The New Cyanotype (1995)

Every other formula in this book is a list of what goes into the bottle. This one is a list of what goes into the beakers. Between the mixing and the make-up, about 15 g of a green solid crystallises out of Mike Ware’s New Cyanotype and is filtered off and thrown away, and the sensitiser that remains is a different substance from the one the weights below would suggest. That discarded solid is the whole invention: it is potassium iron(III) oxalate, and getting rid of it is how a 1994 chemist turned Herschel’s 1842 sensitiser into something that prints in two minutes instead of half an hour, keeps for years instead of growing mould in a fortnight, and holds a tonal scale long enough to print the same negative you would send to platinum.

Ingredient	Quantity	Form the source specifies
Ammonium iron(III) oxalate	30 g	the trihydrate, which Ware writes as (NH4)3Fe(C2O4)3.3H2O with a formula weight of 428.07; General Purpose Reagent grade, about 98 to 99 per cent, is adequate. Dissolved in exactly 30 cc of purified water at about 50 C, with any small residue of colourless crystals ignored
Potassium ferricyanide	10 g	dissolved in exactly 20 cc of purified water at about 70 C and added hot to the oxalate solution. Most of its potassium leaves again as potassium iron(III) oxalate before the sensitiser is made up, so what remains in the bottle is substantially ammonium ferricyanide
Water	to make 100 mL	Purified water - distilled, de-ionised or pharmaceutical. Only 50 cc of it is used in the preparation itself, 20 cc for the ferricyanide and 30 cc for the oxalate; the balance is added at the end to bring the filtrate, about 60 to 62 cc, up to the 100 cc mark. This is the one entry in this formulary whose ingredient list is a list of starting materials rather than of what is in the bottle, and the page says so wherever it matters.
Added to the working bath, not to the stock
Citric acid (anhydrous)	2 g per 100 mL of working solution	optional, and added only to the portion of sensitiser about to be coated, never to the stock bottle, where Ware states it shortens the shelf life. It is his remedy for chemical fogging - the "greening" of a coating caused by impurities in the paper - and it lowers contrast, so that a negative of density range 2.6 can be accommodated — Ware's chemicals list gives 2 g against the 100 cc of sensitiser; 7.3.7 states the same dose as one drop, 0.05 cc, of a 40 per cent w/v citric acid solution per cc of sensitiser, which works out at about 2 per cent w/v. Either the anhydrous acid or the monohydrate will do

☠One solid that Ware prints and this page does not

Ware’s own chemicals list carries a fourth item: 0.1 g of ammonium dichromate per 100 cc, added to the oxalate solution as a preservative. It is not in the table above, and the omission is a decision of this course rather than a gap in the evidence.

Why. Ammonium dichromate is a chromium(VI) compound, and the course’s chromium policy rules that chromium(VI) is never used, at any level, anywhere — a rule that rests on a notified classification carrying H340 and H350 and on a UK workplace exposure limit of 0.01 mg/m³ as chromium with Carc, Sen and a biological monitoring value attached.

Why this is not the course quietly editing a formula. The originator marks it optional three times, in three publications. The chemicals table of his 2009 workshop notes prints it as “Ammonium dichromate (NH₄)₂Cr₂O₇ (OPTIONAL) 0.1 g”. Cyanomicon’s preparation says the small amount “may be omitted to satisfy prevailing safety regulations, but the contrast and storage-life of the sensitizer will be diminished, and some Prussian blue may be formed as a sediment”. And his section on environmental issues says it plainly: dichromate “is recommended in very small amount in my New Cyanotype as a preservative — but can be entirely omitted”.

What is given up, in his words and not in ours. The keeping of the stock bottle falls from several years to somewhere between a few months and a year; the contrast falls, because Ware records that dichromate “will also tend to increase the contrast of the sensitizer significantly”; and a sediment of Prussian blue may form in the bottle, with fogged highlights on the paper as its consequence. Storage and Troubleshooting below say what to do about each of those, and Variants sets out the version with the dichromate in as the formula Ware actually published.

The 0.1 g figure is printed here once, because a reader has to be able to identify the formula they will meet in Ware’s own books and in a supplier’s kit. It is not an instruction, and no procedure for using it appears anywhere on this page.

Purpose

To make a negative-working print in Prussian blue on plain paper, with roughly a tenth of the exposure the classic cyanotype needs, from a negative of long density range, and with a degree of contrast control at the development stage. The two chemicals that carry the image are the same two elements Herschel used in 1842 — iron and hexacyanoferrate — and everything Ware changed is about the counter-ions, the ligands, and where the potassium went.

It is a print-out process: the image is made by light, not by a developer. The bath the print goes into afterwards is called development by convention, but it does not amplify anything. It releases iron(II) from its ligands so that the blue can finish forming, and it washes out everything light did not use.

Recommended uses

• Pictorial printing from a long-scale negative. This is the use it was designed for. With an exposure scale of about 2.2, it takes the same negative as salted paper, platino-palladiotype, argyrotype and chrysotype. Ware’s point is worth stating in one sentence: one negative can serve all of them, and the negative’s contrast is adapted to the process rather than the other way round.
• Where the maximum black matters. The maximum density verges on black, because the ammonium form of Prussian blue is far less readily peptized out of the paper than the potassium form the classic process makes.
• Where exposure time is the constraint — sun-printing three-dimensional objects, weak light sources, teaching sessions where a class cannot stand around a frame for half an hour.
• Under a cyanotype-over-platinum composite. Ware notes that the New process has about the same exposure scale as platino-palladiotype, so a single negative registers for both printings; the classic sensitiser cannot render that density range at all.
• On fabric, where Ware records that the oxalate mordants better than the citrate.

