Nickel(II) sulfate

Every other toner in this encyclopaedia is bought for a colour. This one is not. The colour change nickel(II) sulfate makes to a cyanotype is, in Mike Ware’s words, “rather slight” — a blue that becomes a slightly greener blue. What the bath actually buys is time: a Prussian blue image that has been soaked in a nickel(II) salt survives an alkaline attack that destroys an untreated one, by about a factor of ten.

Nothing else in the cyanotype literature this course has read defends the image that well against the thing most likely to destroy it, and the course still will not have the salt in a darkroom. The reason is not that nickel is dramatic. It is not: under the harmonised European entry, swallowing it is classified harmful rather than toxic, and the solution is a pale green liquid that looks like nothing much. The reason is that nickel sulfate is a Category 1A carcinogen by inhalation, a respiratory sensitiser and a skin sensitiser — and that the last of those three is not an injury to the print, or even, in the ordinary sense, an injury to the photographer. It is a permanent narrowing of which chemicals that photographer may handle for the rest of their working life.

☠Level D: this page is not a procedure

Nickel(II) sulfate is classified historical study only, and the Level D policy governs what this page may contain. It may give the chemistry, the workflow in outline, the image characteristics, the measured results and the hazards. It may not give a procedure, and it does not: no bath temperature, no immersion time and no sequence of steps appears below as an instruction.

One number does appear, twice, and it is worth saying why. Ware publishes a working strength, and this page repeats it — because the whole argument under Why Level D turns on comparing that strength with a waste-classification limit and with the concentration at which an already-sensitised person reacts to a patch test. A number you cannot see cannot be weighed against anything. It is quoted here as evidence, not as an instruction, and no other part of the method is given.

In photography

The conservation treatment that arrived as a toner

Ware’s Cyanomicon files nickel in section 8.6.3, under Heavy metal incorporation, between the lead toner and the copper toner and immediately after the thallium one. It is a short paragraph, and unusually candid about its own weakness. The colour shift with nickel(II) salts, he writes, “is rather slight — towards a more greenish-blue”. Then the sentence that makes the entry worth a page: “This treatment has the beneficial effect of making the cyanotype much more resistant to alkaline hydrolysis, by at least an order of magnitude (10x) in exposure”. Then the cost: “the application of the nickel solution may cause a slight loss of highlight density”. Then the refusal: “Nickel(II) salts are listed carcinogens. Consult the MSDS.”

Read in the order he wrote them, those four statements invert the usual shape of a toning entry. The colour is the least of it. The reason is set out two dozen pages later, in the chapter on the vulnerability and conservation of cyanotypes, and it is the single most serious weakness of the process: Prussian blue is destroyed by alkali. Ware records, from Holtzman’s work, that a buffer at pH 9.4 completely decolourises Prussian blue in one to ten minutes; that pH 9.4 is the pH of a saturated solution of calcium carbonate, the buffer routinely built into archival papers and boards; and that 0.25 molar sodium carbonate at about pH 10.7 destroys the Prussian blue of a cyanotype in under half a minute. A cyanotype is a photograph that can be killed by the mount it is put on.

Holtzman’s proposed answer was nickel. Ware summarises it: incorporation of nickel ions into the Prussian blue improves its resistance to alkaline environments, best done at the time the blue is formed but achievable retrospectively by soaking the pigment in a bath of a nickel(II) salt. Holtzman found treated Prussian blue “stable indefinitely” at pH 9.4 — or at least for four months, the duration of the experiment — and able to withstand the pH 10.7 carbonate solution for four to five hours.

!What this course has read, and what it has not

The alkali-resistance finding is Holtzman’s, published as Alkali Resistance of the Iron Blues in Industrial and Engineering Chemistry, volume 37, number 9, 1945, pages 855 to 861. This course has not read that paper. Everything attributed to Holtzman above and below is Ware’s account of it, cited to Ware, in the same way that the thallium(I) sulfate page cites Crookes for a solubility that Crookes says he took from Lamy.

Naming the primary source without having read it is not a citation, and pretending Ware’s summary is the paper would be worse. It is recorded here so that anyone extending this entry knows exactly which document is still outstanding.

What Ware measured

Ware tested the treatment himself, on Herschel and Ware cyanotypes, “in case it could offer a possible conservation procedure”. He soaked test strips in nickel(II) sulphate solution, then immersed them in Holtzman’s own pH 9.4 buffer — 0.05 molar in sodium hydrogencarbonate and 0.0165 molar in sodium carbonate — for ten minutes, alongside untreated controls of the major sensitiser types, and re-measured the shadow-tone densities. His losses, in units of 100 × ΔD:

Sensitiser	Initial density	Loss (100ΔD)	Loss %
Smee	0.953	58	61
Herschel	1.355	68	50
Herschel, Ni(II) treated 1 hr	1.352	6	5
Herschel, Ni(II) treated 24 hr	1.122	5	5
Lietze	0.952	65	68
Valenta	1.322	94	71
Ware	1.086	65	60
Ware, Ni(II) treated 1 hr	0.929	12	13
Ware, Ni(II) treated 24 hr	1.335	4	3

Untreated, every sensitiser type in the table loses between half and seven tenths of its shadow density to ten minutes at pH 9.4. Treated, the same two processes lose between 3 and 13 per cent. That is the order of magnitude, and it is the strongest permanence result for any post-treatment in the cyanotype literature this course has read.

Three cautions belong with it, and none of them is in the source.

The initial-density column does not behave like a single strip measured twice. The untreated Ware specimen starts at 1.086 and the 24-hour treated one at 1.335, which is higher — although the text says the treatment costs density. The most economical reading is that these were separate strips of differing initial density rather than one strip before and after, which is consistent with the text’s separate account of the density losses but means the table cannot be used to read the cost of the soak. This is the course’s reading of an ambiguity in the table, not a statement in it.

