Thallium(I) sulfate

A cyanotype is a picture made of Prussian blue. Prussian blue is the licensed antidote for thallium poisoning. And thallium(I) sulfate is a toner for cyanotypes — one that works by exactly the mechanism that makes the antidote an antidote, the thallium(I) ion walking into a potassium-shaped hole in the pigment’s lattice and staying there.

That loop is the whole reason this entry exists. Nowhere else in this encyclopaedia has the course met a substance whose photographic use and whose antidote turn out to be the same chemical event, and a reader who understands why the toner works has also understood why a doctor gives a poisoned patient a spoonful of the pigment the print is made of. Everything else about the substance argues for keeping it out of a darkroom: Mike Ware, whose Cyanomicon is the standard modern monograph on the process, describes it as colourless, odourless, tasteless and sufficiently water-soluble to be administered easily to an unsuspecting victim, records that the symptoms do not appear for several days, and writes that a dose of a gram or so is usually fatal because the body cannot get rid of the element, which mimics potassium but blocks its vital enzymic action, especially in the brain.

Where lead(II) acetate is the study-only substance a beginner can still be sold — a cyanotype kit on the market now prints it as an option — thallium(I) sulfate is not sold to photographers at all. A reader meets it in one place: on page 267 of Ware’s Cyanomicon, between the lead toner and the nickel toner, in a paragraph that describes the colour and then refuses to recommend the practice. This page exists so that the encyclopaedia can say what the paragraph is about.

☠Level D: this page is not a procedure

Thallium(I) 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 analytical evidence and the hazards. It may not give a procedure, and it does not: no concentration, no bath temperature, no immersion time and no sequence of steps appears below as an instruction. Ware publishes a working strength; this page says that he does and does not repeat it.

The reasoning is set out against the four criteria of the rubric under Why Level D, in the Handling section. The short form is that the substance meets Level C on its own hazards and is written at Level D because the course teaches a route to the same picture that does not use it.

In photography

The cyanotype toner, which is the only use in which this salt touches a photograph

Ware’s Cyanomicon gives thallium three sentences, and they are worth reading in the order he wrote them. A bath in thallium(I) sulphate solution, he says, “does itself cause a slight but agreeable colour shift in a cyanotype to a ‘cornflower blue’”. Then: “this cannot be recommended as a toning practice for the general public owing to the severe toxic hazard presented by this substance.” A plate later in the same chapter, Figure 8.2, shows four cyanotypes — untoned, lead(II), nickel(II) and thallium(I) — so the colour is not a report at second hand; he made the print.

Three things about that are worth a reader’s attention.

The shift is small. “Slight but agreeable” is doing real work in a book that elsewhere describes the lead toner’s violet as a colour “sometimes rather inaccurately described as ‘lilac’ or ‘lavender’”. Cornflower blue is a bluer blue, not a different hue family. Set against a thallium bath’s hazards, a modest shift within the blues is the least favourable exchange in the whole toning literature, and saying so is part of teaching what a toner is for.

It is the same act as the lead toner, with a different guest. Ware’s Appendix II.2 states the mechanism for both together: “The toning of cyanotypes by Pb²⁺ or Tl⁺ salts is very probably due to cation exchange in the cubic voids of the lattice.” Note “very probably” — this is his inference from the structure and from the X-ray evidence he has for lead, not a measurement on a thallium-toned print, and this course reproduces the hedge rather than quietly dropping it.

Ware’s own general rule predicts which metals work. Under the heading Heavy metal incorporation he conjectures that a metal must have two accessible oxidation states to intervene in the electronic charge-transfer transition that gives Prussian blue its colour: lead (+2, +4), thallium (+1, +3), nickel (+2, +3) and copper (+1, +2) all shift the colour, while alkali and alkaline-earth cations of similar size and charge do not. He labels the conjecture as one. It is worth keeping because it is falsifiable — it says which cations should fail — and because it explains why the family of cyanotype metal toners is a family of heavy metals rather than a cheaper and safer set.

