Alright, I caved. I had been ogling the website of Kremer Pigmente for a few weeks and ultimately I decided to order some pigments from them. I think it actually makes sense, despite the warnings I was given by multiple people on the forums. After all, dry pigments are challenging to work with. Or so they say…
Why on earth mess with powdered pigments?
First: why does this move make sense? It dawned upon me when I was preparing my previous set of color carbon tissues. For 10 magenta tissues of 5×7″ I needed pretty much all of a small 5ml tube of Winsor & Newton Quinacridone Magenta watercolor. One of those retails at around €7, which boils down to nearly €0.70 per tissue. Just for testing! Alright, I do think those tissues were loaded a bit too heavily with pigment and maybe something like half that amount would be more appropriate, but it still adds up.
What’s the rationale behind watercolors, then? Well, mostly it’s a convenience item. The pigment comes neatly packaged in a format that makes it relatively easy to weigh out and dissolve in water. All the necessary dispersants, binders and probably preservatives are right there in the tube. But to be brutally honest – much of that stuff you don’t really need when making carbon tissue. If you mix in your pigment (in whatever form), pour the glop and let it dry over the course of one or two days, there’s really no need for preservatives. And any binders or fillers that are in there to make the product (a) affordable or (b) convenient to use for watercolorists don’t add any value for carbon printers.
The stuff I got, and the story behind yellow
So dry pigments it was. Well, the watercolor paints have been very useful so far for sampling pigments, so I knew what to look for. This is the set I ended up buying:
The yellow could do with some clarification, I suppose. Actually, it’s a story I promised back when I first wrote about balancing the pigments – which we’ll get to in a minute, hold on! Until now, I’ve actually been working with PY150 for the yellow pigment instead of the PY154 you can see I ordered as a dry pigment. While Kremer also carries PY150 (they call it ‘Indian Yellow imitation’), I thought PY154 to be a better choice. The reason is actually mentioned on handprint.com:
An apt name for PY150 might be “new gamboge deep”, as it has the same two toned appearance: in masstone it is a dull deep yellow (chroma of around 65) very similar to genuine gamboge (NY24); in tints it rises to a sunny, noticeably cooler and more intense (chroma 75) middle yellow.Bruce MacEvoy on handprint.com
I actually only read the passage above a few days ago when I had already stumbled upon this the hard way, as they say. You see, PY150 (nickel azomethine yellow) is very nicely transparent, apparently has pretty high tinting strength, and overall it seems quite suited to color carbon – if not for the fact that it’s not a single color. In high densities, it’s more brownish yellow and it only gets its nicely bright pure-yellow hue when it’s present in lower densities. That doesn’t seem ideal for color printing. I imagine it would probably work, of sorts, but color balancing would only become more complicated. So that’s why I jumped the gun and got the PY154 (a member from the benzimidazolone family) instead. Sadly, it’s much less transparent than PY150, but that’s apparently the price I’ll need to pay for better hue stability.
I also got some additives that should make my life easier: some Orotan 731K dispersant and some defoamer. I haven’t tried the defoamer yet, but I think it’ll come in handy. The Orotan tends to foam up things pretty severely, although so far in tissue-making it hasn’t been a real problem yet. Oh, and one more thing: I got me a proper paint muller. I went for the small one because these are strangely expensive items and I figured I would be working mostly with very small volumes anyway. Turns out that was a good choice, because so far it has worked very nicely for me!
Working with dry pigments: disperse!
So from that you can already glean that the dry pigment endeavor wasn’t exactly as daunting as some people apparently believe. The main challenge is accurately weighing out the stuff. For a typical test batch of glop, which is 100ml or 200ml (5 to 10 5×7″ tissues), I only need around 100-300mg (yes, 0.1g to 0.3g) of pigment. That’s not a whole lot. I think this is still on the high side; the first batch of test tissues looks pretty darn dense for how little pigment is in them. Once I’ve figured out the actual quantities, I think I’m going to make some concentrated pigment stocks at around 10% concentration or so, that should make life easier in terms of measuring things out.
How does it work, as far as I can tell from just one or two days of mucking about with these colorful powders? Well, I use my cheap Chinese jeweler’s scale to weight out the dry pigment and dump it onto a matted piece of glass. I happened to have a piece lying about that originally was intended as a diffusor for my color enlarger project, but it had some, err…issues. Turns out it’s pretty usable in this application! I then add one or two drops of Orotan, a couple of drops of water and carefully grind everything together using the muller. It only takes a minute or so to wet the pigment and to create a thick paste that can then be diluted with water.