The citation itself is slippery. The text says the losses “are shown in Table 14”; the caption on the same page reads “Table 9.5 Effects of alkaline hydrolysis at pH 9.4 for 10 minutes”. Two numbers for one table in one document is exactly the kind of thing that turns into an untraceable citation three sources downstream, which is why this page names both.

Two sensitiser types were treated, not five. Smee, Lietze and Valenta appear in the table only as untreated controls. Whether the protection extends to them at the same ratio is not established by this experiment — and Valenta, which loses the most of the five, is the one it would be most useful to know about.

🧮The price, in density, of the protection

Ware states the cost separately from the table. Soaking a cyanotype test strip in nickel(II) sulphate solution for one hour lost 10 units of mid-tone density from the Herschel specimen and 15 after a 25-hour soak; the Ware specimen lost 16 and 21 respectively. The units are the same 100ΔD.

16 ÷ 100 = 0.16 density units

A loss of 16 units, in density

One stop is 0.30 in density, so a change of 0.16 is a little over half a stop — and it comes out of the mid-tones, which is the part of a print a viewer looks at. Set that against what it buys:

60 % loss ÷ 13 % loss = 4.6 times better; and for 24 hours, 60 ÷ 3 = 20 times better

Ware sensitiser, untreated against treated for one hour

So the exchange is a permanent, visible reduction in mid-tone density, plus a hue shift, against a four- to twentyfold reduction in the damage a single ten-minute alkaline exposure does. Whether that is worth making is a conservation judgement about a specific object, not a darkroom preference, and it is the reason the treatment belongs in a chapter on vulnerability rather than in one on toning.

What the nickel ion is actually doing

Ware sets out the mechanism in Appendix II.2, and for nickel it is not quite the mechanism he gives for lead and thallium.

For those two he writes that the toning “is very probably due to cation exchange in the cubic voids of the lattice” — Pb²⁺ and Tl⁺ walking into the sites that hold K⁺ in “soluble” Prussian blue, KFe[Fe(CN)₆]. Both guests are singly charged, or can be: thallium(I) is a heavy potassium, and Ware suggests lead enters as the singly charged hydroxo species PbOH⁺, which is why its incorporation improves in slightly alkaline conditions.

Nickel(II) is not singly charged, and Ware treats it in a different paragraph and for a different reason. “‘Soluble’ Prussian blue may also be rendered less liable to peptize by treatment with a variety of multiply-charged cations, most notably Fe³⁺ and Ni²⁺, which are also believed to be partially exchanged for the K⁺ ions in the lattice. Rendering the Blue ‘insoluble’ is important, because it confers a greater degree of resistance to decomposition by alkalies.”

Three things follow, and the order matters.

The lattice is porous, and that is the whole basis of the chemistry. Prussian blue’s face-centred cubic framework of alternating iron(II) and iron(III) centres bridged by cyanide leaves cubic voids large enough to admit cations of about potassium’s size — which is what the potassium of KFe[Fe(CN)₆]·H₂O is sitting in. In the “insoluble” variety, which Ware describes as a defect version of the ideal lattice, a quarter of the hexacyanoferrate(II) units are absent altogether and water completes the coordination around the iron(III) left behind, so the framework is more open again. The Prussian blue page carries the structure. It is the same porosity that makes the pigment a licensed antidote for thallium and caesium poisoning, which is the loop the thallium(I) sulfate page is built around.

A doubly charged guest cannot be a one-for-one substitute for K⁺. Charge has to balance: one Ni²⁺ entering the lattice must displace two K⁺, or be accompanied by something else. Ware says “partially exchanged”, not exchanged, and does not resolve the stoichiometry; neither does this page. What can be said is that the consequence he reports — a less peptizable, more nearly “insoluble” blue — is what you would expect from replacing two singly charged ions with one doubly charged one that bridges two sites at once, because it is the loose cations in the lattice that let the pigment disperse as nanoparticles in water in the first place. That last sentence is the course’s inference from Ware’s two statements, not a claim either of them makes.

Alkali resistance and peptisation resistance are the same property here. This is the part worth holding on to, because it explains why one treatment does two apparently unrelated things. Prussian blue is destroyed by alkali because hydroxide attacks it and leaves iron(III) oxide-hydroxide behind; the more readily the pigment disperses, the more surface the hydroxide has to work on and the less it behaves as a coherent solid inside the paper fibre. Making the blue harder to peptize therefore makes it harder to hydrolyse. Ware states the link in one sentence and this page has unpacked it; the unpacking is interpretation.

⁇Ware says the nickel soak causes peptisation, and that it prevents it

The two statements are in the same book, twenty pages apart, and they look like a contradiction.

In §9.2 the density lost during the nickel soak is “presumably due to peptisation of the pigment”. In Appendix II.2 Prussian blue treated with Ni²⁺ is “less liable to peptize”. Both are Ware’s.

They can be reconciled, and the reconciliation is a fact about when rather than about whether: a one-hour soak is an hour of immersion in a large volume of water, during which the pigment peptizes as it always does on wet processing — the loss the cyanotype’s own washing problem already accounts for — while the nickel that ends up in the lattice protects what remains from doing so again. On that reading the density loss is the price of the soaking, not of the nickel, and Ware’s own comparison supports it: the 24-hour treated strips are the best-protected in the table.

That reconciliation is this course’s, not Ware’s. He does not address the apparent conflict, and the experiment that would settle it — soaking a control strip in plain water for the same hour — is not in the text this course has read. A reader should treat the two statements as an open question and not as a resolved one.

Nickel in a hexacyanoferrate lattice, four times over a hundred and seventy-six years

The toner is not an isolated curiosity. Put a nickel(II) salt next to a hexacyanoferrate ion and something coloured happens, and the photographic literature stumbled onto that four separate times without ever connecting them.

1844, Robert Hunt. In Researches on Light, chapter IX, Hunt reports that paper coated with nitrate of nickel and ferroprussiate of potash — potassium ferrocyanide — changes “from a delicate light blue to a pea green” after fifteen minutes to half an hour of sunshine. He is describing a nickel hexacyanoferrate coating and a photochemical colour change in it, two years after Herschel published the cyanotype and in the same investigative sweep.