⚗The potassium-shaped hole, and why the same hole is a medicine

Prussian blue is not a dense solid. Its lattice is a cubic framework of iron atoms bridged by cyanide, and the variety the cyanotype process forms is a defect lattice, with systematic vacancies where a proportion of the ferrocyanide groups are missing and water molecules and anions sit in their place. The result is microporous: a framework with holes in it, sized to admit a cation about as big as potassium.

Ware’s footnote to that passage is the whole of the chemistry in one line: the cations of caesium and thallium, Cs⁺ and Tl⁺, are both singly charged and of a size similar to potassium, K⁺. A thallium(I) ion is not a heavy metal ion in the way a lead(II) or a mercury(II) ion is. It carries one positive charge, it forms relatively few complexes except with halogen, oxygen and sulfur ligands, and HSDB summarises the consequence: in this and several other respects thallium(I) has chemical properties similar to those of the alkali-metal cations. It is, chemically, a very heavy potassium.

Everything else follows from that one substitution.

It is why the lattice takes it. Ware sets out at length how Prussian blue was fed to sheep in North Wales, Cumbria and Scotland after Chernobyl, and added to the salt licks of Norwegian reindeer in 1992, to lock up caesium-137 in exactly this way; how the pigment can be eaten in gram quantities and excreted without ill effect, no toxic effect of administering it having been found in animals or humans; and how the German pharmacologist Horst Heydlauf of Karlsruhe proposed it as the preferred antidote for thallium poisoning, “The thallium(I) cation can enter the Prussian blue lattice in place of potassium. There it becomes trapped, and can be safely excreted by the body.”

And it is why the ion is a poison in the first place. The same mistaken identity that gets it through the door of the lattice gets it through the door of a cell: Ware writes that the body cannot rid itself of the element “which mimics potassium but blocks its vital enzymic action, especially in the brain”. A toner and a toxin, by one property.

The affinity is strong enough to be an instrument. Ware records that Prussian blue’s considerable affinity for Tl⁺ has been used to make a sensitive analytical electrode — a “reactrode” — for determining thallium(I) ions in solution. When a course says a pigment has an affinity for a cation, that is usually a qualitative statement; here somebody built a meter out of it.

📷What the course does instead, and how a reader actually meets this substance

The course’s route to a toned cyanotype is the tannic acid toner and the bleach-and-tone chemistry on the toned cyanotype process page, which reach magenta, brown and near-black by converting the pigment rather than doping it. Those give a larger colour change than cornflower blue from a reagent a reader can buy in a food shop.

The comparison with lead(II) acetate is the useful one, because the two substances fail differently. Lead is a live commercial hazard: a cyanotype kit on sale now offers violet tones from either a borax solution or a warm lead acetate solution, prints a working strength, and gives a reader no way to tell the two options apart. Thallium is not sold that way. HSDB records that it has not been used as a rat poison in the United States since 1972 and that the labelling rule for the American product became “for use by government agencies only”, with sale prohibited in some localities; CAMEO records it as not registered as a pesticide in the United States, and the EU pesticides database as not approved.

So the risk here is not that a reader buys it. It is that a reader meets the name in the standard modern monograph on cyanotype — a numbered section, between two toners the same book does recommend — and has nowhere to look it up. The planned Level D formulary entry lead, thallium and nickel cyanotype toners is the page that will treat the three together; this entry is the substance behind one of the three names.

The salt that was put on a plate and did nothing

In July 1905 Sheppard and Mees found that a plate soaked in a solution of uranium nitrate largely lost its sensitiveness. Other nitrates did not do it and other uranium salts did, “so that the action must be ascribed to the metallic ion” — and they then went through the periodic table looking for other metals that would do the same. Their 1907 book prints the result as a list:

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.

The salts that did desensitise were cupric, ferric, mercuric and uranic — copper sulfate at a tenth-molar strength made a plate sixty-four times slower after one minute. Stannous salts reduced the silver bromide outright, platinum bichloride induced fog, and gold chloride produced total reduction on development.

A negative result is worth a paragraph because it bounds a claim. Thallium is a heavy metal that soaks into everything, and it would be easy to assume it must do something to a silver halide plate. The one systematic experiment this course has found says it does not: at whatever concentrations Sheppard and Mees used — they do not give them for the negative list, and this page will not invent them — a thallium salt neither desensitised the plate nor fogged it nor reduced the bromide. The pattern in their positive list is a set of ions with an accessible reduction, and thallium(I) has none at those potentials. That last sentence is this course’s reading of their table, not a statement either author makes.

