Words and phrases, when overused, can be annoying. I have to say that I’m tired of hearing people call things “iconic” for one. I’m also tired of hearing about people “going down a rabbit hole,” especially when going down a rabbit hole is described as a chore in tones of feigned exasperation, but diving deep into a subject out of curiosity is a joy. I spend half my waking hours down rabbit holes and, thanks to the Internet, exploring underground tunnels is much easier than it once was. There’s no longer much of an excuse for not knowing anything you might want to know. The best one can do, it seems, is to point to the number of hours in a day when making excuses for ignorance.
I’m down a rabbit hole as I write this. Or perhaps I should say this post is the partial record of a rabbit hole expedition. You've been forewarned.
I’ve read that more than half the people in the United States never read a book after leaving high school (poking my head briefly underground into the opening of a new burrow, I found estimates ranging from 33% to 68% with the most credible-sounding estimate I could find – from The National Endowment for the Arts – being 58.5%). Thus, I’m in the roughly 40% of the population that does continue to read. At present, I’m in the middle of The Age of the Horse (Atlantic Monthly Press, 2016), by Susanna Forrest, a book that traces the history of the equine species and their interactions with human beings through history. On page 230, in a section of the book discussing equiphagy and how horse carcasses have otherwise been consumed, I came across a line that, yes, sent me down a rabbit hole – the rabbit hole I am in at present. The line “The hoofs themselves were boiled for glue, combs, toys, or Prussian Blue” stopped me in my tracks. Horse’s hoofs and Prussian Blue?
I started searching for a connection between horse hoofs and Prussian Blue. Being a dabbler in the arts, I thought myself quite familiar with Prussian Blue. I had no recollection of any connection with horse’s hoofs and it seemed to me that a horse’s hoof is probably mostly keratin (the stuff our fingernails are made of), which seemed to me unlikely as a source of Prussian Blue. And so, I began digging.
First, I looked into hoofs. I confirmed that a horse’s hoof is, in fact, mostly keratin. Setting that aside for the moment (keratin seemed rather inert), I dug into the history of Prussian Blue. If you’re not immediately familiar with the color, it’s what Crayola began calling “Midnight Blue” in its big crayon boxes in 1958 (because they thought most children by that time no longer understood a reference to Prussia). It is the blue of cyanotypes, blueprints, and the blue of Picasso’s Blue Period.
It was discovered by accident in Berlin in 1704 (some sources say 1706, apparently in error) when pigment and dye supplier Johan Jacob Diesbach, attempting to make Florentine Lake, ran out of the potash used in the Florentine Lake process. He called on another Johan, Johan Conrad Dippel, who topped up his potash supply. Dippel sold him potash contaminated with animal blood (apparently, that was accidental; Dippel used potash in the production of “Dippel’s Oil,” which involved boiling animal parts). When Diesbach went back to making Florentine Lake, he was surprised when the end result was not the reddish pigment he was trying to make but a deep blue, slightly greenish pigment – what came to be known as Prussian Blue or Berlin Blue – in chemical terms an oxidized ferrous ferrocyanide salt (Fe 4[Fe(CN) 6]) or simply iron ferrocyanide. In the Color Index Generic Name Code System (more about that later), Prussian Blue is PB27.
At this juncture, where to turn? What did Diesbach and Dippel do when it became clear that the contaminated potash caused the reaction resulting in Prussian Blue? Why did the animal contamination cause the unexpected reaction? What exactly is Florentine Lake? What is a lake? Although I already knew something about several of these topics, I wanted to know more.
Being a painter, I had some idea of what a lake is – essentially, a lake is a pigment made from a dye. So, what’s the difference between a dye and a pigment? The color in a dye is dissolved in the medium that carries it, while a pigment is a very finely powdered solid. Pigments are made into paints by suspending their particles in a medium but the particles are not dissolved in the medium. A lake is made by binding the dissolved color in a dye with a mordant. A mordant combines with the dye chemically to create a stable, colored substance that, unlike a dye, will not wash away easily – a pigment. Dyes, particularly dyes derived from organic sources, are often fugitive (that is, they are not colorfast; they fade badly). Most traditional lakes have been replaced by more permanent synthetic pigments that are chemically different but mimic their historical antecedents.
I wasn’t familiar with Florentine Lake. I learned that it is a transparent reddish pigment with bluish undertones and that it was made from a dye derived from kermes insects (a group of scale insects, in this case mainly Kermes vermilio), dried and crushed. Kermes vermilio is a parasite that lives on the Kermes Oak (Quercus coccifera) and the Palestine Oak (Quercus calliprinos), both native to the Mediterranean area. This is the source of a crimson dye that has been used since antiquity. “Kermes,” I learned, is the ultimate source of our word “crimson” in English. Florentine Lake then is a red pigment made from this crimson dye. (Genuine Florentine Lake is still available from at least one source, The Alchemical Arts, in Australia.)