1911, Cassell’s Cyclopaedia. Under “Cobalt, Printing with”, a pale rose image of cobalt ferrocyanide is treated with a second metal salt to change its colour: “By treatment with an iron salt, a blue image is obtained; a nickel salt gives a red image.” That is cation exchange into a hexacyanoferrate framework, used deliberately as a toning step and printed as a recipe rather than as chemistry.

1933 and 1935, H. D. Murray. Ware’s §8.7 sets out Murray’s British and United States patents for a two-tone cyanotype: the first wet bath after exposure contains a metal salt that forms an insoluble ferricyanide with the unreacted ferricyanide still in the sensitiser, so the highlights take a different colour from the blue shadows and the print is fixed at the same time. Murray’s list of metals and highlight colours puts nickel(II) at bright yellow, between cobalt(II) at brick red and zinc(II) at pale yellow.

1945 and after, Holtzman and Ware. Ni²⁺ into the potassium sites of Prussian blue itself, for alkali resistance, as above.

⚗Why Hunt's nickel salt goes blue-green and Murray's goes yellow

The two results look inconsistent and are not, because the anions are different and so is the oxidation state of the iron in them.

Hunt used ferroprussiate of potash, potassium hexacyanoferrate(II) — the ferro compound. The solid that forms with nickel(II) is a nickel hexacyanoferrate(II), and Hunt saw it as a delicate light blue turning pea green.

Murray relied on the ferricyanide left over in a cyanotype sensitiser, hexacyanoferrate(III) — the ferri compound, one electron short. The solid that forms with nickel(II) is a nickel hexacyanoferrate(III), and it is the bright yellow Murray reports.

The general rule behind both is the one the potassium ferricyanide and potassium ferrocyanide pages set out: a hexacyanoferrate anion plus almost any transition-metal cation gives an insoluble, intensely coloured mixed-metal cyanide, and the colour depends on both metals and on the oxidation state of the iron. Prussian blue itself is the case where the second metal is iron(III). Nickel is a much weaker chromophore than iron in this role, which is precisely why the cyanotype toner shifts the hue only slightly instead of replacing it.

The identification of Hunt’s and Murray’s products as nickel hexacyanoferrates is this course’s reading of what the reagents make. Neither source names the compound, and no analysis of either is cited here.

The one place a nickel sulfate bath is named in a nineteenth-century process

Hunt’s Manual of Photography of 1854 sets out his own chromatype, built on Mungo Ponton’s bichromate paper: sulphate of copper and saturated bichromate of potash on paper, printed out in sunshine until the brown gives way to a positive yellow image, then washed over with silver nitrate, which turns it into “a very beautiful positive picture”.

Then, in one sentence: “Mr. Bingham remarks on this process, that if we substitute sulphate of nickel for the sulphate of copper, the paper is more sensitive, and the picture is more clearly developed by nitrate of silver.”

That is, so far as this course’s corpus goes, the only historical photographic formula in which nickel sulfate specifically is named as an ingredient rather than as one option among “a nickel salt”. It is worth three observations.

It is a substitution claim without a mechanism: Hunt reports Bingham’s result and does not explain it, and neither this page nor anything else read for it can say why a nickel(II) salt should sensitise a dichromate coating more than a copper(II) one. Hunt’s own results elsewhere give a hint that the course will not upgrade into an explanation — he found nickel salts on paper photosensitive in their own right, darkening to a light brown, and that nickel nitrate with ammonia gave a coating which “darkens quite as readily as the nitrate of silver”.

It is not a formula the course can reproduce, and the reason has nothing to do with nickel: the oxidising agent is a dichromate. See chromium in the darkroom.

And it is Bingham’s, at second hand. Hunt does not identify him further, gives no formula for the variant, and reports no comparison the reader could check. The course records the claim as a claim.

The salt that was put on a plate and did nothing

Sheppard and Mees found in July 1905 that a plate soaked in uranium nitrate largely loses its sensitiveness, established that the effect belonged to the metal ion rather than to the nitrate, and then tried a list of sixteen more. Nickel is second on it, after lead: “Salts of the following metals were then tried, and gave negative results, though a slight effect was suspected with zinc chloride: — lead, nickel, thorium, chromium, molybdenum, tungstan, ferrous iron, cobalt, silver, zinc, manganese, magnesium, barium, hydrogen (nitric and hydrochloric acids), thallium, platinum.”

A negative result is evidence, and this one does the same work on three pages. It is the sentence that closes the question for thallium and for lead(II) as well, and the three metals of the planned Level D toners entry are the first, second and fifteenth names in it. In the one systematic test the course has read, none of them did anything to a silver halide plate. Whatever nickel does, it does to a pigment made of iron.

Nickel where it is a metal rather than a reagent

The word turns up constantly in period photographic literature and almost never as chemistry. Cassell’s 1911 entry headed NICKEL is about electro-deposition in the graphic arts: zinc etchings nickelled against corrosion, stereotype plates nickel-faced to print better, nickel and nickel-steel electrotypes deposited directly on the wax mould. Burnishers and rolling presses ran polished nickel rollers; Abney notes of one daylight developing machine that “as the apparatus is nickel-plated throughout, it is not corroded by the solutions used”, and suggests his own emulsion-coating box could be made of nickelled iron. Wall’s selenium toner, following Namias, fuses the element with caustic soda in a nickel or iron crucible.

Two of those touch this course’s own pages. Ware warns repeatedly that gold salts attack most metals and will corrode a stainless steel or nickel-plated spatula, which is why the chloroauric acid and gold(III) chloride pages specify plastic. And the Platinomicon describes a genuinely photographic use of a nickel(II) salt as a metal source rather than a colourant: a very dilute palladium sensitiser is exposed to give a latent image of palladium nanoparticles, which is then treated in a bath of about 4 per cent nickel(II), cobalt(II) or copper(II) chloride with 2 per cent sodium hypophosphite as reductant, and the non-noble metal plates out catalytically onto the palladium. The authors of the 1972 patent Ware cites claim “outstanding images which have excellent black rendition and which are very stable”; Ware notes that the contrast is high enough to suit line work rather than continuous tone.

