Ammonium citrate (tribasic)

Mike Ware’s Simple Cyanotype is made by dissolving citric acid, adding an iron(III) salt and then carefully neutralising the whole thing with strong ammonia. Ware’s correspondents told him that some of them could not obtain strong ammonia solution and some could not tolerate handling it. This salt is what he offers those readers instead: the neutralisation already done, sold as a white powder, weighed on a balance rather than pipetted out of something that fumes. Everything else on this page follows from that — including the reason it is so easy to buy the wrong jar.

In photography

One published formula uses it, and that is the whole of its photographic career. Section 7.5 of the Cyanomicon heads it “Mike’s cyanotype” and explains why it exists in two sentences: correspondents told Ware they could not obtain, or could not tolerate handling, strong ammonia solution, and this variant is for them. It produces — odourlessly, as he puts it — a sensitiser that should be identical with the one his Simple Cyanotype produces with ammonia. The course’s page for that formula is Mike’s Cyanotype.

The salt is the first thing into the beaker and the largest weight in the formula:

• 15 g of triammonium citrate, which Ware also names ammonium citrate, tribasic, dissolved in 40 cc of purified water. He notes the crystals are colourless, that the solution cools as they go in, and that it may be warmed to help them.
• Then 12 g of iron(III) nitrate nonahydrate, added slowly and never dissolved in water on its own.
• Then 10 g of ground potassium ferricyanide, and water to make 100 cc.

It is doing three jobs at once, and only the first is obvious. It supplies the citrate ligand, at two moles per mole of iron, which is what holds iron(III) as a single well-defined ion instead of letting it hydrolyse and condense into the polymers that make commercial ferric ammonium citrate so variable a material. It supplies the ammonium ions — three per citrate — which in the parent formula have to be titrated in as ammonia. And its citrate is the ligand light destroys, because in an iron-based printing process the electron that reduces the iron comes off the organic ligand bound to it.

⚗What 'tribasic' means, and why it is the whole story

Citric acid is triprotic: three carboxyl groups, each able to give up a proton, with the third and weakest leaving at a pKa the IUPAC dataset puts at 5.68 at 20 °C in 0.1 mol/L sodium perchlorate. Neutralise all three with ammonia and you have this salt, with nothing acidic left on the molecule:

C₆H₈O₇ + 3 NH₃ → (NH₄)₃C₆H₅O₇

Citric acid plus three ammonia: the neutralisation this salt is sold as, already done

Neutralise only two of the three and you have the dibasic salt, (NH₄)₂HC₆H₅O₇, which keeps one proton and gives a mildly acidic solution. That salt has its own page at ammonium citrate, its own CAS number and its own ECHA record. It is a different substance sold under a name that differs by one word.

The difference is not cosmetic and it is not small. The tribasic salt carries three ammonium ions per citrate; the dibasic carries two and an acidic proton. In a formulation whose entire design is the careful neutralisation of a citrate solution, that is the difference between the recipe working and the recipe being about a third of an equivalent short of base per citrate.

🧮Checking Ware's substitution: does 15 g of the salt really equal the ammonia?

Ware states that the two preparations give the same solution. The claim can be checked against his own published quantities, and it is worth checking, because the whole point of the variant is that nothing is lost by making it this way. The arithmetic below is the course’s, not his.

The low-contrast grade of the parent formula (§7.4.2) takes 12 g of anhydrous citric acid and 33.0 cc of 10 per cent w/w ammonia, for which Ware’s own instruction sheet gives a specific gravity of 0.959. The variant (§7.5.1) takes 15 g of this salt instead of both.

• Citrate from the salt: 15 ÷ 243.22 = 0.0617 mol. Citrate from the acid: 12 ÷ 192.12 = 0.0625 mol. The variant is 1.3 per cent low.
• Ammonium from the salt: 0.0617 × 3 = 0.185 mol. Ammonia pipetted in the parent: 33.0 cc × 0.959 g/cc × 0.10 = 3.16 g, ÷ 17.03 = 0.186 mol. The variant is 0.4 per cent low.

So the substitution is equimolar to within about one part in a hundred on both counts — far inside the weighing tolerance of a 0.1 g balance on a 12 g weighing. Ware’s claim survives the arithmetic, which is the strongest statement this course can make about a formulation nobody has published an independent test of.

