Direct carbon on glass process

[AndrewBurns]

Hey all, first off I'm making this thread here because this alt-process area of the forum seems the most active, although my process is not 100% analogue and so I don't know if this should be put into the hybrid area. Technically this could be done in a purely analogue way and the thread is more about the chemistry of the process rather than the method of exposure, so I figured it would fit.

As you may know I've been playing with a UV projector/enlarger for a while now and one aspect I've been keen to explore is making exposures through a sheet of glass, which is something you can't do with a contact-printing process without introducing a lot of blur from light diffusion through the thickness of the glass (unless you had a perfectly collimated light source).

I've tried this with PVA-SbQ so far and it works well, although the process of coating the glass with the PVA-SbQ emulsion doesn't work nearly so well yet which is why I'm looking at carbon. I'm calling this a direct carbon process rather than carbon transfer because there is no transfer, I pour the carbon glop directly onto the glass sheet, expose the image through the glass so that the gelatine hardens from the glass surface upwards, and then develop the glass plate in hot water, leaving only the layer of gelatine that hardened onto the glass surface.

I figured that this technique might allow me to side-step some of the issues with the tonal threshold of continuous-tone carbon printing, particularly when using DAS as the sensitiser like I am. My theory is that the very thin layers of hardened gelatine making up the highlights of a normal carbon transfer print, which are likely to break away and be lost during the image transfer or development, are more likely to stick around in this technique because they're hardened directly onto a rigid support and are never put under any mechanical stress. So far I think my results are promising, although I don't think I'll completely eliminate it (but hopefully move it high enough up the scale that it's not an issue).

Here's a test print of my best result to-date, backed by a piece of white paper. Ignore the few bubbles and specs of dirt that managed to land on the print during development... The tonal scale is a bit dark-heavy and needs an adjustment to be linear, impossible to know if this is related to my process or a characteristic of the LCD screen in my projector.

The dmax is not particularly great, and this is the issue I've been struggling with the most. I've tried increasing the pigment concentration, which seemed to make things worse. I tried decreasing the gelatine loading while keeping the same amount of ink to make a very thin high-contrast tissue and that also didn't seem to work. This most recent attempt was made with the following recipe:

(Wet percentage values)
10% Gelatine
2.5% Sugar
1.2% Pebeo indian ink
0.3% DAS
86% Water

Poured to a wet height of between 1.2 and 1.5mm thick (basically as thick as I could using surface tension, to go thicker would require some dams).

My UV source is about 380 nm, which should have reasonable penetration into the glop. I've tried a range of exposures and it seems that the dmax just plateaus at a certain point, like the gelatine only hardens to a tiny fraction of the total layer thickness. Before developing the dry glop layer is basically entirely opaque so if I managed to harden through most of the thickness my dmax should be very high. I'm suspicious that the problem is due to the low power of my UV projector (current exposure time for ~8x10" is 40 minutes), possibly there's a certain threshold for UV energy below which the DAS just can't harden the gelatine no-matter how long you expose it for, in which case I'm not sure what I could do. The low dmax is particularly bad if you try to back-light the glass, because then the light only travels through the pigment layer in one direction, as opposed to when you put the glass on a piece of paper and the light has to travel through the pigment twice (once to get to the paper and once when reflecting back off the paper).

Here's the same glass but backlit by an LED pad, as you can see the contrast is much lower.

My eventual goal for this process is to create 4 different glass plates of a single image, each with either the C, Y, M or K layer of a colour separation. I'll then stack these glass plates on top of each other and backlight with LEDs. I think that this could result in an interesting effect where the colour layers combine to create a correct-looking image when viewed straight-on, but shift relative to each other when viewed from different angles. I'm hopeful that this will allow me to deal with lackluster dmax, as the shadows of the final image will be built up of 4 different pigment layers rather than just one and so should have much higher optical density.

Any thoughts on ways that I could improve my shadow density? Every other aspect of the process is super nice and easy, without having to transfer the image between substrates there are a lot fewer steps and a lot less to go wrong.

 

[koraks]

Very interesting experiment and the results so far are quite promising, although it has to be said that even with DAS it's quite easy to get this tonal threshold behavior (with this dmax) even with a normal transfer process. Things get tricky if you want, as you're trying, to expand the tonal scale into a higher dmax.