When another formula is preferable

• When you have never coated a sheet of paper before. Use the classic cyanotype sensitiser. It is two solids, two bottles, no heat, no filtration, no crystallisation, and Ware — who devised the formula on this page — writes that the classic chemicals “are not dangerous and can be safely handled by children under supervision”. Nothing on this page is that.
• When your negatives are normal. A negative made for a grade 2 or 3 silver-gelatin paper has a density range of about 0.9 to 1.2. The classic process is built for exactly that. Printing it on this sensitiser will give you a flat, muddy image using a fifth of the tonal capacity you paid for.
• When you cannot heat solutions, weigh to 0.1 g, or filter. The preparation needs a hotplate or a basin of very hot water, a thermometer, a measuring cylinder, a funnel and filter paper, and an overnight wait. Ware is candid that competent practitioners find it “somewhat challenging” and that crystallisation under ill-defined conditions gives a variable yield, so that batch-to-batch concentration varies and the most quantitatively demanding digital-negative workflows need recalibrating between batches.
• When you want contrast control without acid. The Simple cyanotype of 2019 and Mike’s cyanotype are Ware’s own later answers to the same problem, built from iron(III) nitrate and citric acid. His summary of the five formulations sets the Simple one — “easy to make; cheap; contrast control; non-toxic”, which is not a hazard classification and is not one this course adopts on his word — against this one’s “faster; finer; higher Dmax; toxic; chemically tricky”.
• When the formula you meant was the one with the dichromate in. That is Ware’s improved Classic cyanotype, a different formula on a different iron salt, recorded here at Level D with no procedure at all.

Mixing

The preparation is a fractional crystallisation, and the point of it is to throw something away. It is set out here as chemistry, in the order Ware gives it, with the reason for each step; the numbered instructions, the vessels and the timings are his.

Work under dim tungsten light — 40 to 60 watts — and not under fluorescent tubes or daylight. The sensitiser is only sensitive to the near ultraviolet and the deep blue, which is why an ordinary incandescent or warm LED lamp is safe, but a fluorescent tube and a window are not.

1. 10 g of potassium ferricyanide in exactly 20 cc of purified water, heated to about 70 °C and stirred until dissolved. Kept hot. Twenty millilitres is a very small volume for ten grams: this is a near-saturated solution and it will not stay in solution when it cools, which is the point.
2. 30 g of ammonium iron(III) oxalate in exactly 30 cc of purified water, heated to about 50 °C and stirred until the green solid has gone. A small residue of colourless crystals is ignored — it is impurity, not undissolved salt.
3. The hot ferricyanide solution is added to the oxalate solution, stirred well, and left in a dark place to cool to room temperature and crystallise. Ware allows at least two hours for this quantity and says it may be better left overnight, especially at a larger scale.
4. The green crystals are separated from the green liquid by decanting and filtration — a conical funnel, a Whatman grade 1 paper, running into a 100 cc measuring cylinder. About 15 g of crystals should be left in the paper and about 60 cc of filtrate in the cylinder. The crystals are potassium iron(III) oxalate and they are discarded.
5. The filtrate is made up to 100 cc with purified water, mixed, and put into a stoppered, labelled and dated brown bottle.

Why it is not one beaker. If you simply mixed the two salts at the concentrations a pictorial sensitiser needs, potassium iron(III) oxalate would crystallise out anyway — in the bottle, in the syringe, and worst of all in the paper, where Ware describes it as producing “quite pretty, but totally unwanted fern-like patterns” and small white spots typical of surface crystals. The preparation does not avoid the crystallisation. It provokes it deliberately, hot and concentrated, where it can be filtered off.

Why there is no mixing order in the data above. The two solids never meet in the same vessel as solids. They are dissolved separately, at two different temperatures, in two measured volumes of water, and the order that matters is the order of addition — hot ferricyanide into oxalate — not an order of dissolving. A mixingOrder listing both would state something the source does not.

🧮Where 30 and 10 come from, and what is actually in the bottle

Ware derives the two weights rather than choosing them, and the derivation is worth following because it also tells you what the finished sensitiser contains.

Step one: the target. Prussian blue as this process makes it is taken to need iron(III) and hexacyanoferrate in a 1 : 1 molar ratio for the photochemistry. Removing the potassium needs a further 1 : 1, because potassium iron(III) oxalate takes one iron(III) with every three potassiums, and there are three potassiums in every ferricyanide. Add the two together and the oxalate must be in 7 : 3 excess over the ferricyanide — four parts for the picture, three parts to carry the potassium away.

Step two: into weights. The relative molecular masses are 428.07 for the oxalate trihydrate and 329.26 for the ferricyanide, so

7 × 428.07 : 3 × 329.26 = 2996.49 : 987.78 = 3.034 : 1

Ware's weight ratio

which rounds to 3 : 1 — hence 30 g and 10 g.

Step three: what is left. Working it through in moles, 30 g of the oxalate is 0.0701 mol and 10 g of the ferricyanide is 0.0304 mol. If every potassium leaves as potassium iron(III) oxalate trihydrate, that consumes 0.0304 mol of the oxalate and weighs 0.0304 × 491.2 = 14.9 g, against Ware’s stated yield of about 15 g. What remains, made up to 100 cc, is 0.0397 mol of iron(III) oxalate and 0.0304 mol of ferricyanide:

0.397 M ammonium iron(III) oxalate · 0.304 M ammonium ferricyanide

The finished sensitiser

Ware states the sensitiser as 0.4 molar ammonium ferric oxalate and 0.3 molar ammonium ferricyanide. The arithmetic and the discarded 15 g both land on his figures, which is about as good a confirmation of a published preparation as a reader can get without a balance.

The arithmetic in this box is the course’s; the 7 : 3 ratio, the 3.034 : 1 weight ratio, the 0.4 and 0.3 molar figures and the 15 g yield are Ware’s. Two caveats are his as well. The separation of potassium from ammonium is not complete — some potassium stays in solution and a little ammonium goes into the crystals — so 0.4 and 0.3 molar are nominal. And Ware notes that the true Prussian blue of a cyanotype may be closer to the “insoluble” Fe₄[Fe(CN)₆]₃, a 4 : 3 ratio of iron(III) to iron(II), than to the 1 : 1 assumed above; he says the relative proportions are not critical, and it is a pleasing accident that 4 : 3 is exactly what the photochemistry is left with after the precipitation has taken its three parts.

🧮How this compares with the classic sensitiser, molecule for molecule

Concentrations quoted in per cent w/v hide the comparison, because the two processes use different iron salts of very different molar mass. In moles per litre, the classic sensitiser as mixed — 10 per cent ammonium iron(III) citrate, 5 per cent potassium ferricyanide — works out at about 0.35 M iron(III) and 0.15 M ferricyanide, taking Ware’s figure of about 290 g per mole of iron for the green citrate. This page’s sensitiser is 0.40 M and 0.30 M.

