Trying Zerochrome-SbQ (PVA-SbQ)

[AndrewBurns]

Nope these were exposed with my LCD screen contact printing setup, so true continuous tone and pretty well collimated light. You should probably ignore the aluminium plate I posted before because as I mentioned I used totally the wrong exposure times by accident (I used the inverse exposure times I was using when exposing my 6 test strips, rather than exposing each layer for the correct time during the stacking process).

That said, I did another test print with the 'correct' exposure times and the test pattern still came out far too dark, but this is probably because I really need to spend more time fine-tuning my exposure times for each layer so they get the correct amount of overlap rather than just trying to go straight to a final print.

Here's my second attempt at this print on aluminium. This time I put down a layer of liquitex matte acrylic varnish diluted 1+1 with water before coating with emulsion and that made a huge difference to how well the PVA adhered.



Previously I had to lay down very wet sloppy coats and if I tried to brush more than once the previous layers would start to tear off the aluminium, but using the varnish to improve adhesion gave me tons of time to even out the coats. I didn't do as good a job making even coats for the lighter layers which resulted in some pretty visible non-uniformity. You can also see some registration error here too, but I think that's easily solved.

What I am happy about here is how clean I managed to get the print to develop. Kees has mentioned a technique called 'spray development' on the Zerochrome page which amounts to just spraying water onto the print with a spray bottle during development. I've noticed that I keep having situations where the PVA melts but then stays on the surface even during violent agitaton, and I've had to use a soft brush to get it off, which always seems to leave some stain and risks damaging the image. This time I tried the spray development idea and it worked brilliantly. The spray bottle I was using seemed just perfect at jetting enough water onto the surface with enough force to push off the unexposed PVA without doing any damage to the exposed areas, I'll definitely try this method again.

I'm quite hopeful that with properly hardened gelatin sized paper and the spray developing technique I should be able to get good prints with minimal stain, even more-so with aluminium sheets which naturally resist staining. Now I need to go back and be a lot more methodical in working out the relative exposure times of my different layers so that they combine to make a tonal range much closer to linear than what I'm currently getting...

 

[koraks]

pretty well collimated light

I just bumped my head myself on how well "pretty well" really is in practice...

 

[AndrewBurns]

I just bumped my head myself on how well "pretty well" really is in practice...

Interesting that in your carbon with a halftone screen you see an influence from exposure time. Conceptually, the cone of light coming through the screen doesn't get larger with time, and so there's no reason for the printed dot to get larger in this respect. I assume the gain from exposure time comes from a halation/irradiation type effect where light diffuses out sideways once it enters the light sensitive layer. I've seen this in my PVA-SbQ prints too, when following the layer stacking system the lighter layers are massively over-exposed and fine details like text and thin lines can end up several times thicker than they should be. I think in my case the effect is even more pronounced because the layers that get the most exposure also have the least pigment load, which means there's less opaque material preventing the light from spreading out sideways. If that's true then again I'd imagine that effect could be reduced by using a thinner tissue with higher pigment load (which you do anyway for other reasons).

 

[koraks]

I assume the gain from exposure time comes from a halation/irradiation type effect where light diffuses out sideways once it enters the light sensitive layer.

That's my thinking, too. Now, what I haven't tested is whether you get the same kind of effect if you have a shorter exposure time but a stronger light source. I assume it'll work out the same. So I think time isn't really the factor, it's basically the degree of (over)exposure.

fine details like text and thin lines can end up several times thicker than they should be

Yeah, and in that case, you can even visually tell that the structures are bigger than they should be. In my case, visual inspection of the halftone dots showed no anomalies, but the average density gain was still pretty massive. Which goes to show how much dot gain can be 'hiding in plain sight', so to speak.

 

[AndrewBurns]

Made another test print on aluminium sheet in an attempt to dial in my exposure. The per-layer exposure values for this test were less than half what I made my previous test print with and the exposure still isn't right, so yeah the last test I did was definitely very over-exposed on every layer.

Here's a plain photo of the 6 different layers, each with 21-step scale exposed at the correct time for that layer. They look very light in this photo, but they get significantly darker when stacked on top of each other.