A dopant a 1984 emulsion patent names, in a list, without saying why

Fuji’s US Patent 4,469,784 describes silver halide emulsions, and its general description carries a sentence that is worth quoting because it is the only modern photographic appearance of thallium the course has found:

The step of precipitating silver halide grains or the step of physically ripening the silver halide grains may be performed in the presence of a cadmium salt, a zinc salt, a lead salt, a thallium salt, an iridium salt or a complex salt thereof, a rhodium salt or a complex salt thereof, or an iron salt or a complex salt thereof.

The patent repeats the same list for doping the core of a core-shell grain, and adds that the amount of metal ions used is usually at least 10⁻⁶ mol per mol of silver halide — one metal ion for every million silver ions, or fewer.

Three things must be said about this, and the third is the important one.

It is a class, not a compound. “A thallium salt” is what the patent says. It does not name the sulfate, and neither does this page.

It is a permission, not a recipe. The sentence reserves a range of dopants the applicant might use. No worked example in the parts of the specification this course has read names a thallium salt: Example 1 precipitates a monodispersed silver bromide emulsion by simultaneous addition at 60 °C with the pH at 6.5 and the pAg at 9.3, and sensitises it with sulfur and gold.

The patent does not say what the thallium is for. It gives a list and a minimum quantity and no mechanism. This course will not supply one, because the general account of what a heavy-metal dopant does to a silver halide grain — deepening or shallowing electron traps, altering the balance between latent-image formation and recombination — is a real body of chemistry that is not in this document and that the course has not yet read a source for. What can be said is the shape of the thing: a dopant works at a level a thousand times below a sensitiser’s, it acts on the crystal rather than on its surface, and it is put in during precipitation because afterwards there is no way to get it inside.

Thallium as light rather than as a reagent: the green line in the darkroom

There is one thing photographers really did use thallium for, and it is not chemistry. It is a wavelength.

Crookes discovered the element in 1861 by its spectrum, and his 1863 Philosophical Transactions paper is largely an argument that the green line is single. He borrowed a nine-prism instrument being built for Gassiot, capable of separating the two sodium lines by an eighth of an inch at ten inches, and under that magnification the thallium line “was still seen single, being as fine and sharply defined upon the black ground as either of the constituents of the double sodium-line.”

A single sharp line at a known wavelength is exactly what a spectroscopist needs to calibrate an instrument, and the darkroom manuals of the next fifty years say so plainly. Wall’s 1912 dictionary tells the reader to set a prism at the position of minimum deviation “for a line of medium wave-length, such as the green thallium line”. Cassell’s 1911 Cyclopaedia, under Monochromatic Light, gives thallium as its example of a source from which a single Fraunhofer line can be isolated — “a single isolated line at λ 5348” — and under Spectrometer gives the working procedure: photograph the lithium, sodium, thallium and blue strontium lines onto a panchromatic plate, using a yellow or green screen for the thallium line, measure the distances between them, and plot wavelength against distance to make an interpolation chart.

That is thallium doing photographic work: not acting on silver, but marking a place on a plate so that everything else on the plate can be measured against it. It is also the first appearance in this encyclopaedia of a substance used purely as a standard, and the sensitometry parts of this course are built on the same idea.

One proposal went further, and it is a curiosity rather than a practice. The 1906 British Journal Photographic Almanac abstracts an English patent of 1904 by Dr G. Krebs for non-explosive flashlights meant to make yellow, blue, green and red filters unnecessary for orthochromatic and trichromatic work: to a magnesium or aluminium powder are added “copper, zinc, lithium, strontium, barium, soda, calcium, caesium, rubidium, thallium, iridium, or other metals or their salts”, so that the colours of objects are correctly reproduced on ortho- or panchromatic plates. The course has read the almanac’s abstract and not the patent, and records the idea rather than endorsing it: fill the gaps in a flash powder’s spectrum with metal emission lines and you may not need a filter. Thallium is on the list because its line is the strongest single green available.