It's difficult for me to get a feeling for how much UV power you are actually dumping into the tissue as I've only used contact-printing processes. What's your typical exposure time for a classic cyanotype with this same setup? Maybe that gives a basis for comparison.

Every other aspect of the process is super nice and easy

Yeah, there's a lot less fuss if there's no transfer. Although there are some more minor things to iron out and they might be quite challenging as well. For instance, I see that the tonal threshold on the top row starts one patch to the right of that on the lower rows. This suggests to me the tissue is thinner in that place; i.e. the glop wasn't poured to a constant thickness across the plate. Especially when doing color this will become a problem; in monochrome it's not a major issue unless you use that part of the plate for linearization. I can also see some 'waviness' in the mid-grey tones suggesting either problems with the pigment dispersion or perhaps simply artifacts from spreading the glop around during coating (or both).

For higher dmax, I wonder if 380nm will cut it. I personally use 365nm for highlights and 395nm for the black point; I've not tried any in-between wavelength. I'm not sure how feasible it is to swap in a longer-wavelength light source?

 

[AndrewBurns]

It's difficult for me to get a feeling for how much UV power you are actually dumping into the tissue as I've only used contact-printing processes. What's your typical exposure time for a classic cyanotype with this same setup? Maybe that gives a basis for comparison.

For classic cyanotype at this size I'd probably be doing a 30 minute exposure.

Yeah, there's a lot less fuss if there's no transfer. Although there are some more minor things to iron out and they might be quite challenging as well. For instance, I see that the tonal threshold on the top row starts one patch to the right of that on the lower rows. This suggests to me the tissue is thinner in that place; i.e. the glop wasn't poured to a constant thickness across the plate. Especially when doing color this will become a problem; in monochrome it's not a major issue unless you use that part of the plate for linearization. I can also see some 'waviness' in the mid-grey tones suggesting either problems with the pigment dispersion or perhaps simply artifacts from spreading the glop around during coating (or both).

I don't think that's the case. In this test print each row has a scale running from left to right, but each row is slightly different to the next, with the top row being on average slightly lighter than the bottom row. So you could imagine the tonal scale stating from top left and going down the column until the bottom and then continuing at the top of the next row, all the way until the patch at bottom right. I
would expect the first step of tone to appear on one of the rows before the others.

For higher dmax, I wonder if 380nm will cut it. I personally use 365nm for highlights and 395nm for the black point; I've not tried any in-between wavelength. I'm not sure how feasible it is to swap in a longer-wavelength light source?

I do have a 395-400nm LED module the same size and power as the 380-385nm one I'm currently using, but to swap it in and out easily I'd have to have another water cooling block machined up and to put some quick-connects in the cooling water lines (not that hard to do, and something I should probably do).

Ages ago when I was using my contact printing setup with a 405nm light source to perform 'normal' carbon transfers I was having relatively poor dmax as well even with very long exposure times (about 3 times as long as classic cyanotype with that setup) however at the time I assumed the very long exposure and poor dmax was a result of 405nm really being too long for DAS.

 

[koraks]

For classic cyanotype at this size I'd probably be doing a 30 minute exposure.

Okay, I would agree with your analysis that the main problem is likely just a lack of sheer UV power density at the exposure surface. For comparison: with my 365nm light source, classic cyanotype is something like 2-3 minutes and the highlight layer for DAS carbon takes about the same up to maybe one stop more. However, the shadows layer at 395nm easily involves 10 minute or longer exposures at a similar electrical power level (both 100W lensed COB LEDs).

each row is slightly different to the next

Thanks, that explains it well.
Still, keep an eye out for any slope-off in density towards the edges of the tissue. That one has bitten me several times before and it's frankly very difficult to resolve entirely, without resorting to just not using a large rim around the center of the tissue.

the very long exposure and poor dmax was a result of 405nm really being too long for DAS.

Perhaps, although I'd be surprised if the difference between 395nm and 405nm would be so dramatic. Then again, even at 395nm it takes a heck of a lot of power to get good shadows; see above.