Here's a version I edited to make the scales stand out more to make them easier to read. I've added red boxes around the dynamic range of each layer.



Ideally, the dynamic range of each layer would slightly overlap with the layers before and after it, such that when they're all stacked on top of each other there's a smooth transition through the whole range. Clearly that isn't yet the case here, but it's not terrible which is nice to see.

Lots of interesting stuff going on here.

• The 1.2 second exposure for the darkest layer looks about right, the darkest and second darkest patches are slightly different and it shows a surprising amount of tonal range (8 steps out of 21)
• The next layer, at 2 seconds, shows no overlap with the first layer meaning it needed less exposure. The dynamic range has also dropped down to 4 steps.
• Layers 2,3 and 4 show a fairly linear shift in the tonal range but that flattens out massively for layer 5
• Layers 5 and 6 are the same, there must be some non-linearity here (I think the density of the LCD screen is non-linear towards the darker tones)
• You can clearly see the thickness of the lines and text increasing with exposure, this is the halation/irradiation effect we were talking about before. It might even be worse when printing on metal sheet which could be expected to reflect a fair bit of light back up into the emulsion.

In this case each layer was 0.58 times the previous one as per Calvin's PMF formula but clearly my PVA-SbQ doesn't really work the same way. I think I'm going to keep my darkest layer at 1.2 seconds as that looks right, but move my lightest layer to the same exposure time as my second layer in this test (10 seconds). This works out to each layer being 0.65 times the previous one. This will have the effect of pushing the 'red boxes' of all of the layers except the first one up the scale, which should make them all slightly overlap each other.

 

[AndrewBurns]

Best result so far, but still some work to do.

6 layers of PVA-SbQ tinted with lampblack on gelatin sized and hardened Hahnemuhle platinum rag (still wet straight out of the bath):



I think I need to reduce my exposure of the lighter layers and maybe add another dark layer because the scale is currently too dark overall but the dmax isn't as dark as I would want it either.

I also had to liberally use 'spray development' to clear the paper as much as I did and there's still a bit of staining that didn't want to shift. A clear coat does seem to help with the staining so I might try that next, I'm also curious about putting down a white pigmented base coat, which would hopefully serve the same role but also prevent what I think the clear coat does in terms of reducing dynamic range.

 

[AndrewBurns]

Out of interest, here's the measured vs. target density for the above test print. Definitely over-exposed for most of the tonal range which I'll fix, but also I think if I had to calibrate and use it as-is I'd get decent results (it's better than the cyanotype curve I've been using for ages).



Also the measured L* value for the densest patch is about 43, so nowhere near as dark as I'd like. Weirdly this same stack of layers gave me an L* of about 30 for the darkest layer during earlier testing but I think that my brushing technique is different now and it's resulting in thinner (and lighter) layers.

I've noticed that when brushing on the darker layers after the earlier ones have been exposed there's a very clear latent image on the page which I believe is because the hardened layer swells slightly resulting in a raised surface (you can see a slight relief before developing). This raised and hardened surface means that as you brush on later coats the layer of emulsion over the top of previously exposed areas is thinner than it would otherwise be because the brush drags more emulsion off the surface of the raised areas vs. the lower unexposed areas. So I guess the takeaway here is that spectro measurements of the emulsion brushed onto paper won't necessarily match what you get when making a real image and brush technique makes a big difference.

 

[koraks]

Out of interest, here's the measured vs. target density for the above test print.

Looks like you have serious dot gain going on.

 

[AndrewBurns]

Been a while since I worked on SbQ however I recently got probably my best test-print results yet, at least in terms of no/minimal pigment staining from normal development (not having to resort to using a spray bottle or manual brushing to clear the print) and good single-layer dynamic range.

The biggest issue that has been haunting me so far with SbQ is getting the print to clear, even in the unexposed areas I can see the emulsion softens and partially melts away but there has always been a stubborn 'gelled' layer right on the paper surface that can't be shifted even with vigorous agitation, which results in heavy staining. I've had to resort to using a spray bottle to spray the print which shifts the gelled surface layer but also damages and degrades the hardened print layer as well. I've tried a number of things to fix this, including printing on what is considered to be the gold-standard substrate (Hahnemuhle platinum rag sized with hardened gelatin), adding a sub-layer of clear SbQ etc.

