Previously, I have written about my attempts to make DAS carbon work well for continuous-tone negatives. The main challenge is highlight rendition, with flaking problems destroying delicate highlights. After testing many factors, I think the only conclusion left is that there is a limit to what DAS can do. Moreover, this limit is different (and more, well, limiting) than with dichromate carbon transfer.
Mission statement
Let’s recap for a minute what I’m actually trying to do. Take this 8×10″ negative here – it’s Fomapan 100 film, developed in Pyrocat HD, then intensified using chromium intensifier and redevelopment in again Pyrocat HD. If memory serves…my note keeping isn’t always perfect. This is what it looks like:
I’d call it ‘moderately contrasty’, and no, I don’t have a transmission densitometer, let alone one that will measure UV transmission specifically, so it’s going to remain a subjective assessment. I figure this would print OK at around grade 1 silver gelatin paper or so. Due to the dye stain from the Pyrocat developer, it’ll present some more contrast to a UV-exposed process (such as carbon transfer) than it does to the green/blue light silver gelatin works with.
This negative is significant, because I’ve done a truck load of testing with it over the past few months. There are two main reasons for this choice of a negative. Firstly, it’s fairly typical of the kind of negative I make using this film. Without spending too much thought on tailoring the negative to the process, this is what I get. Yes, that’s a haphazard approach, and yes, that’s biting me in the posterior right now. And yes, I was aware that it might, so I’m not surprised or disappointed.
Secondly, this negative features something very useful: a dense area that represents a highlight section with a fairly smooth gradation, and some very low density areas close to it that represent deep shadows. In other words: it encompasses a representative (insofar as that exists) tonal range, and there are some areas that can be printed as delicate highlights. The latter is the Achilles’ heel of any pigment process, so this negative is very useful as a ‘torture test’.
And yes, I might have just used any old step wedge negative with decently closely spaced density steps, but I happen to not have one handy with a useful density range. Besides, printing step wedges for some reason makes me depressed, whereas printing a negative of my fiancée’s legs compensates a little bit for the inherent dullness of endless technical testing. I’m only human, after all.
What I’d ideally like, is to be able to print any negative like this one with a DAS-based carbon process. Simple as that. I can do this without too much difficulty using dichromate. But this involves the inherent problems of dichromate – it’s a nasty chemical to begin with, being toxic, carcinogenic and dangerous to pretty much all life forms. Then there’s the consistency problem associated with brushing a sensitizer onto each individual tissue, which makes it challenging to make two identical prints – or even print a test strip and then reproduce the desired exposure and contrast with a full print. And I find it inconvenient to sensitize a tissue, then having to wait until it dries before I can continue printing. So DAS is nicer in that it can be incorporated into the tissue, and this DAS-sensitized tissue is then perfectly consistent and ready to go at any minute.
So in testing DAS carbon, I’ve done a lot of work adjusting the ratio of the different tissue constituents to get the desired contrast range as well as good highlight rendition in the same print. And I’ve pretty much failed in making this happen, at least for this negative. What’s up with that?
Solutions to the problem
It’s not like the problem is due to some previously unknown suspect. It’s the “good old” tonal threshold, as Calvin Grier calls it. Pigment printing processes where the pigment is suspended in a colloid (e.g. gelatin, or gum arabic) layer all suffer from this. To render highlights, you effectively deposit a very thin layer of this colloid onto the paper. Since it’s thin, there’s not a lot of pigment to absorb incoming light, and consequently, it looks as a low density.
There’s a limit to how thin a layer of colloid will reliably survive processing. This means that there’s also a limit to the lowest discernable optical density you can print. Below that limit, the colloid layer breaks down and only pure paper white is left. This is what Calvin calls the ‘tonal threshold’: the lowest optical density that will reliably print.
There are a couple of obvious solutions to this problem – although they are effectively workarounds only. They boil down to two main approaches, as far as I can tell:
- Make the layer of colloid a little thicker for the same optical density, so it withstands processing better.
- Make the colloid matrix sturdier, so that thinner layers will still survive processing.
As to (1), the most obvious solution that is often employed is to reduce the pigment load. After all, for the same layer thickness, a lower pigment load will render a lower optical density. At certain point, the thinnest layer of colloid that will withstand processing will contain so little pigment that it is barely distinguishable from paper white. In this case, the tonal threshold is shifted so close to no density at all, that the problem effectively vanishes.