So the iron is much the same and the ferricyanide is doubled. The ratio of iron(III) to ferricyanide falls from about 2.3 : 1 in the classic mixture to about 1.3 : 1 here — which is essentially the 4 : 3 of Prussian blue itself.

That single number is Ware’s sixth remedy stated arithmetically. The classic sensitiser carries more than twice the iron it can use; the excess iron(II) made in the shadows diffuses sideways into the highlights and prints there, which is the defect called bleeding. Mix the two in chemically equivalent amounts and there is no excess to migrate. The comparison arithmetic is the course’s; the diagnosis of bleeding and its remedy are Ware’s.

Two things are added to the sensitiser and never to the stock bottle.

• Citric acid to about 2 per cent, where the paper fogs. One drop (0.05 cc) of a 40 per cent w/v solution per cc of sensitiser, mixed in just before coating. Ware is explicit that it must not go into the stock, because it shortens the shelf life.
• Tween 20 to about 0.1 to 0.25 per cent, where a hard-sized paper coats badly. It is a non-ionic surfactant, it is not in the table above because this course has no chemical encyclopaedia entry for it, and Ware’s own position is that with this sensitiser it is “usually quite unnecessary” — the oxalate solution penetrates cellulose fibres far better than the citrate does. He adds two warnings: it may interact unfavourably with gelatin-sized papers, and the stock solution should be replaced if it shows a deposit.

Coat about 1.5 cc per 10 by 8 inch area, four to six passes of a glass rod, on an unbuffered paper. Blot off any excess, which would otherwise crystallise on the surface and damage a negative. Dry for one to two hours at room temperature in the dark, or with warm air at about 40 °C for five to ten minutes.

Behaviour

It prints out, and the print-out is a reliable preview. The classic process prints out too, but so much of its image washes away that what you see in the frame is not what you get. Here there is very little loss of image substance during processing, so the appearance in the frame — high values green, mid-tones blue, deepest shadows reversed to a pale grey-blue — is a good indication of the final print. That is the practical difference between a process you can judge by inspection and one you have to test-strip.

It solarises on purpose, and the solarisation is the exposure meter. The reversal in the shadows is not a fault. A hinged-back contact frame lets you lift a corner and look for it.

The reversed shadows come back. During the acid bath they begin to regain density quickly — nitric acid particularly assists this — and they recover fully over about 24 hours of drying in air. If you cannot wait, half a minute in 0.3 per cent hydrogen peroxide does it at once and, Ware states, makes no ultimate difference to the final densities.

Contrast is set in the first bath. This is the behaviour that most distinguishes it from the classic process, which has no development stage to control. Water alone gives the longest scale and the weakest black; 1 per cent nitric acid gives the shortest scale and the deepest black; anything between is available by dilution.

The coating is a paper test. Ware calls this sensitiser “a very sensitive test of paper quality”. A coated sheet left for some hours in the dark at ordinary humidity should stay light yellow; a drift towards green, or worse blue, is chemical fog, and it condemns the paper rather than the batch.

Heat-drying and “darktime” both move the scale, and Ware advises against both. His step-tablet tests give the numbers: heat-drying the coating shortens the exposure scale by about a stop, from 2.3 to 2.0, and may slightly coarsen the texture; leaving the exposed print in the dark for 19 hours before processing lengthens it by about a stop, from 2.3 to 2.6, but weakens the maximum density, flattens the shadows and causes edge fog at the tideline. Doing both returns the scale to where it started with the maximum density lost, “so it is best to do neither”.

⁇Does heat-drying change contrast?

The disagreement is between two printings by the same author. The 2009 workshop notes say that “the dryness of a cyanotype paper does not appear to influence image colour or contrast”. Cyanomicon 7.3.6 says heat-drying “does not appear to influence image colour but it may increase contrast slightly”, and 7.3.12 reports the step-tablet test that measures the change at about one stop of exposure scale.

What this course uses and why. The later statement, because it is the one backed by a published step-tablet comparison in duplicate rather than by impression, and because it is Ware revising his own earlier account after making the measurement. A reader working from the 2009 handout should know that its sentence has been superseded by its own author.

Image characteristics

Colour. An “ammonium Prussian blue”, which Ware describes as of very good colour and finer than the classic process gives. The lattice of Prussian blue accommodates a large cation in its cavities, and which cation sits there changes both the hue and — much more importantly — the ease with which the pigment disperses in water.

Maximum density verging on black. Not because more pigment is made, but because more of it stays in the paper. This is the single largest practical difference from the classic process.

A tonal scale of about 2.2, against about 0.9 for the classic sensitiser. Ware’s exposure-scale figures are the ones to carry away, because they are what decides which negative fits which process. An exposure scale of 0.9 is about three stops — comparable, he notes, with a grade 3 silver-gelatin paper. An exposure scale of 2.2 is about seven stops.

Delicate gradation in the high values. The classic process loses the high values because the image substance in the thinnest deposits is the first to be peptized away, which truncates the scale at the light end and confers what Ware calls an “artificially high contrast”. Retain the pigment and the high values come back.

No staining, no image tone to speak of, and no grain in the silver sense. The image is a pigment in the fibres of the paper. Its texture is the paper’s texture, which is why the choice of sheet does more visible work here than in any silver process.

The mechanism

Four reactions, in order, and the third of them is the one the acid bath exists for.

One: light reduces the iron. The trisoxalatoferrate(III) anion absorbs in the near ultraviolet, an electron moves from an oxalate ligand to the iron, and the iron(III) becomes iron(II) while the ligand decarboxylates. Ware gives the overall stoichiometry:

2 [Fe(C₂O₄)₃]³⁻ → 2 [Fe(C₂O₄)₂]²⁻ + C₂O₄²⁻ + 2 CO₂

Photoreduction of trisoxalatoferrate(III), overall

The mechanism first proposed by Hatchard and Parker goes in two steps. Light makes a radical anion inside the complex, and that radical then reduces a second, unexcited complex — so one photon can deliver two iron(II):

[Fe(C₂O₄)₃]³⁻ + C₂O₄⁻ → [Fe(C₂O₄)₂]²⁻ + C₂O₄²⁻ + 2 CO₂

The second electron: the oxalate radical anion reduces a second, unexcited complex

Two: the iron(II) reduces the ferricyanide. This is not the reaction most accounts give, and Ware’s evidence for it is a simple experiment: irradiate a trisoxalatoferrate solution, then add it to a ferricyanide solution, and little or no Prussian blue forms. The iron(II) is still wearing its oxalate ligands and cannot enter the Prussian blue lattice while it does. What it can do is hand an electron across:

[Fe(C₂O₄)₂]²⁻ + [Fe(CN)₆]³⁻ → [Fe(C₂O₄)₂]⁻ + [Fe(CN)₆]⁴⁻

Simple electron transfer, favoured by 0.356 V against 0.02 V

Three: acid strips the ligands and releases the iron(II). Iron(II) does not hold oxalate strongly, and hydrogen ions compete for the oxalate to make undissociated oxalic acid. Ware gives the two constants and multiplies them: the overall equilibrium

[Fe(C₂O₄)₂]²⁻ + 4 H⁺ ⇌ Fe²⁺ + 2 H₂C₂O₄

The equilibrium the development bath drives

has K = 4.2 × 10⁴ and therefore lies well to the right. That is why the first bath is dilute mineral acid and why it deepens the blacks. It is not developing anything in the photographic sense; it is dissolving the ligands off the photoproduct so that free iron(II) is available to build pigment.

Four: the pigment forms. Free iron(II) is then competed for by both oxidation states of the hexacyanoferrate present:

Fe²⁺ + [Fe(CN)₆]³⁻ → Fe[Fe(CN)₆]⁻

Prussian blue

Fe²⁺ + [Fe(CN)₆]⁴⁻ → Fe[Fe(CN)₆]²⁻

Prussian white

Where exposure has been light, ferricyanide predominates and the product is blue. Where exposure has been heavy, the ferricyanide has been largely reduced and the product is the colourless Prussian white — which is the beginning of the explanation for solarisation, below.

Deeper: why the oxalate is four to eight times faster than the citrate

Two numbers account for most of it, and both are quantum yields Ware collates from the photochemical literature.

The quantum yield of a photoreaction is the number of molecules changed per photon absorbed. For trisoxalatoferrate(III) in aqueous solution it is about 1.2 between 250 and 420 nm — above one, because of the two-electron mechanism above. For the citrate system it is about 0.4. For hexacyanoferrate(III) photolysing on its own it is about 0.01, which is why Herschel’s proto-cyanotype using ferricyanide alone needs hours rather than minutes.

A ratio of 1.2 to 0.4 is three, and Ware’s stated advantage is four to eight. The rest comes from the second effect, and it is one every cyanotype worker should know about.

The ferricyanide is an internal filter. It absorbs strongly in exactly the region where the iron salt has to be exposed, so a large part of the light that enters the coating is absorbed by a molecule that does nothing useful with it. Ware calculates the fraction of incident light actually absorbed by the photoactive component in a typical cyanotype coating: at 365 nm it is 0.65, but at 420 nm it has collapsed to 0.03, because the ferricyanide absorption rises as the trisoxalatoferrate absorption falls. Two consequences follow. The process is far more efficient under a 365 nm UVA source than under a source weighted towards violet, and doubling the ferricyanide — as this formula does — buys pigment at some cost in speed, which is part of why the practical gain is not the full ratio of the quantum yields.

Deeper: why the potassium had to go

Prussian blue is not one compound. It is a family of cubic lattices with variable water and variable cations in the cavities, running from the so-called “soluble” KFe[Fe(CN)₆]·5H₂O to the “insoluble” Fe₄[Fe(CN)₆]₃·15H₂O. Ware is careful about the word: all of them are highly insoluble, and the “solubility” of the first is an illusion produced by peptization — the pigment disperses as colloidal particles fine enough to make a blue suspension that looks like a solution.

The cyanotype makes the peptizable form. In the classic process that is a disaster in slow motion: the wash that removes the unexposed sensitiser also carries a significant fraction of the image down the sink, taking the high values with it and leaving an image that has to be grossly overexposed to survive. Substitute ammonium for potassium in those lattice cavities and the resulting blue is, Ware states, “more resistant to peptization and alkalies”.

That is the entire reason for the fractional crystallisation. The obvious way to get an ammonium blue would be to use ammonium ferricyanide, and Ware says so — but it is not commercially available, is described as a rare chemical, and is troublesome to make pure. So the potassium is removed from the mixture instead, using the low solubility of potassium iron(III) oxalate as the tool, and the ammonium ferricyanide is generated in situ.

Deeper: solarisation, and the two explanations that compete

The shadows of a well-exposed cyanotype reverse to a pale grey-blue. There are two published mechanisms and Ware weighs them.

Suzuki’s: prolonged exposure photoreduces the ferricyanide directly to ferrocyanide, which then makes colourless Prussian white with the iron(II) photoproduct. Ware’s objection is quantitative — the quantum yield for that photoreduction is about 0.01, far too low to account for what is actually seen, and it offers no explanation of why blue already formed is observed to disappear.

Murray’s: excess iron(II) oxalato complex in the heavily exposed areas reduces the Prussian blue that is already there to Prussian white. The redox potentials support it — +0.46 V for the blue/white couple against +0.02 V for the iron complex — and, importantly, the iron(II) does not have to shed its ligands to do this, so it can happen in the dry coating.

Ware’s own test, and his addition to both. He irradiated a solution of ammonium trisoxalatoferrate, then immersed a finished cyanotype step tablet in it: the strip bleached, and a control in the same unirradiated solution did not. So the photoproduct does reduce Prussian blue. But when he repeated the experiment with the strip in the solution under the lamp, the bleaching was much faster and more complete even at high densities — from which he infers that Prussian blue itself becomes photochemically activated and easier to reduce. He presents this as his own conviction and as the connection between solarisation during printing and the reversible light-fading of finished prints, which is a hypothesis rather than settled chemistry, and this page marks it as such.

The practical upshot is the same either way: the reversal is a reduction, not a loss, and air or peroxide oxidises it straight back.

Function of every ingredient

Ammonium iron(III) oxalate, 30 g — the light-sensitive salt, and the potassium scavenger. It is triammonium trisoxalatoferrate(III) trihydrate, bright green crystals of a known and reproducible structure. Its job in the exposure is the only genuinely photochemical job in the formula: iron(III) is reduced to iron(II) in proportion to the light that falls, at the expense of an oxalate ligand which decarboxylates and leaves as carbon dioxide.