The contrast additive that nobody recommended

One more use recurs and is worth recording precisely because every source that names it also dismisses it. Cassell’s printing-paper emulsion formulae note that “the addition of a small quantity of the chloride of uranium, of nickel or of cobalt, to any of the above formulae will shorten the scale of gradation” — that is, raise contrast — “but the most satisfactory agent is calcium chromate”. Wall’s Photographic Emulsions of 1929 says the same thing twice, once as “greater contrasts can be obtained by adding small quantities of uranium, nickel or copper salts, but they are not so satisfactory as calcium chromate”, and once crediting Valenta for “small additions of uranyl, copper or nickel chlorides … for the hardening of the contrasts in the chloride emulsions”. The ozotype patent Abney reprints allows “the cupric, cobaltus, and nickel salts” in place of manganese salts as metallic salts capable of giving oxides on exposure to light with a chromium sensitiser.

The salt named in all three is the chloride, not the sulfate. That distinction is the reason this page is careful about which nickel salt any given source means, and the reason the register deliberately does not treat nickel chloride as another name for this substance.

Properties

Which substance this entry is, and which one arrives in the jar

The identity here is unusually tangled, and getting it wrong changes a dose by seventy per cent.

The canonical substance is the anhydrous salt, NiSO₄, CAS 7786-81-4, EC 232-104-9, PubChem CID 24586, relative molecular mass 154.76. That is the substance ECHA’s record 100.029.186 is built on, and it carries the CLP Annex VI index number 028-009-00-5 and therefore a harmonised classification rather than only notified ones — the same reason the lead(II) acetate page is the anhydrous acetate and not the trihydrate a supplier ships.

The substance a photographer weighs is the hexahydrate, NiSO₄·6H₂O, CAS 10101-97-0, relative molecular mass 262.85. Ware’s workshop notes name it explicitly in their table of processing solutions: “Nickel sulphate NiSO₄.6H₂O 5% w/v — Dissolve 50 g in 1 litre of water.” It is blue to emerald green rather than yellow-green, less than half as dense, sweet and astringent to taste, and somewhat efflorescent. A heptahydrate also exists, CAS 10101-98-1, relative molecular mass 280.87.

All three are one substance to a regulator and three articles to a balance. ECHA files the two hydrate CAS numbers on the same substance record as the anhydrous salt, and the ILO-WHO card ends its notes with the sentence that settles the safety question completely: “The recommendations on this Card also apply to Nickel sulfate hexahydrate (CAS 10101-97-0) and heptahydrate (CAS 10101-98-1).” The hydrates carry ECHA list numbers of their own — 600-152-3 for the hexahydrate, 600-153-9 for the heptahydrate — rather than EC numbers, and only notified classifications, which is why the harmonised entry sits on the anhydrous salt and the read-across has to be stated in words on the safety card.

🧮Six waters, and the seventy per cent error waiting in them

Start with the two relative molecular masses PubChem gives and check them against each other.

262.85 − 154.76 = 108.09

The water of crystallisation, by subtraction

6 × 18.02 = 108.12

The water of crystallisation, by counting

Agreement to three parts in ten thousand, which is what a correctly assigned hydrate looks like.

Now the two conversions a reader needs.

50 g × (154.76 ÷ 262.85) = 50 × 0.5888 = 29.4 g

What Ware's 50 g of hexahydrate is, as anhydrous salt

262.85 ÷ 154.76 = 1.70

What the same weight of anhydrous salt would be, as hexahydrate

So a reader who takes the “10 % w/v nickel(II) sulphate” of Ware’s §9.2 — which does not say which hydrate — and weighs out the anhydrous salt puts 1.70 times as much nickel into the bath as the same stated strength of the hexahydrate his workshop notes specify. That is not a rounding error. It is the difference between a bath and a bath and two thirds of another.

There is a second, independent way to see which salt the literature means, and it comes out of the toxicology rather than the chemistry. HSDB reports the minimum level of metal needed to provoke a patch-test reaction in nickel-sensitive subjects as “of the order of 112 ppm nickel (0.05% nickel sulfate)”. Take those two numbers as a pair and ask what nickel fraction they imply:

112 ppm ÷ 500 ppm = 0.224

Nickel fraction implied by HSDB's own conversion

0.224 × 262.85 = 58.9

What that fraction weighs, per formula unit of the hexahydrate

Cassell’s Cyclopaedia of 1911 and Wall’s Dictionary of 1912 both print nickel’s atomic weight as 59. So HSDB’s “0.05% nickel sulfate” is the hexahydrate, and it says so only in arithmetic. Had it meant the anhydrous salt, 0.05 per cent would be about 190 ppm of nickel, not 112.

The colour, and why the bath is green when the print goes green

Nickel(II) in water is the hexaaqua ion, and it is green. OpenStax’s coordination-chemistry section puts the reason plainly: ions with a partly filled d subshell usually form coloured complexes, while d⁰ and d¹⁰ ions usually form colourless ones. It illustrates the point with a photographed row of [M(H₂O)₆]ⁿ⁺ solutions in which Ni²⁺ is the d⁸ member, flanked by Co²⁺ (d⁷) and Cu²⁺ (d⁹) — and with the two colourless ends of the same row, Sc³⁺ with no d electrons at all and Zn²⁺ with a full set. Nickel sits in the middle, so its aqua complex absorbs in the visible and the solution is green.

That is worth a moment because it is easy to draw the wrong conclusion from it. The greener blue of a nickel-treated cyanotype is not the green of the bath left behind in the paper. A cyanotype that had merely been stained by residual nickel sulfate solution would wash out. The hue shift Ware measures — from 21D8 to 23D8 in the Herschel specimen and 20E8 to 21E7 in the Ware specimen, in Methuen notation — survives washing, because the nickel is inside the pigment lattice and the absorption that changed is Prussian blue’s own. The two greens have the same cation and completely different origins.

