yield a ghostly color image that ban barely be seen. At least I'm still getting color every time now.

Last night I diluted down the emulsion again, so now it has 1/8 the original AgX content, and this STILL seems like too much. The contrast looks really nice, but after second development the results are far too dark. Diluting from 1/4 to 1/8 seemed to give a speed increase of about a stop for the same processing. Increasing exposure seems to give diminishing returns as far as final brightness is concerned. Processing time for bleaching has gone WAY down, much closer to how I remember. A bleaching takes about 5 minutes now.

I'm noticing now that with the reduced AgX content, the developer seems to be becoming too active - a mere two minutes is my sweet spot, 4 minutes is completely unusable (with the 1/4 emulsion, 4 minutes was good, 6 was iffy and 8 was totally blown out). So next week when I dilute it again (1/16), I'll probably have to reduce the amount of hypo too. Every once in a while I see what it looks like with normal D-76, but the results are basically opaque after second dev.

I've got time for one more plate this afternoon before I'm gone for the weekend, so I'm going to go ahead and really nuke it (6 stops exposure increase) just to see what happens.

 

[ThePhotoChemist]

Okay, I'm much closer to having decent results! After bouncing a few ideas off of Frederic Mocellin (check out his website here for neat autochrome stuff), I found out I was massively overdeveloping my plates. The amount of KSCN I was adding was too low relative to the developing action. My new developer is:

250mL D-19
3g KSCN
750mL distilled water (to make 1L)

Developing time is around 5 minutes at 17C

I'm still playing with the thickness of the coating vs KSCN content in the developer. I know I can get a larger dynamic range without sacrificing shadow density, so there are a few things I need to tweak still.

From the few outdoor shots I took this weekend, it looks like blue is a little stronger than green and red, but it 's pretty darn close! I'll probably up the pinacyanol and erythrosine in my next run of emulsion.

I was metering for 3 seconds at f/5.6, so if I did my math right the speed is about ISO 0.05. Despite the slow speed, at least I can say it's still faster than Lippmann plates! The tulip rows plate yellowed a bit, I believe because I didn't use fresh sodium sulfite during clearing. It helps balance the blue out somewhat, so it's no big deal. The tulip rows was shot on a beta test for my 3d printed plate holder, which was officially its first test, so that's quite exciting!

 

[Dr Croubie]

Ooh wow, I'm very tempted to try this myself one day (or I was, until I read a few posts here saying how hard it was, maybe I'd better get home-made B+W emulsions working first).
At any rate, very impressive.

Just wondering, is the colour of the sky in the last one actually mottled like that? Or is that just scanner noise? Do they project any good?

 

[ThePhotoChemist]

Ooh wow, I'm very tempted to try this myself one day (or I was, until I read a few posts here saying how hard it was, maybe I'd better get home-made B+W emulsions working first).
At any rate, very impressive.

Just wondering, is the colour of the sky in the last one actually mottled like that? Or is that just scanner noise? Do they project any good?

Best of luck! Definitely start with B&W emulsions first... my emulsion recipe now isn't all that much different than the original one I learned on, just with a few spectral sensitizers tossed in. I will always post all my thoughts and findings publicly, so that others can make their own autochromes too!

No scanner noise, that's all due to the large starch grains. I don't (yet) filter out the large ones, so the results are quite grainy compared to original autochromes. In person you can see that effect all over the plate, but digitally it seems to show more readily in objects with less detail.

 

[iandvaag]

Very impressive! These are starting to look really fantastic. Keep up the great work!

 

[bvy]

These are amazing...

 

[Punker]

Stunning!

 

[Photo Engineer]

Very very nice.

 

[ThePhotoChemist]

Alright, I have to admit I'm officially driving myself crazy now. I've shot well over 100 plates over and over again of the same scene, and quite a few different ones outside just to make sure there's not something wrong with my light.

The problem: I can reliably achieve color, but I just can't get anywhere near the dynamic range that original autochrome plates had. Bright colors always involve blown out highlights, and oftentimes weak shadows.