🕰Crookes's corpse-green room, 1863

The most photographically useful sentence in Crookes’s paper is not about photography at all. Having made a flame strongly coloured with thallium by passing hydrogen over the chloride at a high temperature and igniting it, he describes what the room looked like:

The absolute monochromatic character of the light renders everything illuminated by it either green or jet black: coloured sealing-wax, ribbons, a bouquet of flowers, as well as coloured precipitates, are entirely altered in appearance, whilst the human face assumes a horrible, corpse-like green hue.

Read that beside a modern safelight and it is the same phenomenon stated from the other side. A monochromatic source does not dim the colours of the objects it lights; it deletes them. Every surface returns some fraction of the one wavelength on offer and nothing else, so the scene collapses to a single hue plus black — which is why a darkroom under a deep red safelight is not a red-tinted room but a room in which colour has stopped existing, and why judging a print’s tone under one is a skill rather than a glance.

Crookes was not a bystander to photography. Eder’s history records him taking up spectrography of the ultraviolet on wet collodion plates in 1854, prompted by Stokes’s 1852 work on fluorescence; the 1855–56 investigations of silver bromide’s response to coloured light that Eder calls the most comprehensive of their period; and his introduction of a quinine sulphate filter in front of the lens to absorb the ultraviolet and part of the violet, which Eder credits as the pioneering application of light filters in photography. Neither source connects the two careers — Eder does not mention thallium, and the thallium paper does not mention photography — but the thallium paper was communicated to the Royal Society by G. G. Stokes, the same man whose fluorescence work Eder says sent Crookes to the ultraviolet. This course reads them as one man on that evidence, and flags the identification as its own rather than as something either source states.

Properties

Which substance this entry is

The canonical entry is the simple sulfate of thallium in its +1 oxidation state: Tl₂SO₄, relative molecular mass 504.83, CAS 7446-18-6, EC 231-201-3, PubChem CID 24833, ECHA CHEM record 100.028.365, UN number 1707, ILO-WHO safety card 0336. PubChem’s computed IUPAC name is bis(thallium(1+));sulfate and ECHA’s is the more forbidding bis(λ¹-thallanylium) sulfate; the names a source actually prints are thallium(I) sulfate, thallous sulfate, dithallium sulfate and, in the older literature and on the ILO-WHO card’s synonym line, simply thallium sulfate.

Two identifier traps are worth recording.

A second CAS number, 10031-59-1, is filed as the “parent”. PubChem lists it among the CAS numbers of the same record and its Related CAS field reads “10031-59-1 (Parent)” and “7446-18-6 (di-thallium(+1) salt)”; HSDB adds the note “RN given refers to unspecified Tl salt”. Three further numbers — 14012-92-1, 37333-30-5 and 87993-82-6 — are marked deprecated, and all three also appear on the ECHA record. The identity check for this substance is therefore made on 7446-18-6, which is the number the ILO-WHO card, the CLP Annex VI entry and the EC number all agree on.

“Thallic sulphate” is a different substance, and European law prints it as a synonym for this one.

!The Annex VI entry calls a thallium(I) salt by the thallium(III) name

ECHA’s substance record for EC 231-201-3 carries, in its process-related name field — the field that holds the name a regulatory instrument uses — the string “dithallium sulphate; thallic sulphate”. That is the CLP Annex VI form, and PubChem files the harmonised classification under it. The French, Romanian and Hungarian translations pair the same two names.

By the ordinary convention of the -ous/-ic suffixes, which HSDB states for this very element — thallium “exists in two oxidation states, thallous (Tl(I)) and thallic (Tl(III))” — thallic sulphate means Tl₂(SO₄)₃, thallium(III) sulfate, PubChem CID 27758, CAS 16222-66-5, relative molecular mass 697.0. That is a different compound with a different CAS number, a different formula and a different molar mass.

The course records the fault rather than smoothing it over, because a reader who searches on the Annex VI name will find a legally binding classification attached to a name that denotes something else. thallic-sulphate is deliberately not registered as an alias of this entry: a page that names it should fail, so that somebody decides which oxidation state is meant.