What the substitution costs: the contrast control. The parent formula publishes three grades, and the thing that selects the grade is the volume of ammonia: 33 cc of the 10 per cent reagent gives an exposure scale of about 2.7 at a sensitiser pH of about 4, and 44 cc gives about 1.8 at a pH of about 8, with a middle grade between them. Those figures are published for the parent and measured with the iron and the ferricyanide present. In this salt the ammonium is welded to the citrate in a fixed 1:3 ratio, so adding more of it adds base and ligand together and cannot move the pH. Section 7.5 accordingly publishes one formulation, the low-contrast one. A reader who wants the harder grades has to go back to the ammonia bottle. That is a real loss, stated here because a page that presented this salt purely as a convenience would be selling it.

📷Why a citrate is photosensitive at all, and acetate is not

The light reaction is not the citrate’s alone — it belongs to the iron(III) complex — but the ligand decides whether there is a reaction to have. Ware sets out the mechanism in the Cyanomicon’s appendix: light drives charge transfer from ligand to metal, iron(III) is reduced to iron(II), and the ligand then undergoes oxidative decarboxylation, losing a carboxyl group as carbon dioxide.

Table III.1 of that appendix lists the carboxylic acids whose iron(III) salts are photosensitive — oxalic, citric, tartaric, malic, lactic, mandelic and a dozen more — and Ware names the structural feature that most of them share: an α-hydroxy group, next to the carboxyl that leaves, which lets the decarboxylation proceed without stranding an energetically unfavourable carbonium ion. Citrate has exactly that: a hydroxyl on the central carbon. Ferric acetate, which has no such group, shows no photoactivity at all.

That is the reason this salt and not some other cheap organic ammonium salt is in the bottle, and it is why the same word — citrate — keeps appearing across the iron-based processes.

Where it is not used, which matters as much. The citrate developers and clearing baths of the platinum and palladium literature are a different family, and where that literature is specific it specifies the other salt: writing about the lightest Japanese washi papers, Ware records that alkaline tetrasodium EDTA damages the fibres and that “the use of mildly acidic diammonium citrate as developer and clearing agent is preferred”. Mildly acidic is a property this salt does not have. No source read for this page puts the tribasic salt in a palladium developer, a clearing bath, a toner or an emulsion. Where a supplier’s label or a practitioner’s note says only “ammonium citrate”, the course does not assume it means this one; the CAS number settles it and nothing else does.

The negative finding runs the other way too. Ware’s own Simple Cyanotype instruction sheet of 2022 — eleven pages of it — reprints the parent sensitiser weight for weight and adds a two-bottle arrangement, and it does not carry the ammonia-free route at all. Its chemicals list names citric acid, iron(III) nitrate nonahydrate, ammonia solution with a table of specific gravities, potassium ferricyanide and Tween 20. This salt appears in one document.

Properties

Identity, which on this substance is the load-bearing section. Triammonium citrate, (NH₄)₃C₆H₅O₇, Hill formula C₆H₁₇N₃O₇, relative molecular mass 243.22, CAS 3458-72-8, EC 222-394-5, ECHA CHEM record 100.020.359, PubChem CID 18954, ChEBI 63037. PubChem’s deprecated-CAS field carries three further numbers for the same substance — 113131-83-2, 141102-47-8 and 2568981-92-8, the first two of which ECHA also lists against its record — so a supplier or an old paper quoting one of those is quoting this. It is the food additive E 380, INS 380, classed by JECFA as an acidity regulator and described by ChEBI as having a role as a buffer and a food emulsifier.

Against it: the dibasic salt, (NH₄)₂HC₆H₅O₇, M = 226.18, CAS 3012-65-5, EC 221-146-3, ECHA CHEM record 100.019.225, filed under the name Diammonium hydrogen 2-hydroxypropane-1,2,3-tricarboxylate. Searching the registry on that CAS number returns a different record with a different identifier, which is the cleanest demonstration available that these are two substances and not two spellings.

!The confusion is in the registries, not only on the label

It would be comfortable to treat this as a labelling problem that careful shopping solves. The primary records do not support that.