Ware chose it over ammonium iron(III) citrate for five reasons he lists explicitly, and each has a visible consequence on the print. It is a well-characterised, consistent, inexpensive crystalline compound, where the citrate is notoriously variable in composition from supplier to supplier — so a remote reader can be confident of using the same chemical he did. It does not support mould, where a citrate solution is an excellent nutrient medium and grows a furry mat within a week or two. Its solution penetrates cellulose fibres far better, so the image sits inside the fibres where the wash cannot reach it. It is four to eight times more light-sensitive. And it is non-deliquescent, so the coated sheet does not go tacky in humid air and does not stick to a negative.

More or less of it. The 30 g is not a free parameter: it is fixed at three times the ferricyanide by the 7 : 3 molar requirement, four parts for the photochemistry and three to carry the potassium out. Reduce it and there is not enough oxalate to precipitate all the potassium, so potassium is left in solution to crystallise later on the paper. Increase it and the excess iron(III) has no ferricyanide to pair with, which restores the very defect — highlight bleeding from migrating iron(II) — that the balanced ratio was chosen to remove. Dilute the finished sensitiser instead, as the Variants section describes, if a faster and weaker coating is what you want.

Potassium ferricyanide, 10 g — the precipitant, the internal filter, and the source of the problem the preparation solves. It supplies the hexacyanoferrate that becomes half of the Prussian blue lattice, and it does so without being light-sensitive in any useful degree: its own quantum yield is about 0.01, so it contributes almost nothing to the image formation and everything to the image substance.

It has two further roles, one useful and one not. It is the internal filter described above, absorbing strongly at the violet end of the printing band; that costs speed but it also flattens the spectral response and is part of why the process tolerates a wide range of ultraviolet sources. And it is the sole source of the potassium that this entire preparation exists to remove — 0.0304 mol of ferricyanide brings 0.0912 mol of potassium ions with it, and all of them have to be crystallised out as potassium iron(III) oxalate before the sensitiser is fit to coat.

More or less of it. More ferricyanide deepens the maximum density and slows the print, both because of the internal filter and because there is more pigment to make; it also brings more potassium, which the oxalate excess would then be too small to precipitate. Less ferricyanide is what the commercial blueprint papers did — Ware records that they had to keep it at 5 per cent or below to avoid the crystallisation at all, “so they only yield a maximum density that is too low for pictorial purposes”. That sentence is the best available argument for going to the trouble of the crystallisation: the easy way of avoiding the problem costs you the blacks.

Citric acid, optional, to about 2 per cent of the coating solution — the anti-fog and the contrast softener. It is not part of the sensitiser as stored. It is mixed into the portion about to be coated, at one drop of a 40 per cent w/v solution per cc, and Ware states plainly that putting it in the stock bottle shortens the shelf life.

What it does is hold the coating acid. Paper impurities and alkaline additives push the coating towards the conditions in which iron(III) hydrolyses and Prussian blue forms in the dark — the “greening” that turns a lemon-yellow coating green or blue before it has been exposed at all. Acid suppresses that. The photographic consequence beyond the fog is a reduction in contrast: Ware’s web article states that adding citric acid lessens the contrast of the sensitiser so far that a negative density range of 2.6 can be accommodated, which is a very long negative indeed.

More or less of it. Below about 2 per cent the protection against a marginal paper becomes unreliable; above it, the softening of the gradation begins to cost you the black you came to this formula for, and there is no published figure for where that begins. Ware also uses citric acid at 1 to 5 per cent w/v as a last-resort development bath, which is a different use of the same substance at ten to twenty times the strength.

Water, to make 100 cc. Purified — distilled, de-ionised or pharmaceutical. Only 50 cc of it is measured into the beakers; the balance is what brings the filtrate up to the mark and therefore what sets every concentration in the bottle. It has to be measured in a cylinder, not estimated, because the yield of the crystallisation varies and the make-up volume is the only thing holding the sensitiser to a defined strength.

What is not in the list. Ware’s own formula carries 0.1 g of ammonium dichromate, which this course omits under its chromium policy, and 0.25 cc of Tween 20, which is absent because the course has no encyclopaedia entry for it and Ware calls it usually unnecessary with this sensitiser. Neither omission is silent: the dichromate is discussed in the callout at the head of this page and again under Variants, Safety, Storage and Troubleshooting, and the Tween is described under Mixing. The dichromate has no photochemical function whatever — it is an oxidising preservative that keeps stray reductant from making Prussian blue in the bottle, and its photographic side effect, which Ware records, is a significant increase in contrast.

Interactions

Oxalate with ferricyanide: the interaction the formula is built around. Mixed cold and dilute they coexist; mixed at the concentrations a pictorial sensitiser needs, they throw potassium iron(III) oxalate out of solution. Every practical property of this formula follows from the decision to make that happen deliberately, hot, in a beaker, rather than accidentally, later, in the paper.

Iron(III) with acid: the pH window. The trisoxalatoferrate anion changes speciation with acidity — around pH 2 the emerald green solution becomes light yellow as the diaquabis(oxalato) complex takes over — and Ware notes that this species is still light-sensitive with approximately the same quantum yield. Acid is therefore not something the sensitiser merely tolerates; it is the condition the photochemistry is happy in, which is why citric acid can be added to the coating and why a mineral acid bath does no harm to the unexposed sensitiser it is washing out.

Iron(II) with hydrogen ions: the development. Set out under The mechanism. The equilibrium constant of 4.2 × 10⁴ is why 1 per cent nitric acid gives a deeper black than water does, and the sliding scale between them is the only contrast control the process has.

Prussian blue with alkali: the destruction. Alkali hydrolyses the pigment to weakly coloured iron compounds. This is why the paper must be unbuffered, why the wash water must not be alkaline, and why mounting board above about pH 9 will slowly eat the print.

Prussian blue with calcium: the same problem arriving in the tap. Hard water damages the image. Ware’s instruction is unambiguous on both counts — alkaline water must not be used, nor hard water containing calcium salts.

Ferricyanide with strong acid: the interaction that is not happening here. A cyanide complex plus acid sounds alarming, and Ware addresses it directly because the claim circulates. His qualification is that the heating has to be very strong and the acid very concentrated to liberate much hydrogen cyanide, and neither is done in cyanotype. This page’s acid baths are between 0.1 and 1 per cent, which is to say between one part in a thousand and one in a hundred of an acid that is itself sold at about two-thirds strength. See Safety and Incompatibilities for where the line actually is.