The three senses that fail

Sugar of lead announces itself. It is sweet, that is where its name comes from, and although tasting it is the last thing anybody should do, the sweetness is at least a property a person could in principle notice.

Thallium(I) sulfate has no such property. The ILO-WHO card gives white or colourless crystals; CAMEO gives white rhomboid prisms or a dense white powder, odourless; Ware’s summary is that it is “colourless, odourless, tasteless, and sufficiently water-soluble to be administered easily to an unsuspecting victim”. None of the three senses that would warn a person about a contaminated bench or a mislabelled jar registers anything at all, and a solution of it in water is indistinguishable by eye from water.

The one property a hand would notice is density: 6.77, against about 2 for most of the salts in this encyclopaedia. A jar of it is startlingly heavy for its size, and the melting point of 632 °C belongs to a substance that is nothing like an organic reagent.

The delay is the other half of the problem. The ILO-WHO card states that the effects may be delayed and that medical observation is indicated, and adds under Notes that “the symptoms of neurological disorders do not become manifest until after a few days”. Ware makes the historical point that the symptoms could be mistaken for encephalitis or epilepsy, with sudden alopecia the clue that gave the diagnosis away. A hazard whose first sign is hair loss ten days later is not a hazard a darkroom routine can be built around.

Solubility, and the anions that take thallium out of solution

The salt is soluble, though not extravagantly so: 4.87 g per 100 mL at 20 °C, which the ILO-WHO card itself labels “moderate”, rising to 18.45 g at 100 °C and falling to 2.70 g at 0 °C. In solution it is simply a pair of free cations and a sulfate ion:

Tl₂SO₄ → 2 Tl⁺ + SO₄²⁻

The salt exists in water only as the free thallium(I) ion, which is the species that does everything on this page

Which means that everything the sulfate does — the toning, the poisoning, the trapping in the lattice — is done by Tl⁺, and the sulfate is only the counter-ion that made it soluble. The choice of sulfate is not chemistry, it is convenience: it is what crystallises when the metal is dissolved in sulfuric acid, which is how Crookes made it and how HSDB says it is still made.

That also says which reagents would destroy a thallium bath, and the list reads like a darkroom inventory. Appendix J of Chemistry 2e gives solubility products at 25 °C for four thallium compounds and none for the sulfate, which is itself the point — the sulfate is soluble enough not to need one:

Compound	Ksp at 25 °C	What it means in a darkroom
TlCl	1.7 × 10⁻⁴	Any chloride — a salting solution, sodium chloride, the chloride a fixer leaves behind — throws thallium(I) chloride down. About 3 g per litre against the sulfate’s 48.7
TlSCN	1.6 × 10⁻⁴	A thiocyanate gold toner would precipitate it just as readily
Tl₂S	6 × 10⁻²²	A sulfide toner removes it almost completely. HSDB’s note that “in reducing environments, thallous may precipitate as a sulfide” is the same fact in a river
Tl(OH)₃	6.3 × 10⁻⁴⁶	Thallium(III), not this substance, and listed here only because it is what an oxidising bath would make of the ion

The practical reading is the one the lead(II) acetate page reaches by a different route: a heavy-metal toning bath and the rest of an alternative-process darkroom cannot share water. The difference is that lead is thrown out by carbonates, phosphates, sulfates and citrates as well, whereas thallium(I) — being a heavy potassium, and forming few complexes outside the halogen, oxygen and sulfur ligands — is precipitated by a shorter and more specific list.

🧮Crookes's solubility, checked against a safety card 150 years later

Crookes gives the solubility of the sulphate in the units of his period: the salt “is soluble in twenty or thirty times its weight of cold water, and in much less when boiling”. Converting is one division.

100 ÷ 20 = 5.0 g per 100 g; 100 ÷ 30 = 3.3 g per 100 g

One part in twenty parts, and one part in thirty

So Crookes’s range is 3.3 to 5.0 g per 100 g of cold water. The ILO-WHO card of April 2013 gives 4.87 g per 100 mL at 20 °C, which sits at the soluble end of his range; HSDB’s 2.70 g at 0 °C sits just below it, which is what “cold water” not meaning ice-water would predict. His “much less when boiling” is 18.45 g at 100 °C, a factor of about four.