 

[AndrewBurns]

Okay, I would agree with your analysis that the main problem is likely just a lack of sheer UV power density at the exposure surface. For comparison: with my 365nm light source, classic cyanotype is something like 2-3 minutes and the highlight layer for DAS carbon takes about the same up to maybe one stop more. However, the shadows layer at 395nm easily involves 10 minute or longer exposures at a similar electrical power level (both 100W lensed COB LEDs).

Do you think that there is some kind of non-linear behaviour in DAS hardening of gelatine? I've seen it stated a few times (by Calvin amongst others) that you 'need' a certain amount of power per area of gelatine for DAS carbon to work at all. In a linear system you'd expect exposure to add up regardless of how little energy you're using (like cyanotype for example seems to still work even when the UV light source is weak enough you have multiple-hour exposures). However if there was some kind of threshold energy to start the reaction I could see how hours of weak exposure may make zero difference, which would explain what I'm seeing. Kinda like reciprocity failure in film I suppose.

I could see how that might put a hard limit on density, because the energy penetrating into the gelatine layer will get weaker the deeper it penetrates, and if at some depth the energy drops under the threshold needed to harden the gelatine then the top layers would continue to expose with time (resulting in darker highlights and midtones) but the shadows would just peak at a certain density and never get any darker.

I think this matches what my testing has shown. I think my only option is to try to make the gelatine layer thinner with a higher pigment loading, but then more pigment is more opaque to UV and so I'm back to square one again.

This ultimately might mean for my purposes I'll need to go back to PVA-SbQ instead, which doesn't have any issues with exposure depth or UV energy but has proven significantly more difficult to get an even coating on glass. Gelatine is really nice in that it sets long before it dries, and when it sets it 'locks' in place the local thickness and pigment distribution. PVA on the other hand stays liquid and mobile as it dries, and I've seen that uneven drying (practically unavoidable on a big plate) results in emulsion and pigment being 'sucked' from the drying areas into the wet areas, leaving very uneven layers.

I wonder if a mixture of DAS gelatine and PVA-SbQ together could work, they're both water-based mixtures at least...

Thanks, that explains it well.
Still, keep an eye out for any slope-off in density towards the edges of the tissue. That one has bitten me several times before and it's frankly very difficult to resolve entirely, without resorting to just not using a large rim around the center of the tissue.

I'll keep an eye on the tissue thickness, I also know that my projector has a pretty defined vignette which I've partially compensated for in software but I could see how that will create extra issues when trying to print colour.

Perhaps, although I'd be surprised if the difference between 395nm and 405nm would be so dramatic. Then again, even at 395nm it takes a heck of a lot of power to get good shadows; see above.

Yeah your experience of needing 4 or 5 times longer for 395nm vs. 365nm sounds like what I'd expect.

 

[koraks]

Do you think that there is some kind of non-linear behaviour in DAS hardening of gelatine? I've seen it stated a few times (by Calvin amongst others) that you 'need' a certain amount of power per area of gelatine for DAS carbon to work at all.

Yeah, I've often wondered about the same, but I can't offer a firm conclusion on it. But my gut feeling is that this is indeed the case. Indeed, you'd expect exposure vs. density to be a fairly linear relationship, but things appear to be fuzzy at both the bottom and the top ends of the curve. At the bottom end, I suspect that the gelatin matrix just doesn't hold together well enough, causing total failure of the matrix to survive the warm water bath. At the top end, we have to contend with the self-masking behavior of the sensitizer; an effect that's particularly strong with DAS.

I think my only option is to try to make the gelatine layer thinner with a higher pigment loading

Increasing pigment load is the logical thing to do, yes. But at 12% ink load of your dry gelatin weight you're in a place where you should be able to get good density at least for a reflective print. For transmission density you'll certainly have to increase the pigment load - lots.

 

[fgorga]

Do you think that there is some kind of non-linear behaviour in DAS hardening of gelatine? I've seen it stated a few times (by Calvin amongst others) that you 'need' a certain amount of power per area of gelatine for DAS carbon to work at all. In a linear

This process is almost certainly non-linear.

The photo-chemistry involves a two photon mechanism which looks (sorta) like this:

Reactant + photon -> Intermediate + photon -> Useful Product

The intermediate is likely to decay to some un-useful side product if it does not absorb the second photon in a timely manner.