Dichromate with everything organic. Not used here, and this is one of the reasons: an oxidising chromium(VI) salt in a bottle with oxalate and citrate is thermodynamically a fire looking for a kinetic excuse. In the sensitiser Ware describes it survives because it is 0.1 per cent and cold; as a solid on a bench it is a strong oxidiser whose reactivity profile names paper, wood and plastics.

Variants

The formula as Ware publishes it, with the dichromate. Thirty grams of ammonium iron(III) oxalate, ten of potassium ferricyanide and 0.1 g of ammonium dichromate, made to 100 cc; the dichromate dissolved into the oxalate solution before the two are mixed, or added as 0.5 cc of a 20 per cent w/v solution where the balance cannot weigh 0.1 g. Potassium dichromate may be substituted. This is the version in Ware’s books, on his website, in his workshop notes and in the commercial kits. Its advantages over the version on this page are a shelf life of several years rather than months, a significantly higher contrast, and no risk of a Prussian blue sediment forming in the bottle. It is not offered here, for the reason given at the head of the page, and no procedure for it appears anywhere in this course.

The dilute sensitiser. Ware’s web article allows the filtrate to be made up to as much as 200 cc instead of 100. “It will be faster to print, but yield a less intense blue.” Everything in the bottle halves; the internal filter thins, which is where the speed comes from, and there is less pigment to make, which is where the density goes. No numbers are published for either effect.

The three neighbours in Ware’s own formulary, which he sets out as a choice of five and characterises in a line each:

Formulation	Ware’s own summary
Classic cyanotype	easy to prepare; uncertain; limited tonal range
Classic improved	more toxic; better range and density
New cyanotype (this page)	faster; finer; higher Dmax; toxic; chemically tricky
Simple cyanotype	easy to make; cheap; contrast control; non-toxic — his adjective is not a classification, and that entry will carry its own hazards
Mike’s cyanotype	Simple cyanotype without using ammonia

The variant that is not a variant: New cyanotype mixed with classic cyanotype. Ware calls this “a chemical nonsense” and the reason is precise. The entire preparation above exists to remove potassium ions; the classic sensitiser is full of them, so mixing the two puts back what was filtered out and invites potassium iron(III) oxalate to crystallise in the coating. He adds a second incompatibility: this sensitiser wants a mineral-acid first bath for its best maximum density, while the citrate sensitisers tolerate only a very weak acid, above which they fog blue in the highlights. He is blunter still about the practice of mixing two established formulae and naming the mixture after the mixer, and about the absence of step-tablet evidence that it improves anything.

⁇Whether a dichromate-free New Cyanotype is the same formula

The disagreement, such as it is, is about status rather than chemistry. Ware publishes the dichromate as part of his formula in Cyanomicon and on his website, and marks it optional in his workshop notes; his preparation says omitting it diminishes the contrast and the storage life. Nobody disputes that the sensitiser works without it — Ware’s own characteristic curve in Fig 7.1 is measured on a New Cyanotype sensitiser without added dichromate, and it is the curve he uses to demonstrate the acid-development contrast control.

What the course does. It publishes the three-solid formula as verified, states the omission and its cost in six places, and does not create a separate “variant” entry — because the schema’s variant shape needs a base formula to point at, and the base here is the same page. That shortfall is recorded in the formulary schema notes as an open item.

What a reader should take from it. If you meet “New Cyanotype” in a supplier’s kit or in Ware’s own books, expect the dichromate to be in it, and expect the bottle to keep for years. If you make it from this page you have a sensitiser with a shorter life and slightly softer gradation, and you get the contrast back at the development stage, where this process happens to have a control that the classic one does not.

Safety

Level B, and the classification comes from the ammonium iron(III) oxalate and from the acid bath, not from the ferricyanide. This is a level above the classic sensitiser, and the reason is a genuine difference in the chemistry rather than caution: soluble oxalates are considerably more toxic by ingestion than iron citrate, and the process asks you to heat concentrated solutions and to handle a concentrated mineral acid in order to dilute it.

Ware’s own assessment, which he sets out with HMIS ratings on a 0 to 4 scale, is the clearest short statement of the hazard profile and the course records it as his:

• Ammonium iron(III) oxalate: health 3, contact 3. “Like all soluble oxalates it should be regarded as highly toxic, i.e. capable of causing death or permanent injury (especially kidney damage) if taken by mouth. However, several grams would have to be ingested to cause such extreme damage.” Non-volatile crystals, so little inhalation risk, but an irritant. The course’s encyclopaedia entry gives the notified classification as signal word Warning with H302 and H312 — harmful if swallowed and in contact with skin — and discusses at length why two other sources describe it more severely.
• Potassium ferricyanide: health 1, contact 1. “Classified as a low toxic hazard, but may act as a skin, eye and lung irritant. It is incompatible with concentrated acids, which may release deadly hydrogen cyanide gas. Used correctly there is no risk.”
• Nitric acid, concentrated: health 3, contact 3, and the reason the development bath is made up through an intermediate stock rather than by dropping concentrated acid into a tray.
• Citric acid: health 2. Severe eye irritant, skin and respiratory irritant, and Ware notes that prolonged or repeated exposure may cause an allergic reaction in some individuals.

Ware’s own summary of the whole sensitiser is worth quoting because it is calibrated against processes a reader may already know: it is “toxic if ingested (more so than traditional cyanotype - but no worse than kallitype or platinotype sensitizers)”, and it will stain skin, wood, clothes and anything else absorbent.