Good colors with longer development often leads to horribly contrasty scenes. Reducing development and increasing exposure time brings down the contrast and sometimes can muddy the colors, but still my midtones are very weak and shadows/highlights tend to overpower the image. Normal exposure times and longer development result in a nice image contrast, but the color is hopelessly washed, if not non-existent.

So far I've tried exposures ranging from 1 - 60 minutes. Developers I've tried:

Rodinal (concentrations ranging from 1+10 to 1+100, dev times ranging from 2 - 25 minutes)
D-23 in varying concentrations and times. Tried it with KSCN from 1-5g/L
D-76 in varying concentrations and KSCN/Hypo content
D-19 in varying concentrations and KSCN content
Dufaycolor FD in varying concentrations and various KSCN content.
Various ammonia based MQ developers from old photographic journals, with varying concentration and development times.

I guess the only developer I haven't really hit these with yet is the good ol' ammonia-pyro formula that was originally recommended for autochromes.

Also in combination, I have been trying various emulsion AgX contents. Generally, I just create my main batch of emulsion, and dilute it down with a 2.8% gelatin solution, and apply 12mL (full draw of a 10mL syringe) to each 4x5 plate. I've tried anywhere between no dilution to 1:16. 1:4 and 1:8 seem the most correct regarding processing times from old autochrome articles, and most of my experimentation has been near those two concentrations.

I don't know, I'm honestly pulling my hair out at this point. I'm guessing I need to start tweaking my emulsion recipe. I'm still using the same one I referenced earlier, and still including the sulfur sensitization. I don't think it's a particularly high contrast emulsion though, but I could nix the sulfur sensitization, and maybe start to think about pulling back on the amount of KI I add too.

I've been hitting The History of Color Photography pretty hard, specifically the chapter on processing screen plates, and even more specifically the section on reversal processing. So at the very least, I'm learning a lot, even if I don't have much to show for it. A lot of scenarios in my experimentation, I've noticed, result in the developer not penetrating all the way to the bottom of the gelatin layer (where it receives the most exposure). In one particularly thick plate, there was additional AgX even after both first and second development, that really didn't reveal itself until I hit it with a reducer.

This explains a phenomenon I've been seeing, where the plate, when immersed in the second developer, appears as a positive akin to a tintype for a minute or so before fading black. The upper part of the layer, where the bleached image is, develops first, and the milky, untouched AgX on at the bottom shows through the bleached areas, so that they appear white.

These upper layers are where the light is scattered most. Typically a silver solvent (like ammonia, KSCN or hypo) eats away at these, while allowing the developer to penetrate into the emulsion and do its job.

I had one scenario where I got a really nice looking plate. I had hit a thicker plate with a 30 minute exposure and a long development in a very diluted ammonia based developer. After second development it was hopelessly dark, with the positive image barely visible even when held against a light bulb. I hit it with a superproportional reducer based on ammonium persulfate and sulfuric acid. To my surprise, though it lost a good amount of highlight detail, it retained very nice colors and dynamic range. This was a one-off experience that I haven't yet been able to reproduce. I'm assuming the long development time allowed the developer to penetrate deep into the emulsion and develop the bottom of the emulsion, and then the reducer ate away at the big blanket of silver at the top of the emulsion that was left there after second dev.

The good news is that I was able to make some progress on sorting the starch grains, which is a small baby step forward at least.

P.S. I just had a crazy thought - shorter exposures (1 minute or so) with development times on the longer side tend to look washed, but generally take on a blue tint. I'm wondering, is it possible that with a screen plate process, are the AgX behind blue grains being exposed first, yet still hang around blocking blue rays while the AgX behind the red/green grains "catch up"? Maybe this is why the longer the exposure is, the more balanced in color it seems. And with shorter exposures, there would be a larger concentration of exposed AgX behind blue grains. Maybe my emulsion isn't very well balanced for color after all?

 

[glbeas]

I wonder if a developer incorporated emulsion would get the deeper AgX to develop more proportionally.