The chances of absorbing the second photon will depend on the flux of photons. Thus, with dim light the accumulation of the desired product is likely to be low. There will be some threshold (of incident light) above which the process will be linear with respect to the amount of exposure. Below that threshold, the amount of desired product formed with be quite small.

Hope this explanation is clear (and useful).

 

[AndrewBurns]

This process is almost certainly non-linear.

The photo-chemistry involves a two photon mechanism which looks (sorta) like this:

Reactant + photon -> Intermediate + photon -> Useful Product

The intermediate is likely to decay to some un-useful side product if it does not absorb the second photon in a timely manner.

The chances of absorbing the second photon will depend on the flux of photons. Thus, with dim light the accumulation of the desired product is likely to be low. There will be some threshold (of incident light) above which the process will be linear with respect to the amount of exposure. Below that threshold, the amount of desired product formed with be quite small.

Hope this explanation is clear (and useful).

Yep that makes a lot of sense and matches my results. Unfortunately it does limit my options when it comes to DAS carbon using this projector as I'm already pushing the bounds of what the system is capable of in terms of power output and this is the smallest size of print I'd be interested in making. The tyranny of physics remains undefeated.

Edit: I wonder if dichromates show this effect and what their relative speed is. I've been avoiding using them but they're actually relatively easy for me to buy and use in NZ so traditional carbon printing is definitely an option if it would allow me to print higher dmax and larger sizes.

 

[koraks]

Dichromate prints significantly faster in my experience and suffers much, much less from the tonal threshold issue. There's a linearity test by Sandy King somewhere on the groups.io that shows it's very linear, too - but it's been a while since I saw it and I don't recall the specific conditions under which he tested.

 

[fgorga]

Edit: I wonder if dichromates show this effect and what their relative speed is. I've been avoiding using them but they're actually relatively easy for me to buy and use in NZ so traditional carbon printing is definitely an option if it would allow me to print higher dmax and larger sizes.

I would think that dichromate would show a much more linear relationship between light intensity and exposure density than DAS.

The photo-reduction of dichromate [Cr(VI)] to an unstable Cr(V) intermediate is a one photo process. It is generally thought that the Cr(V) is further reduced by the organic molecules in the mix (i.e. the gum) to Cr(IV) [and possibly to Cr(III)] which results eventually to the hardening of the gum.

The details of chemistry of the harden process is complex and not at all clear, at least to me. (Most of the literature is about the reaction with simple organic molecules not with complex polymers such as gum.)

Thus, the secondary reactions of Cr(V) with the gum to harden it may not be strictly linear with Cr(V) concentration but, at least, the initial photochemical reaction is close to linear with respect to light intensity. This is clearly not the case with photochemistry of DAS.

Dichromate prints significantly faster in my experience and suffers much, much less from the tonal threshold issue. There's a linearity test by Sandy King somewhere on the groups.io that shows it's very linear, too - but it's been a while since I saw it and I don't recall the specific conditions under which he tested.

Good to know! These 'experimental' observations are at least consistent with my discourse based on the basic chemistry described above.

 

[Yezishu]

This looks like a great result to me! While I’m not familiar with the DAS-gelatin system, I have a few preliminary suggestions. I also suspect the UV projector power might be the bottleneck; usually, we troubleshoot this by running a high-power exposure test (e.g., reducing projection area or using a direct UV lamp) to map curing depth against power and time, as reciprocity doesn't always hold.

If projector power and wavelength are fixed, you might turn to a photoinitiator with higher sensitivity at 380nm or a higher dosage—similar to how we swap Irgacure 2959 (<350 nm) for LAP (365~405 nm) or use a blend of Irgacure 819(365~405 nm) and Irgacure 184(240~340 nm) in acrylate systems(ex, Poly(ethylene glycol) diacrylate) to balance surface and depth curing. I'm not sure if DAS (aryl azides) has a "menu" of derivatives specifically tailored for different wavelengths and solvents, though I suspect they might; additionally, I’ve seen benzophenones used as crosslinkers for proteins.

Regarding pigments, they often compete with DAS for UV absorption, so you might consider change to pigments with lower UV absorption, using the other three colors to compensate for black density, or consider the "Technicolor" method: printing colorless gelatin reliefs and dyeing them post-development.