The specific controls, none of which replaces the chemical pages:

• Nothing enters the mouth, and nothing is eaten in the room. This is the single control that matters for an oxalate. Gloves for the solutions, hands washed afterwards, dedicated glassware that never goes back to the kitchen, and nothing that has held food used for any part of this work.
• Dilute the acid, and dilute it twice. Ware’s method exists to keep you away from concentrated acid at the tray: make a 10 per cent v/v stock by adding 100 cc of the concentrated acid to water and making up to a litre, then dilute that stock a further 10 to 100 times for use. Acid into water, never the reverse. The corrosive fumes of the concentrated acid are the hazard of that one operation. The finished 0.1 to 1 per cent bath, Ware states, “is not dangerous”, though eye protection still belongs.
• The hot step is a burn and a splash risk. Twenty millilitres of near-saturated ferricyanide at 70 °C is a small volume of hot, dense, staining liquid being poured. Pyrex, a stable surface, and no rushing.
• Ferricyanide is a cyanide complex, not a cyanide salt. The cyanide is bound to iron. It is not the hazard that potassium cyanide is and the two must never be spoken of as the same thing. The complex releases hydrogen cyanide only on strong heating or with concentrated acid, and this process does neither.
• Never bring concentrated acid near the ferricyanide bottle or the sensitiser. This is the boundary the previous point defines. The dilute bath is a controlled use at one part in a hundred or less; a splash of the concentrated stock into a tray of sensitiser is a different chemistry altogether.
• Ultraviolet is the hazard people forget. Two to five minutes of direct sun is short, but a UV lamp is close, does not feel warm, and is used repeatedly through a session. Enclose the lamp, protect the eyes, cover the skin.
• No dichromate, and no dichromate bath. Beyond the sensitiser itself, Ware’s strongest safety statement in the whole chapter is that under no circumstances should a bath of potassium or ammonium dichromate be used to re-oxidise a cyanotype, although it was once recommended; he cites the medical literature on an artist who used one for quilting fabrics and notes that people exposed to dichromate can become particularly sensitised to it. Use air, or hydrogen peroxide.
• The discarded crystals are oxalate too. About 15 g of potassium iron(III) oxalate comes out of every 100 cc batch. Ware’s instruction is to wash them away with hot running water and states that they are not an environmental hazard; the course’s position on that is under Waste.

Storage

One brown bottle, stoppered, labelled and dated, in a cool dark place. The single-bottle keeping is the property the formula was designed for, and the label should carry the date because the version on this page has a life measured in months rather than years.

Expect six months to a year without the dichromate — and read the three printings before you rely on the number. Cyanomicon’s preparation gives six months to a year; its next section says a few months; the 2009 workshop notes say at least a year. The failure mode is the same in all three: with no oxidant in the bottle, traces of reducing impurity slowly convert the ingredients into Prussian blue, which shows first as a sediment and then as fog on the print. Look at the bottle before you coat. A clear yellow-green liquid is good. Blue in it, or on the bottom of it, is the process ending its own shelf life.

Neither the citric acid nor the Tween 20 goes into the stock. Citric acid shortens the shelf life, which is Ware’s own statement; the Tween solution is not indefinitely stable and should be replaced when it shows a deposit. Both are made into the working portion for the session and no further.

Coated paper is a short-term store, and the paper decides how short. Use within a few hours if you can. A cool, dark, desiccated enclosure extends it. The colour of the coating is the test: light yellow is sound, green or blue is fog.

Store the solid oxalate dry and in the dark. It is light-sensitive in the bottle as well as on the paper, and its encyclopaedia page carries the rest.

Keep the finished prints away from alkaline board and buffered tissue. Ware’s permanence note is short and worth repeating: Prussian blue is destroyed by alkali, so mounts and wrappings above about pH 9 must be avoided; cyanotypes can fade somewhat in bright daylight but recover on dark storage in air, regaining their density fully after a few days; and exhibition at low light levels causes no measurable fading.

Incompatibilities

Alkali, at every stage and in every form. Buffered paper, alkaline sizing, alkaline tap water, ammonia, household cleaners, buffered mountboard, ISO 9706 “permanent” paper — which by definition contains an alkaline reserve and is therefore exactly the wrong sheet. Where a buffered paper cannot be avoided, Ware’s pre-treatment is a bath of 5 per cent v/v hydrochloric acid, or 5 to 10 per cent w/v sulfamic acid, for about ten minutes, followed by a wash, to destroy the calcium carbonate.

Hard water, for the calcium, at development, wash and everywhere else.

Concentrated acid with ferricyanide, as set out under Safety.

Acetic acid, specifically, as a development bath. Ware rules it out on chemical grounds rather than safety ones: it tends to form insoluble, strongly coloured basic iron(III) acetate. This is a useful reminder that “an acid” is not a category you can substitute freely within.

Classic or Simple cyanotype sensitiser, mixed with this one, for the potassium reason and the developer reason under Variants.

Dichromate, in the sensitiser and above all as a re-oxidising bath.

Silver, and everything in the darkroom that touches it. Ferricyanide is the oxidising half of Farmer’s reducer. A shared measuring cylinder, a shared tray, a splash in a sink where prints are washed — any of them will attack a silver image. Ware’s own warning about the discarded oxalate crystals belongs here too, from the chemical page: soluble oxalates and silver compounds are incompatible, with explosive silver oxalate named as the product, so oxalate waste is never bottled with anything silver-bearing.

Gelatin-sized papers with Tween 20, which Ware notes may interact unfavourably.

Metal — trays, tongs, clips, sinks. An iron-and-hexacyanoferrate chemistry will find a trace of iron and print it.

Waste

Three streams, and none of them is silver. The discarded crystals, about 15 g per batch of potassium iron(III) oxalate; the spent development bath, dilute mineral acid carrying peptized Prussian blue and dissolved sensitiser; and the wash water, which carries the rest of the unexposed sensitiser.

Neutralise before anything else. The acid bath is the one genuinely reactive stream. Ware’s general instruction is that if waste seems very acidic — test it with pH paper — it can be neutralised with a cheap alkali such as lime or chalk. Do that in a labelled container, not in the sink, and expect the Prussian blue in it to be destroyed by any excess alkali, which is chemically fine and visually disconcerting.

Never acidify the ferricyanide-bearing streams further, and never put a chlorine-releasing household cleaner down the same drain.

The crystals are oxalate. Ware’s instruction is to wash them away with hot running water and his statement is that they are not an environmental hazard. The course records that and does not adopt it as an instruction: the chemical encyclopaedia’s entry for the salt notes that CAMEO records the primary hazard of ferric ammonium oxalate as the threat to the environment, with immediate steps advised to limit its spread, so nothing from this process goes onto a garden, into a soakaway or into a watercourse. Bottle it, label it, and follow the general chemical waste SOP and the course’s disposal ruling.

⁇How hazardous is cyanotype effluent?