The solution is acidic, and that happens to suit Prussian blue

The ILO-WHO card states it in one line: “The solution in water is a weak acid.” CAMEO agrees from the other direction, filing the substance in the reactive group Salts, Acidic and noting that it gives an acidic solution when dissolved in water. The Merck Index puts a number on the hexahydrate: aqueous solution acid, pH about 4.5.

The mechanism is the one OpenStax’s section on the hydrolysis of salts names in its own learning objectives — the acid ionisation of hydrated metal ions. The sulfate ion is the conjugate base of a strong acid and does essentially nothing; the acidity comes from the cation, whose coordinated water molecules are polarised by the 2+ charge enough to give a proton to the solvent.

[Ni(H₂O)₆]²⁺ + H₂O ⇌ [Ni(H₂O)₅(OH)]⁺ + H₃O⁺

Why a solution of a neutral salt is acidic: the aqua cation is the acid

For this particular application that is a convenience rather than a problem, and it is worth saying why. Prussian blue is destroyed by alkali and tolerates acid — the new cyanotype is developed in dilute acid, and the troubleshooting entry on alkaline bleaching is about what happens at the other end of the scale. A bath at pH 4.5 sits about five pH units below the 9.4 at which Holtzman found the pigment destroyed in one to ten minutes, which is the opposite end of the scale from the pigment’s vulnerability rather than a step towards it. Compare the lead(II) acetate toner, whose incorporation Ware says improves in slightly alkaline conditions — that is, the lead toner works best where the pigment is most at risk, and the nickel one does not.

What takes nickel out of a solution

Nickel(II) is precipitated by the three families of anion that take most divalent transition metals out of solution, and OpenStax’s Appendix J gives the solubility products.

Solid	Ksp at 25 °C	What it means on an alternative-process bench
Ni(OH)₂	1.6 × 10⁻¹⁶	Any alkali throws the hydroxide out of solution
NiCO₃	1.4 × 10⁻⁷	A carbonate bath removes it, though far less completely
NiS (α)	4 × 10⁻²⁰	A sulfide toner removes it almost entirely
NiS (β)	1.3 × 10⁻²⁵	The second form has a solubility product five orders of magnitude smaller again

Appendix J has no entry for the sulfate, and the absence is the information: a solubility product describes a sparingly soluble salt, and this one is freely soluble.

The practical reading is the one both heavy-metal toner pages reach by different routes. A heavy-metal bath and the rest of an alternative-process darkroom cannot share water. But nickel adds a failure mode that lead and thallium do not have, and it runs the other way. Ammonia does not precipitate nickel(II); it dissolves it, replacing the water ligands with ammine ones — the hexahydrate is recorded as very soluble in ammonium hydroxide, and OpenStax’s table of standard potentials lists the hexaamminenickel(2+) couple separately from the aqua one, at −0.49 V against −0.257 V, which is a measure of how firmly the ammonia holds on. A nickel residue that has met ammonia is in solution and will pass through a filter, a rinse and a drain, whereas the same residue met by carbonate is a solid you can see.

Handling

The classification is harmonised, which changes how it should be read. Most GHS blocks in this encyclopaedia are aggregates of what notifying companies chose to say. This one is not. ECHA’s substance record carries the CLP Annex VI index number 028-009-00-5 and lists Harmonised C&L among its regulatory processes, so the classification below is legally binding across the classes it covers. The harmonised classes are Carc. 1A, Muta. 2, Repr. 1B, Acute Tox. 4 * twice — oral and inhalation — STOT RE 1, Skin Irrit. 2, Resp. Sens. 1, Skin Sens. 1, Aquatic Acute 1 and Aquatic Chronic 1, with the signal word Danger. See GHS and CLP for what the class names mean.

Three different marks appear in that entry and none of them is decoration. The Environment Agency’s WM3 carries the clearest table of them this course has read. A single asterisk means a minimum classification for that class — “Actual classification may be higher” — and it is used in waste assessment, so both Acute Tox. 4 entries are floors rather than ceilings. A double asterisk “relates to route of exposure” and is not used in waste assessment; H372 carries it. A triple asterisk is “assigned to reproductive toxicity hazard statements where one attribute is not applicable”, which is why the code reads H360D, may damage the unborn child, rather than H360F or H360FD: the fertility attribute is not being asserted, and its absence is deliberate rather than an omission.

The notified layer splits on exactly the two statements the harmonised entry qualifies. The ECHA C&L Inventory aggregate for EC 232-104-9 runs to 2,582 reports from 26 notifications — a substance in a REACH tonnage band of 100,000 to 1,000,000 tonnes a year is looked at by a great many companies. H334, H317, H341, H372, H400 and H410 sit at or within a whisker of 100 per cent, which is the harmonised set being restated. But 87.2 per cent of notifiers file the unrestricted H350, may cause cancer, against 12.7 per cent for the harmonised H350i, may cause cancer by inhalation; and 85.2 per cent file the unrestricted H360 against 15.3 per cent for H360D. In both cases the majority of industry labels the substance more broadly than Europe’s binding entry does. Japan’s NITE-CMC has classified it five times and its 2024 revision goes further again, raising oral toxicity from Category 4 to Category 3 — toxic rather than harmful if swallowed — and naming the central nervous system, respiratory system, kidney, testis and liver as target organs.

The occupational regimes measure the element, and the lowest limit and the highest differ by a factor of sixty-seven. HSE’s EH40 splits the row: water-soluble nickel compounds as Ni at a long-term limit of 0.1 mg/m³, nickel and water-insoluble nickel compounds at 0.5 mg/m³, neither with a short-term limit. OSHA’s permissible limit for soluble compounds is 1 mg/m³, with a vacated 1989 limit of 0.1 still enforced in some states. NIOSH treats nickel metal and other compounds as a potential occupational carcinogen and recommends 0.015 mg/m³, the lowest of the four, with an IDLH of 10 mg Ni/m³. ACGIH sets 0.1 mg/m³ for the inhalable fraction.