 

[Nodda Duma]

Good colors with longer development often leads to horribly contrasty scenes.

So you need to use this combination of exposure and development times as your baseline starting point... i.e. get good colors then tune things from there. By "tune" I mean reduce the contrast of your emulsion. So look up the methods to reduce an emulsion recipe's contrast. Gold sensitization instead of sulfur sensitization, dilution of emulsion to reduce the silver halide concentration, tweak the amount of KI in an Bromide-Iodide emulsion, etc.

Reducing development and increasing exposure time brings down the contrast and sometimes can muddy the colors, but still my midtones are very weak and shadows/highlights tend to overpower the image.

i.e. overexposure

Normal exposure times and longer development result in a nice image contrast, but the color is hopelessly washed, if not non-existent.

overexposure plus the effects of halation (light scattering when it reaches the emulsion layer after passing through your screen) in your emulsion.

 

[ThePhotoChemist]

So you need to use this combination of exposure and development times as your baseline starting point... i.e. get good colors then tune things from there. By "tune" I mean reduce the contrast of your emulsion. So look up the methods to reduce an emulsion recipe's contrast. Gold sensitization instead of sulfur sensitization, dilution of emulsion to reduce the silver halide concentration, tweak the amount of KI in an Bromide-Iodide emulsion, etc.



i.e. overexposure



overexposure plus the effects of halation (light scattering when it reaches the emulsion layer after passing through your screen) in your emulsion.

It's not quite that simple, though. Those examples of overexposure are actually wayyyy on the lower end of my exposure times, and if I reduce the time any less I don't get any detail at all (say, about 30 seconds at EV9). Just to confirm, I shot a plate just now at 10 seconds, and it was almost entirely blank. They get more saturated the longer the exposure is (approaching 5 minutes). But now I need to reduce development times to ridiculously short (or keep diluting it) otherwise the image is all blown out. This helps with colors and contrast in general, but I'm still not getting a decent dynamic range.

Does gold sensitization explicitly lower contrast, or does it just not cause contrast increase like sulfur? I don't need my plates to be fast (yet), I just want them to be consistent and reliable. Diluting apparently doesn't have much of an effect here, or not enough of one to help -- I've diluted all the way down to the point where I can no longer get a solid black anymore.

 

[Nodda Duma]

Gold sensitization does not cause as much a contrast increase as sulfur.

I didn’t realize you hadn’t started sensitizing yet. That explains the long exposure times.

I’m not sure what emulsion recipe you’re using, but

a) 12 ml on a 4x5 plate is A LOT of emulsion, and you were probably forced into doing that because

b) 1:4 - 1:8 is VERY diluted.

I think trying to replicate what their recipe was exactly including processing times, exposures, etc is causing you problems. For one, they used active gelatin which we don’t have available. So you start with inactive gelatin and then you *have* to sulfur sensitize to get close to what their emulsion was. The materials we have now aren’t the same, so you have to use a formula made for modern ingredients that gets you the same *final* result with modern developers.

For a ~5-6% gelatin emulsion, you really only want about 5ml on the plate, undiluted or (if sulfur sensitizing) no more than 1:1 dilution. Dig into period literature and you’ll find this is what plate makers settled in on. Make a test emulsion using that guideline, Bromo-Iodide with close to 3% iodine (no ammonia digest), coat a 4x5 with 5ml, expose a step density wedge, and plot out the characteristic curve to calculate the gamma when you’ve developed it in a standard developer for reasonable times...like D-76 for 6-7 minutes. Then, tune your emulsion for a gamma to about a value of 1 (iirc...look up the gamma value for E100. Developing as a negative is the inverse of the slide film’s value). Add KI after wash to increase contrast, increase precipitation time and mix faster to decrease contrast

So then you have proper contrast, reasonable speed, *significantly reduced halation because you’re not diluting it like crazy*, and then you coat the emulsion, reverse process and tweak the contrast.

 

[ThePhotoChemist]

Gold sensitization does not cause as much a contrast increase as sulfur.

