The chromium policy

Two substances share a word, and one of them is a carcinogen. That is the entire reason this page exists, and it is why no page in this course is permitted to write “chromium” without saying which chromium it means.

The rule was settled on 4 September 2026 and it separates the oxidation states.

	Chromium(VI)	Chromium(III)
In this course	Dichromates and chromic acid	Chrome alum, KCr(SO₄)₂·12H₂O
Status	Never used, at any level, anywhere	Permitted at Level B, as an in-emulsion hardener, where a lesson genuinely requires one
Signal word	Danger	Warning
Pictograms	GHS03, GHS05, GHS06, GHS07, GHS08, GHS09	GHS07
Carcinogenicity notified	H350, may cause cancer	none
Mutagenicity notified	H340, may cause genetic defects	none
Sensitisation notified	H317 and H334	none
UK workplace exposure limit	0.01 mg/m³ as Cr, and 0.025 mg/m³ process-generated	0.5 mg/m³ as Cr
EH40 notations	Carc, Sen, and a biological monitoring guidance value	none

Fifty times on the limit, plus two notations and a biological monitoring value, separates two entries that a reader skimming for the word “chromium” would take for one substance.

Chromium(VI): never, at any level

Dichromates and chromic acid appear across photography’s history: the dichromated colloids that gave carbon, gum and carbro their images; dichromate bleaches and intensifiers; dichromate as a contrast agent in platinotype sensitisers; chromic acid as a laboratory cleaning solution. The course teaches all of that as chemistry and history in Part XXVI, under the Level D policy, with no actionable procedure at any level.

What the classification says. Potassium dichromate’s notified classification, aggregated from 491 reports across 19 notifications to the ECHA C&L Inventory, carries twelve hazard statements: H272 as an oxidiser, H301, H312 and H330 for acute toxicity by three routes, H314 for severe skin burns and eye damage, H317 for allergic skin reaction, H334 for allergy or asthma symptoms if inhaled, H340 for genetic defects, H350 for cancer, H360 for reproductive toxicity, H372 for organ damage on repeated exposure, and H410 for long-lasting aquatic toxicity. Ammonium dichromate’s record is the same shape. There is no reading of that list on which a home darkroom is the right place.

What the exposure limit says. EH40 sets chromium(VI) compounds at 0.01 mg/m³ as chromium over eight hours, with a separate 0.025 mg/m³ entry for exposures generated by a work process such as welding fume, and marks the entry Carc, Sen and BMGV. The biological monitoring value is 10 μmol chromium per mol of creatinine in urine, post-shift — which is to say that the recognised way of knowing whether the control worked is to test the worker. That is what chromium(VI) control looks like where it is done properly, and it is not a thing a domestic darkroom does. The limit was tightened in the 2020 revision, which is a useful reminder that a figure printed in an older book is not necessarily the figure in force.

The reaction hazard is separate from the health hazard. CAMEO’s reactivity profile records ammonium dichromate reacting readily and, in quantity, violently with reducing materials of every kind, detonating on heat or mechanical shock as a solid, and turning its own self-sustaining decomposition into a burst container if confined. NIOSH’s list for chromates is paper, wood, sulfur, aluminium and plastics — the contents of an ordinary workroom.

☠Two consequences of ruling out chromium(VI) that pages must state honestly

The dichromated colloids are not taught with a substitute crosslinker and called carbon printing. Where the course has no dichromate-free route it can source, the process is study-only and the page says the omission is a decision of the course rather than an absence of evidence.

And the dichromate contrast agent in platinotype disappears for a second, independent reason: the platinum sensitiser it went into is itself no longer handled. Two separate rulings remove the same bottle, and a page that mentions only one of them tells half the story.

Chromium(III): chrome alum, permitted at Level B

Chrome alum is potassium chromium(III) disulfate dodecahydrate, KCr(SO₄)₂·12H₂O, the violet salt that hardens gelatin more efficiently than potash alum and that subbed the first dry plates to their glass. It is a different substance from a dichromate, with a different hazard record, and its encyclopaedia entry carries the properties and the period sources.

The classification, with the caveat that goes with it. The notified classification is signal word Warning, GHS07 alone, and three statements: H315, causes skin irritation; H319, causes serious eye irritation; H335, may cause respiratory irritation. No sensitisation, carcinogenicity or mutagenicity statement is notified at all.