Two things are worth taking from the agreement. The first is that a careful nineteenth-century measurement, stated in parts by weight, survives translation into modern units without embarrassment — which is a reason to read the old literature rather than assume it. The second is a caution about citation: Crookes is not the source of the number. He says so himself — “I have little to add to M. Lamy’s description of this salt” — and the course has read Crookes and not Lamy.

It is neither an oxidiser nor much of a reducer, and that is the difference from the other heavy metals

CAMEO puts thallium(I) sulfate in the reactive group Non-Redox-Active Inorganic Compounds and prints a profile that reads, in full: “THALLIUM SULFATE has weak oxidizing and weak reducing powers. Redox reactions can however still occur.” The ILO-WHO card names only strong oxidants as reactive partners.

That is a real distinction and not a technicality. Lead(II) nitrate is an oxidising nitrate and its page carries a ferricyanide bleach; potassium dichromate and potassium permanganate are in the course because of what they take electrons from; mercury(II) chloride intensifies a negative by oxidising the silver image. Thallium(I) does none of that. It has no redox business with silver or with iron at the potentials a photographic bath reaches, which is why the only thing it does to a picture is move into a hole in the pigment and change the colour of the light coming out.

That last sentence is this course’s inference from CAMEO’s reactivity classification and Ware’s mechanism, not a statement in either source, and it is the kind of inference a reader should be able to test: if it is right, a thallium bath should leave a silver image entirely alone, and the finding of the Part XX page on the toners we study and do not use — that no thallium toning formula for a developed silver image appears anywhere in the corpus behind it — is consistent with it. Consistent is not the same as confirmed.

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. The ECHA substance record carries the CLP Annex VI index number 081-003-00-4 and lists Harmonised C&L among its regulatory processes, so the classification below is legally binding across the classes it covers and the notified spread sits on top of it rather than in place of it. The harmonised classes are Acute Tox. 2 *, STOT RE 1, Skin Irrit. 2 and Aquatic Chronic 2, with the statements H300, H315, H372 ** and H411 and the signal word Danger.

The two marks are not decoration and they do not mean the same thing. WM3’s table of qualifications is the clearest statement of both this course has read. A single asterisk means a minimum classification for that class: “Actual classification may be higher”, it is applied to acute toxicity and to repeated-exposure STOT, and it is used in waste assessment — WM3’s step 4 obliges anyone classifying a waste to go and look for data that would justify a more severe classification. A double asterisk “relates to route of exposure” and is not used for waste assessment. So Acute Tox. 2 on this substance is a floor and not a ceiling; H372’s double asterisk is a note about how the damage is done, not about how bad it is.

The notified layer adds the routes the harmonised entry leaves out. The ECHA C&L Inventory aggregate for EC 231-201-3 runs to 90 reports from 7 notifications. H300, H315, H372 and H411 are at 100 per cent, which is the harmonised set being restated. Below that, three additions matter: H319, serious eye irritation, at 24.4 per cent; H311, toxic in contact with skin, at 21.1; and H361, suspected of damaging fertility or the unborn child, at 20. Japan’s NITE-CMC has classified the substance three times, in 2006, 2016 and 2021, and adds H370 and H410 and a list of target organs the European entry does not print — nervous system, respiratory organs, cardiovascular system, gastrointestinal tract, liver, kidney and skin for a single exposure, and nervous system, skin and “genetic organs (men)” for repeated exposure. Safe Work Australia’s HCIS entry matches the harmonised four exactly.

The occupational regime measures the element, and every one of its limits carries a skin notation. HSE’s EH40 lists “Thallium, soluble compounds (as Tl)” with no CAS number at all, a long-term exposure limit of 0.1 mg/m³, no short-term limit, and the Sk notation. NIOSH’s recommended limit and OSHA’s permissible limit are the same 0.1 mg/m³ with [skin]; the ACGIH threshold limit value is stricter at 0.02 mg/m³ for the inhalable fraction, also with a skin notation; the IDLH is 15 mg/m³ as Tl. The NIOSH entry’s target-organ list is worth reading to its end: eyes, respiratory system, central nervous system, liver, kidneys, gastrointestinal tract — and body hair.