PubChem’s own record for this compound carries the note “See also: Ammonium citrate (annotation moved to)”, pointing at a third, generic record — so some of the annotation for the specific salt has been consolidated under the ambiguous name. ECHA’s substance record is worse. Among the IUPAC names submitted against this substance, EC 222-394-5, the registry lists “triammonium citrate” and “Ammonium citrate, tribasic” as you would expect, but also the bare “ammonium citrate” and, from one submitter, “Di Ammonium Citrate” — the name of the other salt, filed against this one. Its registered trade name is simply “Ammonium citrate”.

So the substance’s own regulatory record contains the mistake this page exists to prevent. The practical consequence: verify by CAS number, not by name, and where a formula matters, verify against the supplier’s certificate of analysis rather than the jar. The two salts differ by 7.5 per cent in molar mass and by one proton, and only one of those differences is visible on a balance.

Physical form and what dissolving it feels like. White to off-white crystals or powder. It is freely soluble in water, and although no source read here gives a figure, Ware’s procedure puts 15 g into 40 cc — about 37 g per 100 mL — without any hint of approaching a limit. The dissolution is endothermic: the beaker goes cold, and Ware’s instruction is that it may be warmed to help the crystals go in. A reader who has only ever dissolved sodium hydroxide will find that startling, and it is worth naming, because a cold beaker with undissolved crystals in it looks like a failure and is not.

One physical description on the PubChem record reads simply “Liquid”, contributed from the United States EPA’s Chemical Data Reporting. That is the industrial aqueous concentrate, not the solid, and it is a useful warning: this substance is traded as a solution as well as a powder, and a formula written in grams means the solid.

pH behaviour, and the honest limits of what can be said. All three carboxyls are deprotonated, so the salt has no acidic proton to give. Citrate³⁻ is therefore the conjugate base of an acid whose third pKa is 5.68, and it will take a proton back from water; the ammonium ion is a weak acid pulling in the opposite direction. The net is a solution that is certainly not acidic and, on the general chemistry of a salt of a weak acid with a weak base, a little to the alkaline side of neutral. No source read for this page reports a measured pH for a solution of this salt, so no number is printed here. The pH values Ware publishes — about 4, 6 and 8 — are sensitiser pH values measured with iron(III) and ferricyanide present, and they are not this salt’s.

That distinction matters in exactly one place, and it is the place the reader will meet it. The dibasic salt is chosen in palladium work because its solution is mildly acidic. This salt is not a substitute for it in that role and the substitution does not run in either direction.

Handling

The classification, as notified. Aggregating 278 reports across 16 notifications to the ECHA C&L Inventory, the signal word is Warning with the exclamation-mark pictogram GHS07, and three hazard statements: H319 Causes serious eye irritation on 97.5 per cent of the classifying reports, H315 Causes skin irritation on 45 per cent, and H335 May cause respiratory irritation on 45 per cent. Seven of the 278 reports state that the substance does not meet GHS hazard criteria at all. The hazard classes are Skin Irrit. 2, Eye Irrit. 2 and STOT SE 3.

There is no harmonised classification, and that is a finding rather than an absence of one. The substance’s ECHA CHEM record has an empty index number — the CLP Annex VI identifier a harmonised entry would carry — and its regulatory processes are the EC Inventory, REACH pre-registration, a REACH registration and a CLP notification, at a tonnage band of 10 to 100 tonnes. Everything above is therefore a notifier’s own view, not an agreed European position, and the split between 97.5 per cent and 45 per cent is what disagreement looks like in this data.

Nobody has set an exposure limit. HSE’s EH40 Table 1 lists no citrate of any kind and no ammonium citrate, and paragraph 6 of its introduction states that the absence of a substance from the list does not indicate that it is safe: exposure should be controlled to a level to which nearly all the working population could be exposed day after day without adverse effect. The NIOSH Pocket Guide has no entry either — neither the CAS number 3458-72-8 nor any citrate appears in its indexes. So dust discipline here is a response to having nothing to measure against, not to a published figure.

Why Level A. Against the course rubric this meets Level A on every criterion: the classification tops out at irritant; a 15 g weighing and a 100 cc make-up are home darkroom quantities; nothing is heated beyond gentle warming to help the salt dissolve; there is no mains work; and the waste is a bottleable solution. Gloves, splash goggles for the powder, and a labelled jar are the controls. Ware’s own instruction sheet for the parent formula opens its safety section with “PROTECT YOUR EYES with safety goggles”, and the eye statement is the one nearly every notifier agrees on, so goggles rather than spectacles is the course’s position for the weighing as well as the mixing.