That figure has to be read with a piece of record-keeping the course does not hide. PubChem’s page for the dodecahydrate actually sold as chrome alum — CID 24596, CAS 7788-99-0 — carries no GHS classification section whatever. The ECHA notifications for this cation sit under the anhydrous salt, as Chromium potassium bis(sulphate), EC 233-401-6, where 53 reports across 5 notifications give H315 and H319 in 100 per cent of reports and H335 in 84.9 per cent. A second ECHA record for the same material, Sulphuric acid, chromium potassium salt, EC 233-627-5, holds only 4 reports, 3 of which say it does not meet GHS criteria at all. So the classification quoted above is transferred from the anhydrous salt to the hydrate, and the two differ in molar mass by nearly a factor of two — 283.22 against 499.4 — which matters because every photographic formula weighs the hydrate. A classification transfers; a dose does not. The GHS and CLP reference explains why a four-report record and a fifty-three-report record are not the same kind of evidence.

What the limit says. EH40 gives chromium(III) compounds 0.5 mg/m³ as chromium over eight hours, with no short-term limit and no notation of any kind. Chromium metal and chromium(II) compounds carry the same 0.5 mg/m³. EH40’s own caution belongs beside it: the Carc, Sen and Sk notations are not exhaustive, and an unannotated entry is not a clearance.

Where the course uses it, and where it does not. The default in Part V is no hardener in the emulsion at all, because that is what the course’s tested source does. Chrome alum goes into an emulsion only where a lesson genuinely requires an in-emulsion hardener — the coating lesson, which cannot teach hardening without hardening something, and any formula whose published finals already contain it. Where processing needs the layer to hold, the route is bath hardening with potassium alum, which carries no hazard classification at all. Duffin gives the chrome alum dose as 0.5 to 2 per cent of the gelatin weight, stirred slowly into the molten emulsion immediately before coating and effective only near the usual coating pH of about 6.

The Level B controls the dose implies. The hazard is dust while the jar is open, and the control is to open it once: the whole packet is made up into a labelled, dated stock solution in a single operation, and thereafter dispensed by syringe. Gloves and eye protection while the powder is out, weighing done in still air over a tray, and a disposable dust mask as a precaution against a dust nobody has quantified for you.

⚗One period fact that is a chromium(III) fact and not a chromium(VI) one

Kodak’s 1928 primer records that chrome alum loses its hardening power in the presence of sulfite, that its solutions are violet cold and turn green on warming, and that the ammonium salt evolves ammonia if the bath goes alkaline. Those are the properties of a kinetically stable chromium(III) complex changing its coordination, and none of them is a property of a dichromate. Where a historical text says only “chrome”, the behaviour it describes is usually enough to tell you which salt it meant.

Why an undifferentiated rule would be worse than useless

The argument matters more than the rule, because a reader who understands it will apply it to the next case the course has not written yet.

Banning “chromium” over-warns, and an over-warning is not a safe error. It would rule out a hardener whose notified classification is three irritation statements and whose exposure limit is fifty times looser than the one that motivated the ban. It would also force the course to alter historical formulas — Baker’s published finals, Wall’s rule, Duffin’s dose, Baker’s glass subbing dip all contain chrome alum — and to present the alteration as if the sources had been followed. Worse, a reader who checks one of those claims and finds it overstated has been given a reason to discount the next warning, and the next warning may be the dichromate one.

Permitting “chromium” under-warns, which is the failure that hurts someone. A reader who has handled chrome alum at Level B, seen a GHS07 label and a 0.5 mg/m³ limit, and then meets a dichromate formula in a nineteenth-century manual has every reason to think the same controls apply. They do not. The gap between the two is a carcinogen, a respiratory sensitiser and a biological monitoring programme.

So the course says which chromium, every time. A page that names chrome alum says chromium(III) and gives the reason; a page that names a dichromate says chromium(VI) and gives the reason; and neither restates the policy in its own words. This page is the one place the policy lives, and every affected lesson cites it.

Waste

Adopting chrome alum puts chromium(III) into coating-station rinse water and, wherever the hardener goes into an emulsion, into processing and wash water as well. The course states the chemistry and names the classification method; it does not compute a threshold, and it does not give a jurisdiction-specific instruction. The Environment Agency’s waste classification guidance treats chromium compounds under its transition-metal group entries where they are classified as hazardous, and chrome alum is classified — so the assessment is not automatically trivial. The disposal caveat sets out what the course can and cannot say, and what a reader has to check locally. The default of no hardener at all is also the option that puts no chromium anywhere.

What the course could not source

No mechanism for chromium(VI) carcinogenicity. The course has not read an account of it in a source that meets its evidence standard, and does not publish one. What it publishes is the classification, the limit and the notations, each attributed.

The chromium(VI) entry in the mirrored occupational-hazard guide has no extractable text, which is one of three known holes in the research corpus. The figures on this page therefore come from EH40 and from the aggregated ECHA classifications rather than from that card.

No chrome alum dose has ever been tested at domestic scale. Every figure the course found is either a fraction of the gelatin weight, which scales by arithmetic, or a commercial batch quantity, which does not. Part V therefore publishes a design rule and a test rather than a tested recipe, and labels it as one.