Another solution is to print multiple layers on top of each other. This really boils down to the same as above, because you’d typically print one layer with a very low pigment load, pushing the tonal threshold as far as possible towards paper white. But with multiple layers, you’d put additional image layers on top of this with higher pigment loads, to achieve the desired optical density for the lower values (lower midtones and shadows). Conceptually, it’s really the same as above.
A third solution in category #1 is the use of halftone screen negatives. By imaging low densities as tiny dots with a lot of white space between them, you can still use a relatively thick layer of colloid and/or a relatively high pigment load for each individual dot, but the highlights will still look delicate. Effectively, it means you don’t really need a printing process that is capable of rendering very delicate highlights on its own. You work around the problem by using white space between individual dots to render highlights. This is the approach propagated by many DAS-carbon printers, including Calvin Grier, Charles Berger (the inventor of DAS carbon), Michael Strickland and Tod Gangler.
The second category is a bit more diffuse in terms of techniques involved – at least as far as I can tell at this moment. Material choices would fit in here, such as different grades of gelatin. I’ve been trying a few gelatins over the past few weeks, and the higher bloom gelatins seem to have a bit more to offer when it comes to building thin yet sturdy layers. By extension, I would also fit the choice, surface treatments and sizing of support materials in this category. While this does not influence the gelatin matrix as such, it does make a difference how well a thin gelatin layer will adhere to the support. Untreated Yupo I still find to work quite well (although Kees Brandenburg assures me that Agfa Synaps works even better), and I have found quite distinct differences in sizing or subbing approaches for plastic (polyethylene, acetate) film.
And I’ve come to realize that in this second category, one choice in particular makes a massive difference: the hardening agent used. It’s not a new discovery on my part and people have said it before, but I can only conclude now, based on my own experiences, that dichromate makes for a lot sturdier gelatin matrix than DAS. Simply put, with dichromate, it’s possible to make very thin gelatin layers that are still very sturdy and robust. With DAS, similarly thin layers will be more brittle and disintegrate a lot easier. I have to admit I had underestimated the magnitude of the difference, and its practical implications.
The DAS contrast/highlight tradeoff
Taken together, this means that there is a limited bandwidth of pigment concentrations in a carbon tissue that will work for single layer transfers. The requirements of interest on such a single layer print are (1) a full tonal scale from convincing black to paper white and (2) smooth rendition of highlights, with an imperceptible transition from visible tone to paper white. These requirements are naturally at odds with each other.
An effective way to get a full tonal range (i.e., large c0ntrast in the print) is load the tissue with a lot of pigment, so it doesn’t take a lot of hardening of the gelatin to print a convincing black. However, a heavily pigmented tissue will require that very light tones consist of vanishingly thin gelatin layers, and those don’t render reliably. If we go the opposite direction and make a tissue with very little pigment in it, it becomes easy to render delicate highlights, but reaching a convincing black will be challenging.
This results in a certain usable process bandwidth in terms of pigment load. The limit at the higher pigment load is the point where the tonal threshold becomes acceptable and smooth highlight rendition is possible. The limit at the end of lower pigment load is where a convincing black becomes possible to be printed.
These limits are quite fluid: several parameters influence where they are. I’ve already mentioned the hardening agent, DAS vs. dichromate. This seems to influence the higher pigment limit considerably, since dichromate appears to allow for thinner gelatin layers to survive, and hence, we can get away with higher pigment loads while retaining an acceptable tonal threshold.
Other notable factors include the thickness of the tissue and the density range of the negative used. After all, to print a convincing black, we could either print a thin layer with a lot of pigment in it, or a very thick layer of gelatin with only a limited amount of pigment in it. The net result would be the same, since there’s going to be just as much pigment encountered by light that hits the paper.
To be able to make a thick gelatin layer, we need to expose it longer (assuming everything else remains the same). And that means that our negative needs to have a longer tonal scale. After all, if we expose longer through a low-contrast negative, our highlights (the dense areas in the negative) will be too exposed in the print, and light tones will end up too dark. As a result, the factors of pigment concentration, tissue thickness and negative tonal scale are intimately related:
The red arrow shows that to an extent, we can compensate for a lower pigment load by increasing the thickness of the tissue. Or, conversely, if we increase the pigment load, we can still make a decent black with a thinner tissue. The blue arrow shows that a higher pigment load allows for the use of a negative with a smaller density range. Or, put differently: if we want to print a low-contrast negative, a higher pigment load is an obvious solution. And, the green arrow: if we want to print with a very thick tissue, we will require a negative of a long density scale. Alternatively, if the negative has a limited density scale, we might as well use a thinner tissue. And if we do so, we will have to choose a high pigment load, because otherwise we won’t be able to make a decent black.