I didn’t realize you hadn’t started sensitizing yet. That explains the long exposure times.

I’m not sure what emulsion recipe you’re using, but

a) 12 ml on a 4x5 plate is A LOT of emulsion, and you were probably forced into doing that because

b) 1:4 - 1:8 is VERY diluted.

I think trying to replicate what their recipe was exactly including processing times, exposures, etc is causing you problems. For one, they used active gelatin which we don’t have available. So you start with inactive gelatin and then you *have* to sulfur sensitize to get close to what their emulsion was. The materials we have now aren’t the same, so you have to use a formula made for modern ingredients that gets you the same *final* result with modern developers.

For a ~5-6% gelatin emulsion, you really only want about 5ml on the plate, undiluted or (if sulfur sensitizing) no more than 1:1 dilution. Dig into period literature and you’ll find this is what plate makers settled in on. Make a test emulsion using that guideline, Bromo-Iodide with close to 3% iodine (no ammonia digest), coat a 4x5 with 5ml, expose a step density wedge, and plot out the characteristic curve to calculate the gamma when you’ve developed it in a standard developer for reasonable times...like D-76 for 6-7 minutes. Then, tune your emulsion for a gamma to about a value of 1 (iirc...look up the gamma value for E100. Developing as a negative is the inverse of the slide film’s value). Add KI after wash to increase contrast, increase precipitation time and mix faster to decrease contrast

So then you have proper contrast, reasonable speed, *significantly reduced halation because you’re not diluting it like crazy*, and then you coat the emulsion, reverse process and tweak the contrast.

Sorry that I've been vague, I'll re-post my whole process real quickly. You're right, I AM using sulfur sensitization, but I'm considering not doing it next time. I'm not replicating the original Lumiere recipe, it's sort of a weird cobbled together technique from a few different guides off of Denise Ross's website.

1. 7g KBr and 30mg of KI are added 60mL of water with 2g of gelatin. 12mL of 1:1000 erythrosine solution is added. Solution is heated to 50C
2. 8g of AgNO3 are dissolved in 60mL of water. With a syringe (smallest tip), this is added over the course of a few minutes, with the tip of the syringe under the surface.
3. Allowed to ripen for 15 minutes, then cooled.
4. Noodle wash
5. Heat to 60C. 5mL of a fresh 0.1% hypo solution is added. Allowed to ripen for 30 minutes. At this point there's usually about 250 - 300mL of emulsion.
6. Emulsion is diluted to 1000mL, keeping a gelatin% of 2.8. 1mL 1:1000 pinacyanol is added.

This is my "stock" emulsion (which I usually call 1:4). I do all my stirring magnetically, so that should all be pretty consistent. That's interesting what you say about stirring speed -- I've had it backwards mentally, so I've been stirring pretty slowly (~200RPM). I'll speed it way up and see if that helps.

I know from digging through tons of old journals that autochrome plates have coatings that are significantly thinner than other standard plates at the time. If I coat a plate with what I would consider "normal", the emulsion is WAAAY too thick. It causes crazy things like 45 minute bleaching/clearing/second dev each, and opaque plates after full processing, regardless of how long exposure and development is. Even now with my 1:4 dilution second dev takes a solid 10 minutes at least, and it's not for lacking exposure. I'm gonna try to reach out to Lavendrine again and see if I can get any information on how much emulsion was applied per area, and from there I should be able to figure out how much AgX to apply.

My low gelatin% and 12mL application is due to coating -- coating on the screens needs to be exactly flat, and even the smallest mistake can cause huge brightness gradients across the plate. Since I have to coat under extremely subdued light, I keep the plates flat on a leveling table and spread the gelatin with the tip of the syringe. The plates take a few minutes until the gelatin sets up, whereas higher gelatin% tends to set up during coating, which more or less ruins the coating.

That's interesting what you say about halation though -- I need to do some reading on that. I can try increasing the AgX content of the emulsion while still retaining a low gelatin% to see if I can apply more AgX per plate. Intuitively I would have thought higher AgX would cause more halation though.