 

[koraks]

a photoinitiator with higher sensitivity

That would be cool. But so far, decades of experimentation by carbon transfer practitioners have not yielded such a silver bullet, although many have tried.

As to the pigments: a CMYK pigment set with sufficient UV transmission is easily available; there's not much need here to reinvent the wheel. For K, carbon black in one of its many varieties works fine. For the others I've used with good success:
Cyan: PB15.3
Magenta: PR122
Yellow: e.g. PY154 or PY151
All of these are extensively used in e.g. inkjet technology and many other applications, so they are readily available and often at low to modest cost. Ready-made dispersions are also easy to obtain.

 

[Yezishu]

That would be cool. But so far, decades of experimentation by carbon transfer practitioners have not yielded such a silver bullet, although many have tried.

I believe their conclusion is that the potassium dichromate(350-420 nm) method......
DAS is safer, but also a much new thing. Maybe no one has optimized the combination of DAS process and LCD projection before. Not too long ago, LEDs were still expensive, and the default high-intensity light source was the high-pressure mercury lamp that could provide a large amount of short-wave ultraviolet light, so the current problem might not exist.

 

[koraks]

Maybe no one has optimized the combination of DAS process and LCD projection before.

I'm sure they haven't; I think on the planet there's maybe 1 or 2 people who have even tried this route - so far I know of only one!
I have done a decent amount of work on DAS carbon transfer, just not with LCD/DLP projection. It's a tricky process; in some ways dichromate is easier, although it has other drawbacks apart from the health, environment and safety aspects. It's not easy to be very consistent with dichromate as you can't reliably incorporate the sensitizer into the tissue and then store that for any appreciable amount of time due to the dark reaction.

 

[Yezishu]

I'm sure they haven't; I think on the planet there's maybe 1 or 2 people who have even tried this route - so far I know of only one!
I have done a decent amount of work on DAS carbon transfer, just not with LCD/DLP projection. It's a tricky process; in some ways dichromate is easier, although it has other drawbacks apart from the health, environment and safety aspects. It's not easy to be very consistent with dichromate as you can't reliably incorporate the sensitizer into the tissue and then store that for any appreciable amount of time due to the dark reaction.

Also, I have a solution that's completely uneconomical: using a special modified gelatin, gelatin-methacrylamide (GelMA), and a water-soluble photoinitiator known to work well at 405nm, such as LAP(Lithium phenyl-2,4,6-trimethylbenzoylphosphinate). I'm confident this formulation will work, as it's almost standard practice for biomaterial 3D printing and cell immobilization today. It should also be fine for immobilizing pigment particles.

The only problem is the price. GelMA isn't currently industrially produced, and the purchase price is around $100-$300 per gram, which is far from ideal for someone interested in this. Unless someone studies how to synthesize it by themselves, or it's not worth paying that much compared to potassium dichromate!

 

[koraks]

using a special modified gelatin, gelatin-methacrylamide (GelMA), and a water-soluble photoinitiator known to work well at 405nm

This reminds me of something I tried some time ago, but I didn't really get anywhere with it: https://www.photrio.com/forum/threads/hardening-gelatin-with-eosyn-and-tea.212869
I don't have (and am unwilling to invest in) gelMA, but the concept remains appealing.

 

[Yezishu]

This reminds me of something I tried some time ago, but I didn't really get anywhere with it: https://www.photrio.com/forum/threads/hardening-gelatin-with-eosyn-and-tea.212869
I don't have (and am unwilling to invest in) gelMA, but the concept remains appealing.

My general view is similar to yours. Eosin method seems to rely heavily on some residual groups in the gelatin, resulting in low curing efficiency, thus requiring a large amount of TEA. If limit to gelatin, I may consider whether known effective gelatin photoinitiators (in this case, DAS) can be molecularly modified to shift the absorption wavelength to longer wavelengths(It is possible that the required molecules are not industrially produced and are therefore expensive), or whether mixing eosin and DAS can transfer energy from eosin to DAS to excite gelatin curing(it was not considered in detail, the possibility of failure is not small, but can have a try).

