Iron(III) nitrate nonahydrate

This is the salt that Mike Ware’s Simple Cyanotype is named after. Not the process, the salt: he called the sensitiser Simple because it starts from what a chemist calls a simple salt — a pure crystalline compound of known composition, containing monomeric hydrated iron(III) ions — instead of from the complex salts the older cyanotypes buy ready-made. Everything interesting about it follows from that plainness, including the one thing you must never do with it.

In photography

It is the iron, and nothing else. In Simple Cyanotype and in Mike’s Cyanotype this salt supplies the iron(III) that light reduces. It does not supply the light-sensitive compound. That is made in the beaker, by dissolving these crystals into a solution that already holds twice their molar quantity of citrate, so that the photoactive dicitratoferrate(III) anion forms in place. The reason for choosing a simple salt is an argument about reproducibility rather than chemistry: the ammonium iron(III) citrate of the classic sensitiser is a polymer of variable composition sold at iron contents anywhere from 14 to 28 per cent, so a recipe written in grams of it does not fix how much iron is in the bottle. A recipe written in grams of this salt does — subject only to how much water the jar has taken up from the air, which is a smaller and far more manageable uncertainty than a factor of two.

⚗Why it must never be dissolved in water alone

Ware states the trap plainly: if this pale violet salt is dissolved in water only, “hydrolysis and polymerization of the ferric species begins immediately”. The hydrated Fe³⁺ ion is a weak acid: it polarises the water molecules bound to it strongly enough that one of them gives up a proton, and the hydroxo species that results condense with one another into the polymeric, ill-defined iron(III) the whole formula was designed to avoid. Put the citric acid in the beaker first and the ligand captures the iron before the water can; put the water in first and you have made, at some expense, a worse version of ferric ammonium citrate.

Fe³⁺ + H₂O → [Fe(OH)]²⁺ + H⁺

The first hydrolysis step, which is where the polymer starts and why the order of mixing is not a style preference

Ware’s instruction sheet makes the same point as a sequence rather than a mechanism: weigh the 12 g, then immediately add it to the citric acid solution, in portions, stirring each portion in before the next. The word immediately is doing two jobs — it keeps the deliquescent crystals off the balance pan and out of the air, and it keeps them out of any beaker of plain water.

The nitrate does not leave when its job is done. Every nitrate ion that came in with the iron stays in the bottle, and at 12 g of the salt in 100 cm³ that is 0.089 mol of nitrate against 0.030 mol of iron. Ware regards this by-product as an advantage and gives four reasons, and the course reproduces his own levels of confidence rather than flattening them into a list of facts.

⁇Ware's four claims for the nitrate, at his own strength

1. It keeps the bottle alive. Stated as a prediction from chemistry: at low pH the oxidising nature of nitrate “should prevent” the formation of iron(II) during dark storage, which would otherwise precipitate Prussian blue with the ferricyanide in the same bottle. On this reading the nitrate is the single-bottle sensitiser’s only preservative, doing what ammonium dichromate does in New Cyanotype without bringing chromium(VI) into a domestic room.
2. It restores solarised shadows. Stated flatly, without a citation: on development in about 1 per cent citric acid at roughly pH 4, the nitrate also helps re-oxidise to Prussian blue the “Prussian white” that heavy exposure leaves in the deepest tones.
3. It may help the image print out. Stated as a possibility: ammonium nitrate left in the dried coating makes it somewhat hygroscopic, and the retained humidity may assist print-out, as it is observed to do in the Malde-Ware print-out platinum-palladium process.
4. It probably inhibits mould. Stated as a likelihood, and set against the classic citrate sensitiser’s habit of growing mould in the bottle within a fortnight.

Ware notes that adding ammonium nitrate to improve the stability and quality of blueprint paper was itself the subject of US patent 2,113,423 in 1938, which is corroboration of the effect and not of the mechanism. This page asserts no stoichiometry for any of the four. The nitrate couples are strongly pH-dependent, the kinetics in a citrate-buffered sensitiser are not published, and writing a balanced equation here would be inventing chemistry to fill a gap the source itself leaves open.

Herschel had it in his hands in 1842 and put it down. His first cyanotype experiments coated paper with potassium ferricyanide alone, and were extremely slow. Ware records that developing the exposed sheet in a bath of iron(III) chloride or nitrate gives greater speed and higher densities but “invariably causes blue fog in the highlights” — because flooding the whole sheet with iron(III) supplies the missing half of the Prussian blue reaction everywhere, not only where the light fell. The same failure is described, from the chloride’s side, on the iron(III) chloride hexahydrate page. It is one of the reasons Herschel abandoned the ferricyanide-only method for the much faster citrate sensitiser provided by Alfred Smee.

And a trade used it long before either. Ware records that nineteenth-century industry dyed cotton, linen, silk and wool in Prussian blue with two baths: first a solution of a ferric salt, “usually the nitrate or chloride”, which mordanted strongly to the fibres, then, after rinsing, an acidic bath of potassium ferrocyanide that formed the pigment in place. That is the industrial ancestor of every cyanotype printed on fabric, and it explains why the iron-based processes suit textiles: they need no gelatin or gum to hold the image, so the cloth keeps its drape.