The story of a scale insect producing a red dye is very familiar. I’ve written here about cochineal on a number of occasions, most recently regarding an extraordinary red garment dyed with cochineal that is housed in the Murakami Kaizoku Museum on the island of Oshima in Japan’s Inland Sea. Cochineal (Dactylopius coccus), like Kermes, is a parasitic scale insect but it lives on cacti in the genus Opuntia, which comprises plants native to the Southwest United States and Central and South America. Cochineal insects are the source of carminic acid from which the dye carmine is made, and, yes, there was a traditional Carmine Lake pigment, now in most applications superseded by synthetic equivalents (although cochineal is still used as a food coloring).
So, Prussian Blue was a mistake. Dippel knew immediately when he heard about the new blue pigment that his potash must have been the cause. According to my Internet sleuthing, nitrogen from blood in the potash Dippel supplied formed cyanide compounds such as potassium cyanide when heated. Mixed with the iron sulfate Diesbach used in making his Florentine Lake, the potassium cyanide participated in a reaction leading to formation of potassium ferrocyanide, which, in the presence of iron sulfate, creates iron ferrocyanide, or Prussian Blue. Apparently Diesbach and Dippel had a monopoly on the new pigment until others figured out how to replicate it. The formula was first published in 1724 (although Prussian Blue had already been replicated by others somewhat earlier) so Diesbach and Dippel controlled the Prussian Blue market for about a decade and a half after the pigment’s discovery. Prussian Blue is often cited as the first modern synthetic pigment.
Something I read noted that genuine Prussian Blue, or pure iron ferrocyanide, is not entirely stable, mainly because it tends to degrade in alkaline environments and with exposure to light (more about the latter below). This led me to material on pigments published on line by Jackson’s Art. It seems that genuine Prussian Blue is acceptably stable when used in making watercolors or oil paints. Acrylic media, however, are alkaline, so true Prussian Blue acrylic paints are rare (Australian paint maker Derivan is about the only paint maker that sells a true Prussian Blue acrylic paint, noting that it is not stable). Acrylics sold as Prussian Blue are almost always labeled “Prussian Blue Hue” and made by mixing other pigments, often Pthalo Blue, Dioxazine Violet, and Carbon Black. That is, to the eye, the hue is replicated but the pigments responsible for the hue are a substitute. True Prussian Blue in acrylic media has mostly been replaced by more stable equivalents – superseded like the Florentine Lake that Prussian Blue stemmed from.
At this point, I thought I had my answer. To say that boiled horse’s hoofs were a source of Prussian Blue seemed inaccurate. Rather than hoofs, per se, it seemed likely that blood in tissues connected to hoofs removed from horse carcasses provided the nitrogen for the Prussian Blue reaction rather than keratin itself. Unless…keratin is a good source of nitrogen. Apparently, it is. I suspect that both blood and keratin were used as a nitrogen source in making traditional Prussian Blue and it probably came from multiple animal sources.
I could have stopped there, but there was more on my mind. Again, as an artist, I’m quite familiar with Color Index Generic Name Codes (usually referred to simply as pigment codes). I knew that pigments used in artist’s colors can be reliably identified by their pigment codes, which operate much like Latin names in biology. They are standardized and recognized globally, eliminating confusion caused by overlapping common names, alternate names, traditional names no longer used, and language differences. According to Jackson’s Art, the Colour Index International is a database of pigments and dyes published by the Society of Dyers and Colourists and the American Association of Textile Chemists and Colorists. The Index was created just over 100 years ago, in 1924.
Prussian Blue has the code PB27. It would be natural to assume that “PB” here stands for “Prussian Blue,” but it doesn’t. It’s a misleading coincidence. The color codes for pigments (as opposed to dyes) all begin with the letter “P.” Blue pigments are all coded “PB” for “pigment blue”. Yellow pigment codes, by the same logic, are all “PY” followed by a number. Red pigments are all coded “PR” followed by a number, etc. The number indicates the order in which each pigment was added to the official database. If Prussian Blue was the first modern synthetic pigment, you might expect it to be “PB1,” but, because iron ferrocyanide was the 27th blue pigment added to the database, its code is PB27. While the numbers tell us little more than how recently a pigment was recognized in this particular coding system, the codes have the virtue of telling artists precisely what compounds are in their paints. A tube of acrylic Prussian Blue marked PB27 is likely to be unstable. A tube marked PB15, PV23, PBk7 is the stable Prussian Blue mimic described above made from Pthalo Blue, Dioxazine Violet, and Carbon Black. Most artists want their work to last, so, that’s a good thing to know. On another tangent altogether, I’ve noticed independently that mixing paints labeled “Ultramarine” and “Viridian” produces a hue very close to what we call Prussian Blue. I will refrain from going into the history of the colors Ultramarine and Viridian here – although I’m tempted to.