!Two regulators looking at the same substance and reaching opposite verdicts on cancer

The ACGIH threshold limit value for soluble nickel compounds carries the notation A4, not classifiable as a human carcinogen. The ILO-WHO card prints that notation in its occupational-limits box and, in its long-term-effects box on the same card, states: “This substance is carcinogenic to humans if inhaled.”

IARC’s evaluation is the one the harmonised classification follows: “There is sufficient evidence in humans for the carcinogenicity of nickel sulfate”, and, on the group as a whole, “Nickel compounds are carcinogenic to humans (Group 1)”. Metallic nickel is Group 2B, possibly carcinogenic — the metal and its soluble salts are not the same object. NIOSH takes the same view as IARC and recommends the lowest limit of anybody.

Two things are worth taking from the disagreement, and neither is that the science is unsettled.

The first is that a classification is a statement about a route and a form. H350i says by inhalation; the epidemiology behind it is refinery workers breathing dust and aerosol, not people handling a solution. The regime that disagrees is the one that sets an airborne limit for a workplace where that limit is enforced and measured.

The second is that a reader cannot resolve this by picking the source they prefer. In Great Britain the binding classification is the harmonised one, and it says Carc. 1A. In this course, an argument that turns on which of two regulators to believe is an argument that has already failed the test in the safety classification rubric, which asks what controls a domestic reader can actually assemble.

EH40 names this substance, by name, as the sensitiser. The comments column of the nickel row reads “Sk, Carc (nickel oxides and sulphides) Sen (nickel sulphate)”. Read the parentheses carefully. The carcinogen notation on that row is attached to the oxides and sulphides; the sensitisation notation is attached to nickel sulphate specifically, out of the whole family of inorganic nickel compounds. The table’s own heading adds that “The Carc, Sen and Sk notations are not exhaustive”, so the absence of Carc beside the sulphate on that line is not a finding — the harmonised classification says Carc. 1A regardless. What the line does establish is that when a British regulator wanted to name the nickel compound that causes occupational asthma, it named this one.

⚠The number that makes a glove list beside the point

HSDB reports the minimum eliciting level for a patch-test reaction in twenty-five nickel-sensitive subjects as of the order of 112 ppm of nickel — 0.05 per cent nickel sulfate, which the arithmetic above shows to be the hexahydrate.

The toning bath Ware publishes is 5 per cent w/v of the hexahydrate.

5 % ÷ 0.05 % = 100 times the minimum eliciting concentration

The toning bath, in units of a patch test

At the 10 per cent strength of his conservation experiment it is two hundred times. Both figures are about a person who is already sensitised; they say nothing about the dose that sensitises somebody in the first place, which is the hazard the classification is actually about.

Two further facts finish the argument.

This is a substance dermatologists patch-test with. The DrugBank description PubChem carries says flatly that it “is used in allergenic testing”, and HSDB records a North American series of 1972–74 in which three thousand subjects were tested against nineteen allergens and nickel sulfate was among the most frequent sensitisers. Read that the way a course has to read it: the readership of a darkroom manual already contains people who will react to this substance, and they did not acquire the sensitisation in a darkroom.

Sensitisation does not undo. The ILO-WHO card’s instruction to anyone who has shown symptoms is to avoid all further contact with nickel, nickel compounds “and other metal compounds of e.g. copper, chromium and cobalt”. For a photographer that sentence closes copper sulfate, chrome alum and a good deal else, and it closes them permanently. No item of protective equipment reverses it, and the injury is not to a print or to a working session but to the rest of a working life.

The asthma is delayed, which defeats the ordinary way a darkroom notices trouble. The card’s notes say it directly: “The symptoms of asthma often do not become manifest until a few hours have passed and they are aggravated by physical effort. Rest and medical observation are therefore essential.” A control that depends on the user noticing an exposure while it is happening has nothing to work with. The other notes belong with it — do not take working clothes home; isolate contaminated clothing by sealing it in a bag; periodic medical examination is suggested depending on the degree of exposure.

Incompatibilities

The violently reactive list is short. Oxidants are the one the ILO-WHO card names, and it says the reaction “generates fire and explosion hazard” — a stronger phrasing than the thallium(I) sulfate card uses for the same class of partner, where the entry reads only that the substance reacts with strong oxidants. Heat is the second, and it is worth being exact about what comes off, because there is no melting point at all and so nothing evaporates. At 848 °C the salt decomposes, and the ILO-WHO card names the products as sulfur trioxide and nickel monoxide, referring the reader on to two further safety cards, ICSC 0926 and ICSC 1202, for them; CAMEO’s profile adds oxides of nitrogen and metallic nickel to the fume. Anything airborne from a hot sample is therefore a decomposition product and not the salt. The third incompatibility is the substance’s own dust.

The precipitations set out under Properties are the practical incompatibilities, and they matter for the reason the thallium page gives: they are not dangerous reactions, they are the reactions that would silently move the metal from a bath into a sludge, so that a worker who assumed it was still in the beaker would be wrong about where it was. Nickel’s version of that trap is the ammonia complex, which moves it the other way — out of a solid and into a solution that looks like nothing.

The last incompatibility is not chemical. Both the ILO-WHO card and the transport rules require separation from food and feedstuffs, and the card adds that the substance is not to be transported with them. As with lead and thallium, a substance is being treated as incompatible with domestic life.

⚠Why Level D

Argued against the classification rubric rather than asserted.

It passes the acute half of Level A’s first criterion, and that is the trap. Level A admits substances whose GHS classification is “at most irritant, harmful if swallowed, or corrosive at the concentrations actually handled”. Nickel(II) sulfate is Acute Tox. 4 by both oral and inhalation routes: harmful if swallowed, which is the exact phrase the criterion permits. A reader who assessed this substance by its acute toxicity alone would place it at Level A and be wrong about six of its eleven hazard classes. The criterion says “at most”, and Carc. 1A, Muta. 2, Repr. 1B, STOT RE 1, Resp. Sens. 1 and Skin Sens. 1 are every one of them more than that. This is the clearest case in the encyclopaedia of a substance whose danger is invisible to the acute reading, and it is why the rubric’s first criterion has to be read as a ceiling on every class rather than on the acute ones.