And that’s also the explanation why I kept running into a brick wall, hoping that it would be less ‘bricky’ than it really is. Remember that I set the negative as a fixed point on the triangle above, because it represented a typical negative I’d like to print. This means I could jiggle with the pigment load, and this in turn would allow or necessitate me to jiggle with tissue thickness.
That there would be a limit to how much pigment I could put into the tissue before running into problems with the tonal threshold, I knew beforehand. What I did not know, was at what point this would happen. What surprised me, is that this happens a lot quicker with DAS than with dichromate. For instance, a tissue with a Kremer Pbk7 pigment load of 2% of the weight of the gelatin (i.e. 0.2g pigment to 10g gelatin) easily prints very delicate highlights with dichromate. With DAS, I never managed to reliably do this – I might luck out once or twice, but it just wouldn’t work.
A note on sensitizer concentration
You might notice that I’ve conveniently ignored something so far: how much hardener / sensitizer is used. The main reason is that I haven’t varied this all that much throughout my DAS experiments. Yes, I did do some testing very early on, varying the DAS to gelatin ratio between 3% and 6%, since this was the bandwidth I saw being mentioned by others. For the vast majority of my tests, I kept it at 4%, since this is also the percentage suggested by Sandy King et al. in their book, where they provide an example DAS carbon tissue formula. Since this happens to be at the 4% mark (4g DAS to 100g gelatin), I figured that it should work. Except that for me, it didn’t. How come?
Well, there was (of course) a fatal flaw in my reasoning. This reasoning went something like: “4% DAS to gelatin ratio is a good one, because it’s proven to work by Sandy King. To print my benchmark negative, I need a pigment to gelatin ratio of around 2%. I know this pigment to gelatin ratio works, because this works perfectly fine with dichromate carbon. Hence, a 4% DAS tissue with 2% pigment loading should work.” But DAS isn’t dichromate.
A few days ago, I went back to another ‘benchmark’ formula for DAS carbon, in this case the UltraStable formulas from Charles Berger. I never really looked into these too deeply, but when I did, I noticed that (1) their pigment loads appear to be pretty darn high, and (2) the DAS concentration is really high, too, at 8% DAS to gelatin instead of King’s 4%.
Concerning the UltraStable pigment loads: these are a bit difficult to interpret, because they’re given as a wide range of around 13% to 67% pigment to gelatin. This sounds awfully high compared to my measly 2%, but the UltraStable tissues were made with pigment dispersions (probably pastes), and such a dispersion consists to a large extent of water. How much water? Your guess is as good as mine. At least half, and that’s a very conservative estimate indeed – I’d put it closer to 20-30% pigment and the remainder being mostly water and a little dispersant (could be gum Arabic, gelatin or some high-tech modern dispersant). Also, the UltraStable formula I’ve got does not specify the color, and they were made in CMYK sets, and the used cyan, magenta, yellow and black pigments would have been present in different concentrations.
As to the DAS loading of the UltraStable tissues: well, quite high indeed, and to be frank, I never tested such high concentrations until very recently. I admit such a test was probably a stupid thing to put off for so long. I mostly did this because (a) DAS is pretty expensive so I wouldn’t want to settle on a tissue with a high DAS load, and (b) doesn’t everyone work with lower DAS concentrations, so why go up that far to begin with? Again, sloppy reasoning, poor assumptions – a mess, in short.
Testing the 8% DAS load with a fairly high 2.67% pigment (powder Pbk7) load resulted in a tissue with rather good highlight retention – and quite low contrast. Just like with dichromate, increasing pigment load results in reduced print contrast. By this, I mean it takes less exposure to render the first visible tone (the toe of the print curve), and the resulting high density is lower (because the exposure is shorter). The highlight retention was significantly better than with tissues with a lower pigment load (e.g. 2%) and a lower DAS content (e.g. 4%). In other words: with the 8% DAS content, thinner gelatin layers can be printed that reliably survive processing. But print contrast will be lower than with 4% DAS tissues.
The contrast ceiling
And this introduced a rather unforgiving trade-off. Increasing the DAS load makes a tissue that enables thinner layers to be printed, so better highlight retention. Since it also reduces contrast, however, pigment load needs to be increased. And that, in turn, we need to print yet thinner gelatin layers to get the same highlights. So we’re again running into problems with the tonal threshold, which we can solve by increasing DAS load, reducing contrast, necessitating higher pigment loads, etc.