Properties

Two fifths of the weight is water. At 404.00 g/mol against 241.86 for the anhydrous salt, nine waters of crystallisation account for 162.14 g of every mole — 40.1 per cent by mass — and the iron itself for only 13.8 per cent. A formula written for the crystals needs 1.670 times the weight if the anhydrous salt is all that is to hand, and 0.599 times for the reverse, on the course’s own arithmetic from the two molar masses. Ware prints the same 404.00 as the relative molar mass, so a recipe in grams and this page agree.

It is deliquescent, and that is not a storage inconvenience but a measurement error. A salt that takes up water from the air weighs heavy in water and light in iron, so the same 12 g weighed from a jar opened fifty times is less iron than the first weighing was. Ware goes further than the usual warning and says the absorbed water reacts with the salt and destroys it irreversibly. The practical rules follow: buy small quantities, keep the lid on, weigh quickly, and do not open the jar in a humid room.

It melts at 47.2 °C. That is a hot day in a conservatory or a shelf above a radiator, and for an oxidiser it is worth knowing: the crystals do not so much melt as dissolve in their own water of crystallisation, and what results is a corrosive syrup that wets whatever it stands on. CAMEO’s datasheet, filed under the anhydrous CAS, gives 117 °F and a specific gravity of 1.7 at 68 °F and a molecular weight of 404.02 — which are 47.2 °C, 1.7 at 20 °C and the nonahydrate’s formula weight, so on the course’s own comparison of the two records all three are the nonahydrate’s figures under the anhydrous salt’s name. That is a reminder that “ferric nitrate” in a data source rarely means the anhydrous salt in practice, because the anhydrous salt is not what anybody sells.

Its solution is acidic and its solubility is not well recorded. The Merck Index entry, held on PubChem’s parent record, says the nonahydrate is freely soluble in water, alcohol and acetone and only slightly soluble in cold concentrated nitric acid, and gives no figure. CAMEO records water solubility as data unavailable. The 150 g per 100 cm³ that PubChem carries from the CRC Handbook is labelled for the hexahydrate and the course does not carry it across to a different salt. What can be said with confidence is that 12 g dissolves without difficulty in the roughly 35 cm³ of citric acid solution Ware’s first step makes, which is all a printer needs, and that the resulting solution is yellow and acid enough to keep itself clear.

Handling

The aggregated notifications PubChem holds classify the nonahydrate Danger, with three pictograms: oxidiser, corrosive and irritant. Seven notifiers in ten give skin irritation and serious eye irritation, six in ten give respiratory irritation and the oxidiser statement, and close to three in ten give the severe pair — H314, causes severe skin burns and eye damage, and H318, causes serious eye damage. Ware’s own entry in the Cyanomicon hazard list reads, in the same order and to the same effect: “Causes serious eye irritation. May cause skin irritation. May intensify fire; oxidizer.”

There is no harmonised classification, and that is a finding rather than an absence of risk. The substance’s ECHA CHEM record carries no CLP Annex VI index number, and its regulatory processes are the legislation-finder inventory, a notified classification and a 2008 pre-registration — with no REACH registration behind them. So every hazard statement on this page is a notifier’s own opinion, aggregated, and the spread between notifiers is wide enough to matter: the same substance is an irritant to seven in ten of them and a skin corrosive to three. The course handles it against the harsher reading. It is also worth knowing what the classification does not cover: PubChem’s toxicological section records secondary hepatotoxicity and methaemoglobinaemia as adverse effects, and the Merck Index summary says it may cause liver damage — none of which appears in any H-statement, because nobody has notified one.

Why Level B. Against the course rubric, three of Level B’s criteria apply at once. It names sensitisers and “still toxic or corrosive reagents”, and this is both: three notifiers in ten classify it as a skin corrosive. It names procedures whose failure mode is “a splash, a burn or a spill of something corrosive rather than merely a spoiled negative”, which is what a knocked beaker of this does. And it is a classified oxidiser weighed out as a solid on a bench that is usually made of paper — a hazard Level A’s ceiling of “at most irritant, harmful if swallowed, or corrosive at the concentrations actually handled” does not contemplate at all, because H272 is not a health statement and no concentration makes it go away. It does not reach Level C, whose criterion applies where a fume cupboard or specialist disposal is the recognised control — goggles, gloves, a clear bench, dust discipline and an open window are the recognised controls here, and they are the ones Ware’s own sheet asks for.

⚠Safety

Goggles before the jar is opened; nitrile gloves; eyewash within reach. Weigh it on a clear bench, not on a sheet of paper or card, and clear away anything combustible before the lid comes off — this is an oxidiser, and the darkroom is full of paper. Add the solid to the acid solution in small portions with stirring, never the other way round, never dry, never warmed, and never at more than the published quantity. Keep it away from every alkaline bottle in the room, from reducing agents, and from the ferricyanide jar, because the two make Prussian blue on contact and the stain is permanent. Mop a spill with sand or vermiculite, not with paper towel: CAMEO names cellulose-based absorbents as potentially incompatible with this reactive group.