Having said all of the above, I hadn’t realized that, because the pigments identified using these codes are sometimes natural pigments (derived directly from earth or clay, for example), they are not always the same hue. A case in point is PBr7 (brown pigment number seven). As it’s a manganese-containing, naturally occurring iron oxide, the hues it produces in paints can vary somewhat. In addition, processing can change the hues expressed by a single pigment dramatically. PBr7 identifies the iron oxide pigment that is expressed as both Raw Umber and Burnt Umber – two obviously different hues. So, while the codes always identify the pigment, they don’t necessarily always exactly identify the hue we see – although they pretty much do in the case of synthetic pigments, which are pretty much uniform.
Then, not wanting to spread misinformation, I started wondering if I remembered correctly that the blue we call Prussian Blue is the same blue formed in the cyanotype process and in traditional blueprints. Upon more research online, I was relieved to confirm that these processes do create iron ferrocyanide, or Prussian Blue.
One odd fact about the cyanotype process is that cyanotypes are known to fade over time with exposure to light. That’s not unusual. Fading is a common phenomenon. What’s weird about cyanotypes is that the depth of color in a faded cyanotype recovers in darkness. That I knew, but why? And is the same phenomenon observed in, for example, oil paintings that use genuine Prussian Blue, or in the Japanese print style known as aizuri-e in which the image is produced almost entirely in shades of Prussian Blue? Prussian Blue when newly available in Japan became known as ベロ藍 (bero-ai) with “bero” derived from “Berlin” and “ai,” meaning indigo (although today you’re more likely to hear it called プルシアンブルー (purushian burū) in Japanese, directly from the English), reflecting its first synthesis in Berlin. The plant-derived pigment we call indigo, however, is not the same as Prussian Blue. Several different plant species have been used as the source of natural indigo over the centuries, notably true indigo (Indigofera tinctoria) and other species in the genus Indigofera, but also the unrelated plants woad (Isatis tinctoria) and Japanese Indigo (Polygonum tinctorium). All of these produce natural indigo, which is C16H10N2O2, unrelated chemically to iron ferrocyanide. It’s only the similar hue that links indigo to Prussian Blue.
But I was wondering about the recovery of cyanotypes in darkness. What’s going on here? The now-ubiquitous AI bot that answers Google queries tells me that with exposure to visible light, Prussian blue undergoes photochemical reduction – resulting in a fading caused by the gradual conversion of ferric ions in Prussian Blue to ferrous ions, which, apparently appear not blue but white. A reduction reaction is the opposite of an oxidation reaction. In reduction, a molecule loses oxygen. In oxidation, the reverse happens. Happily, in the presence of oxygen in darkness, this reaction is reversible, at least in part, with some ferrous ions oxidizing back to ferric ions, restoring the blue hue. As cyanotypes (and aizuri-e) generally are not given a surface treatment, this restorative reaction likely occurs fairly easily, but I’m guessing that oil paintings using Prussian Blue are less able to undergo the reversal reaction because they are often finished with a varnish that blocks oxygen – but perhaps the varnish also inhibits the fading?
Maybe that’s enough for one day. I started writing at about 10:00AM. It’s now going on 4:00PM. I need food. Time to climb back into the open air.
Then, unable to resist, I asked Google if faded aizuri-e prints can be restored in darkness. The bot says no – which leaves me puzzled. I don’t understand why Prussian Blue in a cyanotype would behave one way while the same pigment in an aizuri-e would behave differently. Maybe the bot is wrong. In any case, I need to eat something.
Except that, asked the same question about Prussian Blue in oil paintings, the bot suggests that the pigment in oil paintings can, in fact, undergo the reversion reaction (although the reaction and the initial fading in Prussian Blue oil paint, if any, can be affected by admixtures of other pigments). So why not in the case of aizuri-e? And what about the blues in blue and white Japanese and other ceramics? Is that Prussian Blue?
I’m going now. Really.
[Edit: It occurred to me shortly after posting this that blue and white ceramics in Japan – for example, the famous ceramics from Imari in southern Japan – predate Prussian Blue. Imari became a center of ceramics production in the early 1600s. Prussian Blue was first synthesized about 100 years later and it didn't become widely used in Japan for another 100 years. The heyday of aizuri-e was the 1820s to 1840s. Traditional blue and white ceramics in Japan used (and still use) cobalt-based blue underglazes, as in China. Cobalt blue ceramics first appeared in China during the Tang dynasty (618-907).]
[Additional info: Two days later, still poking around, I came across this interesting tidbit: The term "lake" that I discussed here has the same root ("lac") as the "lac" in "lacquer" and "shellac." This "lac" refers to yet another scale insect, in this case Kerria lacca (formerly known as Laccifer lacca, Coccus lacca, or Tachardia lacca – aparently reclassified a number of times or split). I had known that "lacquer" and "shellac" were related to insect secretions but I hadn't made the connection with "lake" meaning a dye-derived pigment. Kerria lacca, like the cochineal insect and Kermes vermilio, has historically been the source of a red dye. That dye, according to some articles I've read was the first or one of the first to be precipitated as a pigment, or lake, and, eventually, the term "lake" came to be used for any such pigment.]