Dry pigments vs. watercolors
Then what? Well, figure out how these pigments behave in relation to the watercolor paints I’ve become familiar with. Because to be frank, I had, and for the most part still have, not the faintest idea what to aim for in terms of pigment load in a carbon tissue. So I started with simply making some 1% w/v solutions of the pigment in water. The pigment has a tendency to sink to the bottom (especially the PB15:3 does this rapidly), but my own dispersions don’t seem to behave much differently from the watercolors in this respect. They’re more concentrated, that much was immediately clear to me!
I then did some tinting strength comparisons to see how the 1% pigment concentrations related to 1% concentrations of similar watercolors. I forgot to snap a photo, apparently, but what I did was pretty simple: take two identical beakers (200ml in my case), put 50ml of water into both, then add 5 drops of the strongest pigment dispersion (i.e. the dry pigment dispersion) into one. Now comes the matching process: to the other beaker, add drops of the other (watercolor) dispersion and keep doing so until both solutions look identical. This turns out to be not a very exact science – there’s a massive margin of error, no doubt. But it does give a ballpark comparison that allows me to estimate what the pigment concentration is in the various watercolors I had been using. This is what I came up with:
- Winsor & Newton ‘Winsor blue GS’ watercolor seems to contain around 10% w/v PB15:3 pigment.
- Winsor & Newton ‘Quinacridone magenta’ has around 5% w/v PR122 pigment.
- Winsor & Newton ‘Winsor yellow’ might be as much as 20% w/v PY154.
Alright, that last one I’m not so sure about, but I may not be all that far off. For instance, referring once more to MacEvoy:
The proportion of pigment to vehicle in tube paints generally ranges from less than 10% to around 20% of total volume for a finely divided, strongly tinting pigment such as the phthalocyanines, red quinacridones, dioxazine violet or alizarin crimson; from 20% to 30% for prussian blue, carbon black, the “raw” (uncalcinated) black and red iron oxides, zinc or titanium white, yellow quinacridones, benzimidazolones and most other synthetic organic pigmentsBruce MacEvoy, handprint.com
Alright, so less than 10% to around 20% for the PB15:3 and PR122 sounds about the right kind of range if I estimate them at around 5% and 10% respectively. And the benzimidazolones live at around 20%-30% in a typical watercolor, with my estimate being 20%. Interesting!
Which reminds me of the economics. Frankly, I hadn’t really spent much thought on it anymore after ordering the dry pigment, but now we come to speak of it anyway, let’s see how the financials work out. I had that tube of quinacridone magenta paint from Winsor & Newton (a very nice paint, by the way; it makes me wish I was a watercolor artist!) which retails at € 7 for a 5ml tube. This tube should contain around 5% of PR122, which works out as something like 250mg. That’s not entirely accurate as the density of watercolor paint is higher than that of water, so the contents of the 5ml tube will weigh more than 5g. So let’s say I get around 300mg of PR122 from this tube as a rough estimate. That makes for a little over €23 per gram of pigment. Gulp!
The Kremer PR122 costs € 5,14 if you buy the smallest 20g jar they offer (I actually got the 100g jar, which is considerably cheaper on a per-gram basis). This would be a bit less than…€0.26 per gram. For almost a factor 100 price difference I’m prepared to much around a bit and see if those dry pigments work! Besides, I call myself a tinkerer, so I better live up to the expectations that go with the moniker. Seriously though, I honestly hadn’t realized the difference was this gigantic – it’s a high price you pay for convenience!
All sorts of caveats, of course. I mean, is the pigment I got from Kremer the same quality and consistency as the pigment that Winsor & Newton put into their paints? I just don’t know – all I know is what I can see when I mix things up and compare. The PR122 looked pretty much identical; there was no difference I could observe. The PB15:3 I purchased from Kremer looks a little more blue / less green than the PB15:3 in the Winsor & Newton paint I used for my recent batch of tissues. The difference doesn’t worry me, but it’s a sign that we’re not necessarily comparing apples to apples. The Kremer PY154 again looks identical to the one I got from the Winsor & Newton paint. If it walks like a duck, quacks like a duck…well, there still could be very duck-like geese hiding among the crowd. Or a chicken with a severe identity crisis.
Getting rid of the colors again
I promised I would get back to the matter of mixing grey from three primaries. This is one of the things I needed to figure out with the dry pigments as well, so I did the same as I did last time: I started by mixing drops of the aqueous pigment dispersions until I got something close to a neutral grey to get a general feeling for the kind of ratios I would be working with.