On the pairing with citric acid, which the formula deliberately makes. CAMEO’s citric acid datasheet states that citric acid reacts with metal nitrates and that those reactions are potentially explosive, and the ILO-WHO safety card tells you to store citric acid separated from metal nitrates. Ware’s published preparation puts a metal nitrate into a solution of free citric acid on purpose, in water, at room temperature, at a scale of 12 g. The course prints that procedure as its author published it, under the ruling that a formula published by the person who devised it is primary evidence for that formula, and states equally plainly that the warning is real and that nothing here licenses improvisation on it — not a dry mixture, not a heated one, not a larger batch.

Waste

What actually leaves the bench is a small quantity of leftover sensitiser, which is a concentrated nitrate solution, and a much larger volume of wash water carrying the unexposed sensitiser away. Nitrate governs both. A 100 cm³ batch of Simple Cyanotype sensitiser carries about 0.089 mol of nitrate — roughly 7 g expressed as ammonium nitrate — and nitrate is a fertiliser: it does its environmental damage by feeding things, not by poisoning them, so nothing from this process goes onto a garden, into a soakaway or into a watercourse. Solid residue and spill sweepings are an oxidiser: keep them damp, keep them out of a bin of paper and rags, and use sand or vermiculite rather than a cellulose absorbent to pick them up. Kodak’s J-52 publication gives 5.6 to 9.4 as the pH window that sewer codes most frequently set, and any working solution of this salt sits well below it. Ware’s own environmental judgement is that nitrate is of very low toxicity and that dilute cyanotype processing baths do not merit treatment as chemical waste; that is his reading of the dilute baths, it is not a statement about a discarded bottle of sensitiser, and the course keeps its controls either way. ILFORD’s guidance for domestic users in the United Kingdom is to bottle wastes separately, label them and take them to a household waste and recycling centre’s chemical cupboard. Check your local regulations; they govern, and they differ.

History

The first person to make an image with this salt was not a photographer and did not know he was making one. In 1828 John Mercer (1791–1866), whom Ware describes as a Mancunian calico-dyer and printer who became a self-taught colour chemist and finished as a Fellow of the Royal Society, wrote in his experimental notebooks:

I spotted a piece of white cloth with a solution of pernitrate of iron, and exposed it to the sun. On testing it afterwards with a solution of red prussiate of potash it gave a blue — but no blue before exposure. This is worthy of note. I have not seen it noticed by any chemical writer.

🕰What Mercer had actually done

“Pernitrate of iron” is this salt. “Red prussiate of potash” is potassium ferricyanide. The sunlight had reduced some of the iron(III) in the cloth to iron(II), and the ferricyanide then found it. Ware’s note on the passage supplies the missing reagent: although ferric nitrate is not an organic salt of iron, “it was doubtless the cellulose of the cloth that provided the oxidisable material” — the fabric itself was the electron donor. That is the whole of a cyanotype, fourteen years before Herschel named one, and it is why this page does not list the salt among the light-sensitive salts: on its own, in solution, it is not one. It became one because it was lying on cotton.

Fe²⁺ + [Fe(CN)₆]³⁻ → Fe[Fe(CN)₆]⁻

What Mercer's red prussiate revealed: iron(II) plus ferricyanide, before the internal electron transfer that makes the product identical to Prussian blue

Mercer had a business to run and did not follow it up; by the time he returned to photographic chemistry in the 1840s, other and more leisured men had established the field. Ware’s verdict is that if one is looking for an unsung figure of early photography, Mercer is a worthy candidate.

The salt then spent the best part of two centuries as somebody else’s reagent. The Merck Index entry PubChem carries lists its uses as dyeing, where it is a mordant, the weighting of silks, tanning, the inhibition of corrosion and chemical analysis — which is to say, the thing you reach for when you want iron(III) and do not want an interesting anion attached to it.

Its return to photography is recent and is one chemist’s work. In 1998 Matzapetakis and co-workers, in a Journal of the American Chemical Society paper this course knows only through Ware’s citation of it, isolated and characterised by X-ray structure determination the first mononuclear, water-soluble iron-citrate complex, (NH₄)₅Fe(C₆H₄O₇)₂·2H₂O. Their starting material was ferric nitrate nonahydrate, dissolved directly in citric acid rather than in water, neutralised to about pH 8, and then thrown out of solution with a large volume of ethanol because the product was too soluble to crystallise otherwise. Reading that, Ware asked the question the paper does not: if the complex forms in solution, why isolate it at all? Simple Cyanotype, 2019, is the answer — citric acid, this salt, ammonia and ferricyanide, three solids and a bottle of ammonia, under an hour, and a sensitiser about four times faster than a typical classic formula with an exposure scale of about 2.7 against about 0.9. Mike’s Cyanotype followed, reaching the same solution from triammonium citrate so that the ammonia bottle never has to be opened.

The historical symmetry is worth naming. Herschel tried iron(III) nitrate in 1842, as a developing bath, and it fogged his highlights; Mercer had already used it in 1828, as a sensitiser, and could not pursue it. It took until 2019 for someone to work out that the salt belonged at the beginning of the preparation rather than at the end of the exposure — and that the whole difficulty had always been the half-second in which it meets water.