Otherwise, In acrylate system(not gelatin system.....), GelMA(if a nearly completely biocompatible solution is required), HEMA or PEGDA are proven aqueous acrylate monomers that can be well combined with various photoinitiators suitable for various wavelengths to form hydrogels that immobilize nanoparticles. It's not surprising that some people have achieved good results with these monomers.

Further information: These two papers provide more reference for protein (gelatin) cross-linking.

Frontiers | Photocrosslinkable natural polymers in tissue engineering

Natural polymers have been widely used in scaffolds for tissue engineering due to their superior biocompatibility, biodegradability, and low cytotoxicity com...

www.frontiersin.org

Visible-light-induced protein labeling in live cells with aryl azides - Chemical Communications (RSC Publishing) DOI:10.1039/D2CC06987C

From those papers, I believe some DAS-like aryl azides like 4-azido-N-ethyl-1,8-naphthalimide can be used in visible-light-induced protein cross-linking, their research aimed to avoid damage to cells and biomolecules from short-wave UV light, but is expensive. An old German Pat. No. 514,057 disclosed water-soluble bisazido cross-linker such as sodium 4,4'-diazidostilbene-2,2'-disulfonate or sodium 1,5-diazidonaphthalene-3,7-disulfonate.

Another system, Diazirine-based crosslinking agents, are available commercially, for example,

https://www.researchgate.net/publication/287759570_Development_of_Diazirine-based_crosslinking_agents_for_covalently_linking_protein

https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202500414

https://xlynxmaterials.com/wp-content/uploads/2024/03/BondLynx-BLD-201-Instructions-v3.09.pdf

This one, BONDLYNX BLD-201 provides instructions for curing at 395nm. I believe this type of product can also be effectively used for gelatin curing.

 

[PGum]

This is a 2023 test that I performed using DAS and Fish gelatin. I spread it with a comb. It doesn’t solidify at RT as pork or bovine gelatin does. Image is soft as I exposed from an overhead UV lamp, through the 1/8” glass thickness. I had no issue with colour density or scale. I was going to invest in a coating rod but other projects took front row.

SbQ might be your ticket if a coating rod could be obtained. If the SbQ scale is short, you might be able to do multiple registered layers as is often done on paper

 

[PGum]

I never mentioned the use of ferric chloride + acetic acid! In this back of the glass exposure I coated a mix of phthalo blue and quinacridone magenta in iron iii chloride, acetic acid, and fish gelatin. I did a few of these in 2021. The ph has to be very low for this to work. Oddly it would not image on paper. There is some type of free chlorine radical produced as far as I remember on reading about FeCl3 photolysis. The gelatin gets so hard, I has to take a nylon scrubbing pad and plenty of elbow grease to remove it. Exposure is very slow, maybe 5-8 x longer than classic cyanotype. It made the most perfect tonal range however.

 

[koraks]

I never mentioned the use of ferric chloride + acetic acid!

I never heard of that! That's amazing. I've got a big jar of ferric chloride solution here for etching PCB's - if I acidify that, it's supposed to work? Would other acids work as well?

 

[Raghu Kuvempunagar]

I never mentioned the use of ferric chloride + acetic acid! In this back of the glass exposure I coated a mix of phthalo blue and quinacridone magenta in iron iii chloride, acetic acid, and fish gelatin. I did a few of these in 2021. The ph has to be very low for this to work. Oddly it would not image on paper. There is some type of free chlorine radical produced as far as I remember on reading about FeCl photolysis. The gelatin gets so hard, I has to take a nylon scrubbing pad and plenty of elbow grease to remove it. Exposure is very slow, maybe 5-8 x longer than classic cyanotype. It made the most perfect tonal range however.

Very interesting! If I read correctly, no additional step (such as Peroxide bath after exposure) is required for the hardening of gelatin in this approach. Is that right?

 

[Yezishu]

I never mentioned the use of ferric chloride + acetic acid! In this back of the glass exposure I coated a mix of phthalo blue and quinacridone magenta in iron iii chloride, acetic acid, and fish gelatin. I did a few of these in 2021. The ph has to be very low for this to work. Oddly it would not image on paper. There is some type of free chlorine radical produced as far as I remember on reading about FeCl3 photolysis. The gelatin gets so hard, I has to take a nylon scrubbing pad and plenty of elbow grease to remove it. Exposure is very slow, maybe 5-8 x longer than classic cyanotype. It made the most perfect tonal range however.