The above is a bit pessimistic, in the sense that it’s an infinite loop that’s difficult to escape from. A bit more optimistically, I could also put it in terms of what’s possible, yielding something like this:
Put into words, you might say that a certain sensitizer concentration will support a certain pigment load. Exceeding this pigment load will result in inevitable problems with highlight rendition, unless sensitizer concentration is also increased.
The problem with most of my DAS tests in the past few months, is that I was stuck in the red triangle area in the top of the diagram. The consequence of the diagram above is that there is a fairly firm limit to how much contrast a DAS carbon print can have. Increasing pigment without increasing sensitizer results in highlight disasters, and decreasing sensitizer to boost contrast also runs into a firm limit, since pigment load needs to be decreased along with it. If I were to dub a term for this, it would be the “contrast ceiling of the process“. Try to rise above that ceiling and you’ll bump your head!
Graphically, the ceiling might look as above. On the left side, print contrast is increased by a low sensitizer content, but at the same time, the low pigment load limits it. We can’t increase pigment load here, because highlight flaking will become visible. On the right side, we can use a higher pigment load to boost contrast, but in order to sustain that, we need to use a high sensitizer concentration as well. And that limits contrast. So in both cases, there’s a limit to how much contrast the printing process will give, or perhaps better put: how steep its characteristic curve can be made.
Caveats to the ceiling
The ceiling idea is more of a hypothesis than exact science at this point. It’s what I arrive at if I take stock of the tests I’ve done so far, which consist of a couple of prints across a couple of dozen tissue formulations and two generous handfuls of transfer supports / surface treatments.
The first thing that I really doubt is that combinations of low sensitizer + low pigment and high sensitizer + high pigment always end up giving the same maximum contrast. It sounds at least as likely to me that there’s either some kind of linear relationship between them (a or d in the image above), or a curvilinear one (b, c), or even a more complex relationship such as a logarithmic one (not pictured). I have not tested yet, nor have I reasoned it out, but it’s an interesting question to investigate further (given the magical combination of time + inspiration).
The second thing is that the ceiling is probably not very firm. I bet it can be pushed up or down a little by carefully selecting materials (gelatin comes to mind) and process control also plays a role. For instance, very careful processing might allow for the use of a relatively high pigment load with a moderate sensitizer concentration: keep development temperatures as low as possible, limit agitation, etc. It also seems that tissue soak times and the resting period between mating and warm-water development have some influence. This is something I might dedicate a separate blog to at some point, since the topic is somewhat contentious (see here, in particular the exchange between Sandy King and Charles Berger).
On the other hand, I have done many tests with parameters such as soak and resting times as well as development temperature and agitation, and in the end, the differences seemed to be marginal and any positive effects seemed to be very difficult to replicate. All this makes me suspect that the ceiling is real, alright. It’s just a little squishy, perhaps.
And then there’s the matter of the sensitizer itself. I’m focusing on DAS mostly here, and it seems that there are different quality grades this stuff comes in. The DAS I have, from two different sellers, is the lowest brick red quality seen in the black plastic bag in Calvin’s photos in the link above. It wouldn’t surprise me if higher quality, purer DAS would shift the required sensitizer for a certain pigment load down a little. The difference is likely marginal, but may be noticeable.
More importantly, it seems to me that the whole ‘contrast ceiling’ issue is just as true for dichromate as for DAS. I just haven’t run into it, yet (and maybe never will). That is to say, perhaps I have, but have never noticed it. I do know that very low dichromate sensitizer levels tend to do ‘funny things’, but I never bothered looking into it seriously, because they were really (really!) low levels and there was no real point in going there. I do recall that the very first phase of carbon printing I went through, I was using very horrible digital negatives and I was using high pigment loads and low ammonium dichromate sensitizer concentrations. Sure enough, highlights were virtually impossible to get right – but then again, so much was going on at that point…
The dichromate ceiling will be a whole lot higher than that of DAS – that much is clear to me. Pretty much any negative that will print OK at grade 3 paper or so can easily be made to print well with dichromate carbon – and probably thinner negatives too, at that. I just never tried. At some point, I expect you’d run into a similar unsustainable combination of sensitizer strength and pigment load where good highlight rendition becomes impossible.
Practical implications
Summarizing the implications of all this, for my particular use case: SOL! Which is to say: getting a print with good highlight rendition and a full tonal scale with a single DAS transfer from a moderately low contrast negative – it’s apparently just not going to happen.