My most recent testing has resulted in the print completely clearing back to paper-white with vigorous agitation but no spraying or brushing, and also a reasonable dynamic range from a single layer. I've changed a number of things all at once and so I'm not sure what single thing or combination of things has helped, and I'm going on holiday now so I won't have a chance to do more testing for a while. Things I've changed that might have made a difference:

• Adding an extra-thick first layer of clear PVA-SbQ (1+4 with water instead of 1+8 that I have been using), which is dried and then over-coated with the pigment layer before exposure
• Using Calvin's black colour paste in my emulsion rather than india ink or dry lampblack pigment
• Not adding any surfactant to either emulsion
• Developing in water made alkaline through the addition of sodium carbonate

Here are two test prints I made with this method yesterday, one is a simple card where one half was fully exposed and the other half not to test for staining. The other is an exposure test from 0.5 to 4 seconds in 0.5 second increments (ignore the fact that the exposure test runs off the page, the numbers are mirrored and there's a dirty streak through the middle of it!).



I know it doesn't look that impressive but I'm happy with it! The exposure test is still a little dark for my target first layer (L* of 62, while I'm aiming for more like 70). Hard to tell from the photo but the ideal exposure in this case looks to be 2.5 seconds, with 8 steps of dynamic range (out of 21).

When I get back from holiday I'll need to sort out my pigment loads and exposure times for each layer using the new method and see how it goes stacking multiple layers to get a wider dynamic range and better Dmax.

 

[koraks]

That sure does look very clean! Are you planning on figuring out which combination of factors has resulted in the reduced staining? The reason I'm asking is because I strongly suspect that the first and fourth factors do most of the 'heavy lifting' here.

 

[AndrewBurns]

That sure does look very clean! Are you planning on figuring out which combination of factors has resulted in the reduced staining? The reason I'm asking is because I strongly suspect that the first and fourth factors do most of the 'heavy lifting' here.

Yeah, just got back from holiday but I do plan to see if I can separate out the various factors. It would be nice to be able to use dry pigments for example.

 

[koraks]

There's no doubt in my mind that dry pigments will work.

 

[AndrewBurns]

Still working on printing on paper, but tonight I tried a bit of an experiment only really made possible with my UV projector.

One advantage of using a projector for exposure that you might not think about is the ability to project a sharp image onto the back-surface of a sheet of glass. This isn't possible with contact printing because the thickness of the glass between the negative and the rear surface causes a lot of diffusion, but with a projector it's possible to focus the image onto the rear surface of the glass and then it's perfectly sharp.

Also as an aside I found that fluorescent paper like post-it notes make great focus aids as they re-radiate the UV very strongly as visible light:



The reason I would want to project onto the rear surface of a piece of glass is to harden whatever emulsion I'm using from the substrate upwards, rather than from the top downwards. I figure that this would significantly increase the dynamic range of the PVA-SbQ process vs. the traditional method of top-down exposure, allowing me to get a full range from a single exposure and layer rather than multiple coats and exposures.

To test this, I coated a piece of glass with some 2% colour-paste PVA-SbQ (dilute 1+8 with water). I initially tried brushing but to get a thick enough coat to be opaque I had to pour the emulsion on and spread it around more like coating a wet plate.



I tried drying the plate using a fan heater, but the table it was sitting on was not level and the heater dried it unevenly, resulting in a significant variation in emulsion thickness across the plate. I think next time I'm going to try making a much more concentrated PVA mixture as a more viscous layer should be more even.



Nevertheless I tried doing an exposure test, from 6 to 60 seconds, and to my genuine surprise I actually got some really promising results! There's some damage from when I dropped the glass plate into the sink during development, but the PVA layer was actually fine during development, and appears to be well bonded to the glass. There's a bit of base-fog/stain but it's tolerable (a thicker layer with lower pigment loading in the future would reduce this).



Waiting for the plate to dry now, but I'm really encouraged by the results and I think this process definitely has some legs, assuming of course you're happy with the print being on a transparent substrate! Also it looks like I'm getting full exposure in the 10's of seconds range, which is pretty crazy given I'm using a projector.