Amazing......perhaps you are the world's first to do so. The reason why images cannot be formed on paper may be that paper’s cellulose hydroxyl groups that compete for free radicals.

The closest thing I can find is a combination of ferric chloride and PVA for photosensitive applications (as a less-than-ideal alternative to potassium dichromate).
Manivannan, Gurusamy, Rupak Changkakoti, and Roger A. Lessard. "Metal ion doped polymers as media for optical memories." Polymers for Advanced Technologies 4.10 (1993): 569-576.
Olivares-Pérez, Arturo, et al. "Voltage Effect in Holograms of Polyvinyl Alcohol with FeCl3." Polymers 6.3 (2014): 899-925.
For reference, 7% PVA solution: 10% FeCl3 solution = 10:1~10:5 (w/w), pH 2~2.5, coating to 44 um thickness. 445 nm laser exposure, 10~15 J/cm2 energy

One paper proposed a combination of copper chloride and PVA, suggesting that gelatin might also be a potential solution？
Olivares-Pérez, Arturo, et al. "Holograms in polyvinyl alcohol photosensitized with CuCl2 (2H2O)." Optical Engineering 50.6 (2011): 065801-065801.
For reference, 25% PVA solution: 50% Copper Chloride solution = 10:1~10:3 (w/w), pH 2~3, coating to 200um thickness. 442 nm laser exposure, 2~5 J/cm2 energy

 

[PGum]

Yezishu, I haven’t tried pva. Might be worth a shot. Your suggestion about the paper interfering is what I was speculating as well.

if this exposure wasn’t so slow, I might have moved forward with it. I would expect it would work for carbon with some process adjustments and lots of exposure patience. It seems that the quantum efficiency is low for FeCl3, and I don’t think there is a way to speed things up. I tried FAC, FAO as additives but it was no faster and hardening was less. Nonetheless, here are my notes:

My emulsion:

2 ml Fish gelatin (glue) of 22.5% w/v)
( essentially fish glue diluted by 50%)

add pigment to fish glue first and mix well, or you get pigment agglomeration. I used pigment dispersions.

then, mix up:

0.4 ml Ferric chloride 42 deg Baume ( as copper etchant)
mix with:
1 ml acetic acid 10% w/w

add the FC + Acetic acid to the FG+ pigment mix, mix very well. It might gel a bit at first but that will setttle down ( the acid addition is used to eliminate this Fe+ Gel complexing!)

Pour on glass, spread with comb.

I fast dried it with a hair dryer from the bottom of the glass.

If using carbon black, exposure will lengthen further. I would anticipate in that case, 30 mins at least in full summer sun. If using an exposure unit, it needs to put out a lot of UV or its a very long wait.

Without the pigment, the emulsion colour is yellow and bleaches out with exposure.

To develop:

I used 1% sodium bisulfate in water, rocking with the glass face up for about 3-4 mins. You need the acid or the iron will hydrolyze when it hits plain water and nothing will wash out, as it will bind to the gelatin. If done by mistske, you can recover it by going back to the acid bath.

Process is sensitive to these emulsion quantities. It may be pH related?

3 drops 70% glycolic acid works well in place of the acetic acid.

 

[Hologram]

I never mentioned the use of ferric chloride + acetic acid! In this back of the glass exposure I coated a mix of phthalo blue and quinacridone magenta in iron iii chloride, acetic acid, and fish gelatin. I did a few of these in 2021. The ph has to be very low for this to work. Oddly it would not image on paper. There is some type of free chlorine radical produced as far as I remember on reading about FeCl3 photolysis. The gelatin gets so hard, I has to take a nylon scrubbing pad and plenty of elbow grease to remove it. Exposure is very slow, maybe 5-8 x longer than classic cyanotype. It made the most perfect tonal range however.

This is interesting.

If you replace acetic acid by tartaric acid, you will have the system Poitevin introduced around 1860.
At a ~ 4 : 1 (ferric chloride : tartaric acid) weight ratio the warm gelatin solution will not coagulate and can be used for coating.

 

[PGum]

Citric acid wasn’t so great for this. It may be pH effect or radical quenching. I never tried tartaric acid.