Of course, there are several solutions, or at least workarounds. Firstly, the print can be made with dichromate instead of DAS, which has a higher contrast ceiling. Secondly, the negative itself could be changed, by expanding its contrast through one of several means (bleach & redevelop with staining developer, selenium or sepia toning, chromium intensification, etc.) Thirdly, there’s the possibility to simply use a different negative with a larger tonal scale – possibly an inkjet printed one, or even a halftone screen imagesetter negative.
So this isn’t the end of the world, and the only thing that got really shattered was my (perhaps naïve) expectation that DAS carbon would have roughly the same flexibility as dichromate carbon.
How flexible is DAS carbon, then, in terms of delicate highlights? That’s something I’ve been trying to establish over the past couple of days in additional tests. Here’s an example of such a test:
This is a scan of a 4×5″ test image with a step wedge (0.15logD steps) on top. The step wedge area was exposed for around 20 seconds – but that’s not very important. The relevant bit is the gradient that makes up most of the image. This was created by sliding a piece of cardboard over the carbon tissue in a single, smooth motion as it was being exposed under a strong UV light. Hence, the bottom end of the gradient has an exposure time of effectively zero seconds, while the top (in this particular case) was around 2 seconds. (Ignore the significant sensitizer stain; this was not cleared on this Yupo print.)
Ideally, the gradient should be very gradual and very smooth, showing no distinct transition from no tone to the first perceptible grey. In this case, there is an area around the middle of the image where the first tone starts to appear and from there, tone starts building up quite smoothly. I’m not entirely happy with this outcome, but it’s about as good as it seems to get with DAS. This print was made with a tissue with 8% DAS and 3.5% Pbk7 pigment (dry powder) to gelatin.
Yes, that’s a lot of DAS! Twice as much as mentioned in for instance the recipe in the carbon book of Sandy King, Don Nelson and John Lockhart. It seems I get the smoothest highlights with this high concentration; I’ve also tested 6% and 4% DAS to gelatin ratios with lower pigment loads, but they are inferior to the 8% ratio. I have not (yet) gone beyond 8%. So far, my tests suggest that highlights render OK-ish (fingers crossed…) at pigment loads of around 3% dry Pbk7 and 8% DAS to gelatin, or around (max.) 1% dry Pbk7 and 4% DAS to gelatin. The tonal scale of test prints from tissues of these formulations show a comparable tonal scale and contrast – all of which seems consistent with the ‘contrast ceiling’ line of thought put forth above.
Also, I observed in making these gradient tests that pure 365nm light works a little better for smooth highlight rendition than the combined 400nm + 365nm LED source I had been using for most of my tests. On the other hand, shadow densities build with much difficulty under 365nm exposure, so it seems I’ll have to do some more testing with optimal ratios of 365nm and longer wavelengths to get good highlights as well as convincing blacks. Coincidentally, I never quite understood why Calvin Grier emphasizes the dual wavelength approach so much in his carbon printing, but it seems I’ve now come to a point where I can actually see with my own eyes how it works. More testing is in order, though.
The main implication of all this is that I’ll just have to limit myself to much longer tonal scale negatives for DAS carbon. There’s not going to be much possibility of having ‘dual purpose’ negatives that print OK on a lower grade silver gelatin paper as well as carbon. With dichromate, this works fine, but with DAS, it just seems to be outside the process envelope – at least within my current understanding and capabilities.
Here’s a print made from a much more contrasty negative – Fomapan 200 exposed at E.I. 125 and overdeveloped by roughly 100%. This particular print was made with a tissue with 8% DAS and 3 % pigment to gelatin. It’s of course not the kind of torture test as the leggy picture I’ve shown before, but at some point I just wanted to see where my testing has gotten me at this point. So I exposed this portrait last week, developed the living daylights out of it and see how it went. This is a double transfer onto gelatin-sized (10g/m2 gelatin) Schut Laurier, by the way. It’s just air-dried, so a little wavy – hence the wiggly border and apparent unevenness in the background gradient (neither exist in the real print).
I’m not going to claim victory yet, but I hope, at least, that I’m starting to understand the material now in such a way that I can gingerly start making some images with it.
For good measure, I did borrow my fiancée’s credit card to order some Stouffer step wedges, which I hope will arrive in the mail in a week or so, for some more testing. My inkjet digital negatives really don’t cut it for continuous tone tests (way too much halftone screening going on in them), but I’d still like to have a little more solid footing than the DIY tests I can rig up with coincidental negatives and bits of cardboard. I’m still thinking of a way to make a consistent, continuous tone gradient…
Let’s see how it goes from here on.