But I also learned last time that watery dispersions may not always behave the same as gelatin emulsions, color-wise. So I also mixed some ‘test glops’ consisting of 8% gelatin and 1% of each pigment I trialed, and then mixed drops of those to see when I would hit a neutral grey.
The image above is one of the tests I did. They’re drops of mixed pigmented gelatin with varying droplet ratios of cyan, magenta and yellow. The numbers under each blob actually represents this droplet ratio; so for instance, 4/4/12 indicates 4 drops of cyan glop, 4 drops of magenta and 12 drops of yellow. Turns out that this is also the approximate ratio that makes grey.
Just to rule out any problems with drops not being good proxies for a true volumetric measurement (i.e., milliliters), I tested 0.5ml/0.5ml/1.5ml (the same 1:1:3 ratio) in the bottom right corner. I boxed these two swatches; the upper one is a mass tone so to say, so a heavy drop that dries to near black, and the lower one is diluted with water to make a lighter grey. They looked alright to me – I could finetune it further, but I’m not convinced that at this stage, it would make sense to do so. Elsewhere in the process there are all sorts of inaccuracies (e.g. measuring 100mg of dry pigment!), so getting the last percent of accuracy from this step doesn’t seem sensible.
And thus, I had just about everything I needed to make the first batch of tissues with the dry pigments! Well, apart from a concrete answer to the question how much pigment I should put into the glop in the first place. Yes, I knew the approximate ratio, but would a 1%:1%:3% (C:M:Y) pigment load be a good starting point, or 0.1%:0.1%:0.1%…? I opted for the latter as it seemed to approximate what I was doing last time with the watercolors. Keep in mind I also worked out at least approximately how the dry pigments relate in strength to the watercolors.
My previous tissue recipe called for 0.67% of ‘Winsor blue GS’ watercolor, 2.0% ‘Quinacridone magenta’ and 0.50% ‘Transparent Yellow’ – all Winsor & Newton watercolors. Since the yellow is a different pigment (the W&N paint was PY150-based and I now switched to PY154) that one couldn’t work as a benchmark. Also note that somewhere, I’m off in my estimates, as the cyan:magenta ratio I determined as mixing to a neutral grey was vastly different for the watercolors (1:3) than for the dry pigments (1:1) – so in the end I just went for 0.1% of the cyan and the magenta dry pigment as a starting point and 0.3% for yellow, and then just see what happens.
I only made 5 tissues each to do some initial test, and then finetune. They’re currently hanging to dry, and frankly, they look pretty darn saturated. Certainly not any less than the watercolor tissues I made last time, so this is quite promising. I suspect I’ll have to actually drop the pigment load for the next round.
Just prior to making the tissues, I also made a little color mixing triangle with the 1% test glops, for no particular reason than they’re kind of cool to look at. And they do give a bit of a feeling for how the colors mix and what kind of saturation costs I can expect when working with these pigments. I.e., how vibrant will the greens, oranges and violets really be? Here’s what I came up with:
The extremes of the triangle are the pure glops; all other patches apart from these three are mixes. I tried to hit neutral black for the center patch and pretty much got it as well, using the ratios I determined earlier. The other hues look…well, colorful. I haven’t done any measurements on them, but I think it’s pretty clear that especially the greens are not so high in chroma, neither are the oranges. The violets are interesting, but the saturation penalty is visible here as well. Not saying that these pigments won’t be able to produce nice prints. I don’t doubt they could! But it’s also evident to me why additional spot colors would be preferred by some. Indeed, it’s pretty much the reason why printer manufacturers like Epson switched over to ink sets with additional colors besides – a CMYK-OG set: cyan, magenta, yellow, key/black, orange and green. Note how orange and green fill two of the saturation gaps we incurred by relying on mixing magenta and yellow and cyan and yellow, respectively.
One more thing: note the two blobs top left in the image above. They’re supposed to be neutral grey. The top one doesn’t bother me too much; it’s a drop of each of the glops I made for the tissues that are currently drying mixed together. It’s kind of off – pretty far. It’s too green, so lacks magenta. I then made the one below by mixing 0.5ml of each glop, which is generally more precise than taking one drop each. It’s closer, but still not perfect, and apparently not as good was what I could make with the test glops I had before. This goes to show that measurement inaccuracies are still present, and I’ll probably need to solve them. One day. Promise…