 

[koraks]

Maybe it's possible to transfer the print form the substrate to paper, just like we do with double transfer carbon. I've never tried this with PVA, but in my mind, it's not a stretch. Worth a shot at some point.

I do wonder if you really will have much of an advantage in terms of limiting the tonal threshold; i.e. the issue of the flaking highlights that's inherent to pigment processes. Having said that, I do see how an expose-through-the-back approach will work better than a direct print onto paper.

 

[PGum]

Overall those are good results. The exposure range looks a little less than one would get from gelatin+DAS or dichromate, but looks useable. I also performed a similar test of SBQ on glass with back exposure ( Digital negative with UV lamp at a distance) and I see similar results: A narrower linear range, weaker adhesion to glass when wet VS gelatin+. These issues can probably be solved. A coating rod might be helpful here. What I really like is the direct from display-screen approach.

 

[AndrewBurns]

The PVA is pretty well adhered to the glass. You can still rub it off with your thumb but I'm not sure how easily it would transfer onto another substrate. I'm pretty happy with being able to transfer onto glass at this point anyway as I have some images in mind that the process would suit.

Actually I think this technique would work best for carbon rather than PVA, I imagine that having the gelatin harden from the substrate upwards would significantly reduce the issue of highlights flaking away by giving them a ton of mechanical support (in the form of the glass). Again, impossible to transfer because the gelatin is hardened onto the glass surface, but fine for some things like orotone.

 

[AndrewBurns]

The nice thing about glass as a substrate is that I can scrub the PVA off and use it again during testing. My second attempt went way better than the first, true to the scientific process I changed several variables at once so I have no idea on the individual impact of each lol:

• Leveled my glass plate so that the PVA would cure to a more even thickness
• Put down a thin clear coat before the pigmented layer to prevent staining/fog on the glass surface
• Mixed up a much thicker (1+2 instead of 1+8 PVA to water) emulsion with half the pigment loading (1% rather than 2%)

Results are as follows:





Ideal exposure looks to be between 24 and 30 seconds based on this (3rd and 4th scale from the left). Next step will be to print a calibration sheet, although I have no idea how I'm going to use my reflection spectrometer to measure the patches. I can put the plate down onto a white sheet of paper like the second image but then there will be the thickness of glass between the instrument and patch which I'm sure will mess up the results. I might just have to use a camera on a copy stand and see if I can generate an adjustment curve based on eyedropper measurements of a photo of the plate.

 

[PGum]

Andrew,
Are these 1/2 stop increments?

 

[AndrewBurns]

Andrew,
Are these 1/2 stop increments?

Each of the scales is 21 steps from 0 to 255 on the LCD screen used in my projector (where 0 is the highest density the screen can produce and 255 is the lowest density the screen can produce).

Each scale is an additional 6 seconds of exposure, moving from left to right, starting at 12 seconds (the glass wasn't large enough to capture all of the scales). So the left-most exposure was 12 seconds, then 18, 24, 30, 36, 42, 48, 54 and 60 seconds. So it's a linear increase in exposure rather than the exponential doubling you get in a 'stop' scale.

 

[PGum]

Perhaps a reference test with identical glass and known targets would give you the compensation to apply.

 

[AndrewBurns]

Perhaps a reference test with identical glass and known targets would give you the compensation to apply.

I have thought about doing a test like that with the LCD screen set to maximum transparency (displaying a completely 'white' image). However there are two things in the system that need calibration, one is the process (how the polymer hardens in response to UV light, the density of pigment etc.) and the other is the LCD screen itself (optical density of the LCD screen vs. commanded value). I suspect that the LCD screen density is very non-linear at high densities, because I've never been able to have the lighter end of the tonal scale show any exposure, even at significantly longer exposure times than required to get full density at the darker end.

 

[PGum]

Certainly the display gamma compensation is going to make the whole thing very non linear.

Could you not just measure a bunch of greyscale patches on a sheet of paper with your spectrometer, lay glass of same thickness overtop, measure again and get your compensation to apply to the measurements from your printed glass?