It fails Level B, and the reason is instructive. Level B’s list of admitted hazards does include “fine powders that must not be inhaled”, and this substance is one; a mask rated for fine particulates is in Level B’s control list and would be genuinely relevant. But Level B’s other controls — splash goggles, an apron, eyewash within reach, a spill kit — are answers to an event, and nothing here is an event. The asthma arrives hours later, the sensitisation is permanent, and the carcinogenicity is an epidemiological finding about years of exposure. Level B is the right shape for the dust and the wrong shape for what the dust does.

It meets Level C on two criteria, and on the first it meets four of the five named categories at once. Level C’s first criterion names reagents classified as “acutely toxic, carcinogenic, mutagenic, reproductive toxins or sensitisers at the concentrations used, where a fume cupboard or specialist disposal is the recognised control”. This one substance is Carc. 1A, Muta. 2, Repr. 1B, Resp. Sens. 1 and Skin Sens. 1 — carcinogen, mutagen, reproductive toxin and sensitiser twice over — and the recognised control is exactly the engineered one the criterion describes: the ILO-WHO card’s inhalation prevention is “Use closed system or ventilation (not if powder)”, which excludes by name the operation a darkroom would be performing. Level C’s fourth criterion names any procedure whose waste cannot lawfully enter a domestic drain; the Waste section below shows a published toning bath at roughly fifty times WM3’s carcinogenicity limit. On its own hazards this is a Level C reagent.

It is written at Level D because the second Level D criterion is satisfied — but only just, and the gap is worth naming. Level D covers processes for which a safer modern route exists and is taught instead. For the colour, that route plainly exists: the tannic acid toner and the bleach-and-tone chemistry of the toned cyanotype both give a far larger colour change than a slight shift towards green, and both are taught here.

For the permanence, it is thinner, and this page will not pretend otherwise. Nothing else in the corpus protects Prussian blue against alkali by an order of magnitude. What the course teaches instead is avoidance rather than protection — unbuffered paper, unbuffered mounts, unbuffered enclosures, and attention to the wash water — which is what the troubleshooting entry sets out. That is a real answer and a cheaper one, because it costs no density and no hazard; it is not the same answer, because it protects a print from an environment you control and not from one you do not.

So the honest form of the Level D argument here is: the hazard reaches Level C, a home darkroom cannot supply Level C’s controls, and the course has a different and adequate route to the same end. Ware, who published the treatment and measured it, stops at the same place by the same reasoning — he sets the method down, records that nickel(II) salts are listed carcinogens, and tells the reader to consult the safety data sheet. This entry is not correcting its source. It is explaining, at a length a toning chapter has no room for, what that instruction is made of.

Waste

There is nothing to neutralise. An acid can be neutralised because the proton is consumed; a dichromate can be reduced because chromium(VI) becomes chromium(III). Nickel(II) is an element in its stable oxidation state, and every treatment that exists moves it from a solution into a solid that is still nickel waste. HSDB’s disposal entries for this substance are not chemistry at all but the generic precautions for carcinogens — sealed bags, labelled containers, HEPA or charcoal filtration of exhausted air, and the flat statement that “there is no universal method of disposal that has been proved satisfactory for all carcinogenic compounds”. That is what the absence of a route looks like in a reference work.

In Great Britain the assessment runs through the List of Waste’s own vocabulary, and nickel is named twice in it: in the legal definition of “heavy metal” — “any compound of antimony, arsenic, cadmium, chromium (VI), copper, lead, mercury, nickel, selenium, tellurium, thallium and tin … as far as these are classified as hazardous substances” — and again in the definition of “transition metals”. This compound is classified, so both apply.

🧮A published toning bath, measured against WM3's own worked example

WM3 carries a worked assessment of a contaminated soil whose component of concern is nickel carbonate — a different substance, but one whose harmonised classification is the same shape as this one’s: Carc. 1A, Muta. 2, Repr. 1B, STOT RE 1, Acute Tox. 4 * twice, Skin Irrit. 2, Resp. Sens. 1, Skin Sens. 1 and Aquatic Acute 1, with the codes H350i, H341, H360D***, H372**, H332, H302, H315, H334, H317 and H400. Because the concentration limits belong to the hazard statements rather than to the substance, that example prints the numbers this bath would be measured against. Its nickel carbonate sits at 0.06 per cent and clears every one of them.

Hazardous property	Limit WM3’s example applies	Ware’s 5 % w/v bath
HP7 carcinogenic (H350)	0.1 %	about 50 × over
HP10 toxic for reproduction (H360)	0.3 %	about 17 × over
HP11 mutagenic (H341)	1 %	about 5 × over
HP5 STOT (H372)	1 %	about 5 × over
HP13 sensitising (H317, H334)	10 %	under, at about half

The arithmetic is one step. Five grams of the hexahydrate per 100 mL of solution is about 5 per cent by weight, taking a dilute aqueous solution’s density as near enough that of water — a little under 5, in fact, since a litre of it weighs rather more than a kilogram, which makes every ratio above slightly conservative.

5 % ÷ 0.1 % = 50

Against the carcinogenicity limit

Three caveats, because this is arithmetic and not a determination. WM3’s example applies generic concentration limits; a real assessment must check whether a specific concentration limit applies to the substance in hand, which would change the numbers. WM3 assesses a waste, following its own numbered steps, and a spent bath is not the soil in the example. And a domestic reader in Great Britain is not the person WM3 is written for — the household route is a council hazardous waste service, which GOV.UK says exists in England and Wales only.

What the arithmetic does establish, without needing any of that, is the order of magnitude: the sensitising property is the only one of the five that a published toning bath does not engage, and it misses by a factor of two.