⚠The substance is Level A; the formula it goes into is not

A safety level belongs to a procedure, not to a substance, and this is a clean example of the gap. Weighing this salt and dissolving it in water is Level A work. Mike’s Cyanotype is classified Level B, and none of the reasons are this ingredient: they are the potassium ferricyanide, the iron(III) nitrate, the ultraviolet exposure and the used baths. Reading a Level A ingredient as permission to relax the controls on the formula is precisely the error the rubric’s insistence that a level belongs to a procedure exists to prevent.

There is one control that belongs specifically to this substance, and it is easy to miss because it inverts the reason for buying it. This is an ammonium salt, and a strong base liberates ammonia from any ammonium salt. Somebody who has chosen this powder in order not to handle ammonia has every reason not to store it beside the sodium hydroxide, and no reason at all to add an alkali to a citrate solution to “adjust” it.

Waste

A dilute citrate solution is about the least demanding effluent an alternative process produces, and Kodak’s J-52 publication gives 5.6 to 9.4 as the pH window sewer codes most often set — a range a neutral citrate solution sits comfortably inside. That is not what decides where the liquid goes.

In its one photographic use this salt never leaves the darkroom on its own. It leaves inside a spent cyanotype sensitiser, in the citric acid developing bath and in the first wash, carrying iron and hexacyanoferrate with it, and those are the constituents the waste question is actually about. Ware devotes section 7.8 of the Cyanomicon to exactly this and argues that cyanotype processing solutions are dilute and that treating them as chemical waste is a needless consumption of resources; he also notes that if the waste tests strongly acidic it can be neutralised with a cheap alkali such as lime or chalk. His argument is about the process as a whole and is a Tier 2 author’s reasoned position, not a regulatory determination, and the course records it as such.

The course’s own practice is to bottle spent sensitiser and first washes, label the container, and follow the disposal caveat. ILFORD’s guidance for domestic users in the United Kingdom is to take bottled photographic wastes to a household waste and recycling centre. Local regulation governs and it differs, particularly for anyone on a septic tank or a private drainage field, which is the case section 7.8 raises first.

History

The name is much older than the use, and it never meant one thing. Wall’s 1912 Dictionary of Photography has an entry headed “Ammonium Citrate” which prints two formulas under the one heading — the mono- and the diammonium salts — says the article is usually met with as a solution because the solid is so deliquescent that keeping it is extremely difficult, and gives the British Pharmacopoeia’s neutralisation of 12 parts of citric acid with 11 parts of strong ammonia. Its stated photographic uses are as a restrainer in the development of chloride emulsion plates and in the preparation of gelatino-chloride emulsions.

The tribasic salt is not in that entry. Notice what the two eras have in common and where they part. Both make an ammonium citrate out of citric acid and ammonia; both know that the acid and the base have to be brought together in a fixed proportion. But the 1912 darkroom made the salt in the bottle because the solid would not keep, and Ware’s reader buys the solid precisely so as not to open the ammonia — the same neutralisation, approached from opposite ends, a hundred and eight years apart. The family name has outlived every one of the specific substances it has been applied to, which is why the identity section of this page is longer than its chemistry.

How it reached photography. The route is unusually short and unusually well documented. In 1998 a monomeric dicitratoferrate(III) complex was isolated and structurally characterised for the first time, as (NH₄)₅Fe(C₆H₄O₇)₂·2H₂O, its citrate ligands acting tetrabasic and terdentate; Ware summarises that work in §6.7.5 and this page relies on his account of it rather than on the original paper. Inspired by it, he devised the Simple Cyanotype of 2019, which makes that complex in the beaker from citric acid, iron(III) nitrate and ammonia rather than isolating it. Mike’s cyanotype, published in the 2020 Cyanomicon as §7.5, is the last step: the same complex, made by buying the ammonia already combined with the citrate.

So this substance enters photography in 2020, as a convenience rather than a discovery, in a single section of a single book — and it enters the encyclopaedia only because a formula page needed to link it and could not. It was added to the register on 5 September 2026, taking the planned count from 133 to 134.