The disagreement. Ware argues that the processing solutions are very dilute and that treating the effluent as chemical waste is “quite unnecessary from an environmental viewpoint, and an expensive and needlessly troublesome consumption of vital resources”. His grounds are specific to named substances: ferricyanide reduces in the environment to the more stable ferrocyanide, a permitted food additive; Prussian blue is so insoluble that a human can eat and excrete grams a day, which is why it is a clinical antidote to thallium and caesium poisoning; nitrate is of very low toxicity and is spread on farmland by the tonne. He is more careful about the oxalates, saying only that they are poisonous when concentrated and are naturally present in rhubarb, spinach and sorrel at the 1 per cent level.

Who holds what. That is a chartered chemist writing about his own process, in a passage aimed at practitioners working with septic tanks and land drainage. It is not a statement about what a particular sewer authority permits, and he does not offer it as one.

What this course does and why. It publishes no jurisdiction-specific disposal instruction anywhere and makes no exception here. Collect, label, follow the SOP, comply with local regulation — which is the only document that can answer the question. Ware’s argument is recorded because it is well founded and because a reader is entitled to know that this chemistry is benign relative to much of what else is in a darkroom, not because it settles what may go down a drain. The one place the course is firmer than Ware is the oxalate: 15 g of a soluble oxalate per batch is not the same object as a tray of dilute wash water.

Troubleshooting

The sensitiser has gone blue, or has a blue sediment. Reducing impurity has made Prussian blue in the bottle. This is the failure the omitted dichromate was there to prevent, and it is the main cost of the version on this page. Filtering will clear the liquid but will not restore what has been consumed; make a fresh batch, keep it cooler and darker, and use it sooner.

The coating turns green or blue before exposure. Chemical fog from the paper. Ware’s diagnosis is that the sensitiser is a sensitive test of paper quality and that a colour drift in the dark condemns the sheet. The two remedies are a better, unbuffered paper, and citric acid to about 2 per cent in the coating solution. Dry the coating rapidly, which he notes can reduce fogging from impurity.

White specks, or fern-like patterns, in the coating. Potassium iron(III) oxalate crystallising in the paper. Either the crystallisation was cut short — Ware allows at least two hours and prefers overnight — or the filtration was incomplete, or the sensitiser has been contaminated with a potassium-bearing solution such as classic cyanotype sensitiser. There is no rescue for a coated sheet; there may be for the bottle, by re-cooling and re-filtering it.

Weak blacks. Three sourced causes. The development bath is too weak or is water only, which gives the softest gradation and a reduced maximum density by design — go to 1 per cent nitric. The sheet was underexposed, which on a print-out process you can see: the shadows should be reversed. Or the sensitiser has been diluted, deliberately or by a poor make-up.

The print is too contrasty. Weaken the development bath towards 0.1 per cent, or develop in water alone. Add citric acid to the coating solution, which Ware states lowers contrast enough to accommodate a negative of density range 2.6.

The print is too flat. Strengthen the acid towards 1 per cent. If it is still flat, the negative is short — this process wants 1.8 to 2.4 in the ultraviolet, and a normal silver-printing negative will not fill it.

Yellow stain in the highlights after washing. Sensitiser has not cleared. Ware’s test is to hold the print up to a bluish light, or look through a blue filter, and see whether yellow remains inside the sheet; if it does, wash longer. It is a wash fault and not an exposure fault.

Blue staining across the whole sheet after development. The development bath has been used past its capacity and is carrying Prussian blue in suspension. Replace it after five to ten prints per litre; do not re-use it.

The image weakens during washing. Much less likely than with the classic process, and when it happens the causes Ware names are alkaline or hard water, over-rapid drying of the coating before exposure, or a paper whose sizing keeps the sensitiser in the surface pores instead of inside the fibres.

The shadows stay reversed. They are Prussian white and they need oxygen. Twenty-four hours in air will do it; half a minute in 0.3 per cent hydrogen peroxide does it at once. Not a dichromate bath.

Batch-to-batch density changes. Ware names this as a real limitation: the yield of the crystallisation varies with how it was cooled, so the final concentration varies too, and workflows that calibrate digital negatives tightly may need recalibrating when the bottle changes. Record the filtrate volume before make-up — it should be about 60 to 62 cc — as your batch’s fingerprint.

Experiments

1. The contrast control, measured. Coat six sheets from one batch, expose all six identically through a step tablet, and develop one each in water, 0.1, 0.25, 0.5 and 1 per cent nitric acid, and 1 per cent hydrochloric. Read the steps. Ware’s published range is an exposure scale of about 2.6 down to about 2.1, and his statement that the hydrochloric result is identical to the nitric is a claim you can check in an afternoon. This is the experiment that proves the process has a development stage, which most cyanotype writing says it does not.

2. Where the potassium went. Weigh the crystals after filtration and compare with the 14.9 g the arithmetic predicts. Then make a second batch, filter it after only thirty minutes of cooling instead of overnight, weigh those crystals, and coat both. The prediction from the chemistry is a lower yield, a higher potassium content in the sensitiser, and white specks in the coating from the second batch. Record the filtrate volumes, which are the cheapest indicator of what happened.

3. Ammonium blue against potassium blue, in the wash. Coat one sheet with this sensitiser and one with classic sensitiser, expose both to full density, and wash both in the same tray of the same water for the same time, catching the wash water in a white tray. The classic wash goes visibly blue and the print loses its high values; this one should not. This is peptization, and it is the difference the whole preparation exists to buy.

4. The internal filter, if you have the light sources. Print the same step tablet under a 365 nm source and under a source weighted to 400 to 420 nm, at exposures matched for total energy if you can measure it, or simply record both times to a fixed density. Ware’s calculated absorbed fractions — 0.65 at 365 nm and 0.03 at 420 nm — predict a very large difference. This experiment explains why two workers with “a UV box” can report exposures that differ by a factor of ten.

5. Heat-drying and darktime, in duplicate. Ware’s own four-way test: control, darktime only, heat-dry only, and both. His result is that darktime lengthens the scale by a stop at the cost of maximum density and edge fog, heat-drying shortens it by a stop, and the two together return you to where you started minus the black. Reproducing it is the cheapest way to learn to read a step tablet and the fastest way to settle an argument that recurs constantly in alternative-process forums.

6. The dilution Ware offers and does not quantify. Split a batch of filtrate, make one half to 100 cc and the other to 200, and print both to the same maximum density. Record the exposure ratio and the maximum densities. Ware’s sentence — faster, but a less intense blue — has no numbers behind it anywhere in the sources this page rests on, and a step tablet supplies them.