The environmental classification is unambiguous in a way the health one is not. The harmonised entry gives H400 and H410, very toxic to aquatic life with long-lasting effects, at Aquatic Acute 1 and Aquatic Chronic 1 — the most severe category in both — and better than 99.9 per cent of notifiers agree. The ILO-WHO card does not hedge either: “The substance is very toxic to aquatic organisms. The substance may cause long-term effects in the aquatic environment. It is strongly advised not to let the chemical enter into the environment”, and its storage instruction is a location rather than a container — an area without drain or sewer access. CAMEO’s summary of the whole nickel sulfate family puts it first among the hazards: “The primary hazard is the threat to the environment.”

🧮One litre of the bath, against a drinking-water guideline

The EPA’s federal drinking-water guideline for nickel is 100 micrograms per litre. Ware’s bath is 50 g of the hexahydrate per litre, and the arithmetic above fixes the nickel fraction of that salt at 0.224.

50 g × 0.224 = 11.2 g = 11,200,000 µg

Nickel in one litre of the toning bath

11,200,000 µg ÷ 100 µg/L = 112,000 L ≈ 110 m³

Water lifted to the guideline by that litre

A hundred and ten cubic metres is a hundred and ten tonnes of water, from one litre of a bath a print was soaked in for an hour. A guideline is not a threshold of harm and a dilution is not a disposal route; the number is here to give a sense of scale to the card’s instruction about drains, which is otherwise easy to read as boilerplate. It is not boilerplate.

Check your local regulations; they govern.

History

Nickel’s photographic history is longer than thallium’s and much thinner than lead’s, and it has an odd shape: an early flurry of genuine light-sensitivity work, a long middle in which the metal appears in photography almost entirely as plating on hardware, and a late, narrow, well-measured use that is about conservation rather than pictures.

• 1844. Robert Hunt, in Researches on Light, gives nickel a chapter of its own. Nitrate of nickel on paper passes into “a fine light brown”, giving “tolerably good negative pictures”; the same salt washed with ammonia — enough to precipitate the oxide, not enough to redissolve it — gives a coating that “darkens quite as readily as the nitrate of silver”; the change is entirely dependent on the rays of greatest refrangibility; iodide of nickel changes quickly; and nitrate of nickel with ferroprussiate of potash goes from a delicate light blue to a pea green in fifteen to thirty minutes of sunshine. He closes: “Many other salts of this metal have been found to afford similar results.”
• 1854. Hunt’s Manual of Photography records Mr Bingham’s substitution of sulphate of nickel for the sulphate of copper in Hunt’s chromatype, with the paper more sensitive and the image more clearly developed by silver nitrate. It is the only historical formula in this corpus that names the sulfate in particular.
• 1876. A report Eder quotes records the general rule behind the toners: metal chlorides acting on silver ferrocyanide give silver chloride and the corresponding metal ferrocyanides, iron chloride colouring the image blue and copper chloride reddish-brown, “colouring with cobalt and nickel salts is also referred to there”. Eder’s later summary lists the ferricyanide methods as toning in blue, green, yellow, orange and brown with lead, copper, uranium, chromium, nickel and cobalt salts.
• 1905–07. Sheppard and Mees try a nickel salt on a plate, second in a list of sixteen, and record a negative result as a desensitiser.
• 1911. Cassell’s Cyclopaedia gives nickel an entry that is entirely about electro-deposition in the graphic arts. Its photographic chemistry appears elsewhere in the volume, unindexed under the metal: a nickel salt turning a cobalt ferrocyanide image red, and nickel chloride among the additives that shorten the scale of gradation in a printing-paper emulsion.
• 1924 and 1929. Wall records “cobalt or nickel chloride for green” among the second baths of a ferricyanide bleach-and-tone sequence, and twice repeats the emulsion-contrast use while recommending calcium chromate instead of it.
• 1933 and 1935. H. D. Murray patents the two-tone cyanotype in Britain and the United States. A nickel(II) salt at 10 to 20 per cent in the first wet bath gives bright yellow highlights and fixes the print at the same time.
• 1945. Holtzman publishes Alkali Resistance of the Iron Blues and recommends incorporating nickel ions into Prussian blue against alkaline attack. The paper is the origin of everything on this page that matters, and this course has read Ware’s account of it rather than the paper.
• 1972. Calligaris, Callaby and Rossello patent the electroless deposition of nickel, cobalt, copper or iron onto a gold, platinum or palladium latent image. It is the one modern use in which a nickel(II) salt supplies the image substance itself rather than modifying somebody else’s.
• The regulatory era. IARC’s evaluation — which PubChem reproduces without a date, and which this page therefore does not date — concludes that nickel compounds as a group are carcinogenic to humans, Group 1, with sufficient evidence in humans specifically for nickel sulfate, while metallic nickel is Group 2B. Regulation (EC) No 1272/2008 gives the substance the harmonised entry it still carries, index 028-009-00-5: Carc. 1A, H350i.
• 2016 and 2020. Ware’s Cyanomicon sets the toner down in §8.6.3, between the thallium and copper toners, and reports his own measurements of the alkali protection in §9.2 — which is where a reader of this course will meet the substance, and the reason for this page.

The register that plans this encyclopaedia had missed it, and the miss is on the record twice already. When lead(II) acetate and thallium(I) sulfate were added on 6 September 2026, both were added because the planned Level D formulary entry lead, thallium and nickel cyanotype toners names three metals and the encyclopaedia could explain none of them. The thallium page closed by saying that it could now explain two, that no nickel salt was registered at all, and that the gap was being recorded rather than left to be discovered. This page closes it. The reason the miss was easy to make is the reason it was easy for glycerol and for thallium: an inventory built by asking what each substance does to silver will not turn up a compound whose only photographic action is on a pigment made of iron — and Sheppard and Mees found in 1905 that a nickel salt did nothing measurable to a silver halide plate.

Like the lead and thallium pages, this one exists so that a reader meeting the name in the historical literature understands what it is — and why the course will not have it in a darkroom.