⚗Why a glove list would be the wrong instrument

Four features of this substance defeat the ordinary darkroom control set, and they are worth naming separately because each defeats a different control.

It goes through skin. The ILO-WHO card states the route without qualification: “The substance can be absorbed into the body through the skin and by ingestion.” Three separate occupational regimes attach a skin notation to the limit, which is the regulator’s way of saying that measuring the air does not measure the dose.

It is invisible on a surface. Colourless, odourless, tasteless, dense and soluble. A dried splash on a bench is a faint white bloom that could be anything; a spill into a tray of water is nothing at all.

It becomes airborne easily. “A harmful concentration of airborne particles can be reached quickly when dispersed, especially if powdered”, says the card, and its prevention column leads with PREVENT DISPERSION OF DUST. Weighing a fine powder in a room is exactly the operation that does this.

Nothing tells you it has happened. The effects may be delayed; the neurological symptoms do not appear for several days; HSDB records urinary thallium still detectable in a non-fatal case more than a month after the last exposure, and thallium excretion persisting for many weeks. A control that depends on the user noticing an exposure and responding to it has no signal to work with.

Contrast the ordinary Level A argument, which is that gloves and glasses interrupt the route by which the substance reaches you and a splash tells you at once when they have failed. Here the route is not interrupted by anything a darkroom has, and there is no splash to notice.

Incompatibilities

The reactive list is short, because a non-redox-active salt of a soft cation has few violent partners. Strong oxidants are the one the ILO-WHO card names. Heat is the second: decomposition on heating gives toxic fumes including thallium and sulfur oxides, and since the salt melts at 632 °C and has no appreciable vapour pressure, anything airborne from a hot sample is decomposition product rather than the salt itself. The third is the substance’s own dust.

The precipitations set out under Properties are the practical incompatibilities — chloride, thiocyanate and sulfide — and they matter for a reason that is easy to state backwards. They are not dangerous reactions. They are the reactions that would silently remove thallium from a bath and put it into a sludge, so that a worker who assumed the metal was still in the beaker would be wrong about where it was. With a substance whose whole hazard is that you cannot tell where it is, that is a failure mode worth naming.

The last incompatibility is not chemical. Both the ILO-WHO card and the transport rules require separation from food and feedstuffs, and the card’s Notes add “Do NOT take working clothes home”. A substance is being treated as an incompatibility with domestic life.

⚠Why Level D

Argued against the classification rubric rather than asserted.

It fails Level A on the first criterion, twice over. Level A admits substances whose GHS classification is at most irritant, harmful if swallowed, or corrosive at the concentrations actually handled. “Fatal if swallowed” is two categories above “harmful if swallowed”, and Acute Tox. 2 carries the asterisk that makes it a floor. The qualifying phrase about concentration does not rescue it either: a solid weighed on a balance is at 100 per cent whatever the bath is at, and the operation Level A cannot supervise is the weighing.

It fails Level B, and the reason is instructive. Level B is built around concentrated reagents, heat, mists, fine powders and splashes — hazards whose failure mode is a burn or a spill that announces itself. It does list “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 here. But Level B’s other controls — splash goggles, an apron, eyewash within reach, a spill kit — are all answers to an event, and thallium’s failure mode is not an event. It is a body burden accumulating from a route nobody observed, with the first sign arriving days later. Level B is the right shape for the dust and the wrong shape for the poisoning.

It meets Level C on two criteria. Level C’s first 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. Acute Tox. 2 is acute toxicity by the plainest reading, HSDB describes soluble thallium compounds as skin sensitisers, and the ILO-WHO card’s own spill procedure calls for a fitted respirator and full protective clothing. Level C’s fourth criterion names any procedure whose waste cannot lawfully enter a domestic drain and needs a licensed route; the harmonised H411, the card’s instruction to store the substance where there is no drain or sewer access, and the American classification of the discarded chemical as an acute hazardous waste all meet it. On its own hazards, this is a Level C reagent.

It is written at Level D because the second Level D criterion is satisfied. Level D covers processes for which a safer modern route exists and is taught instead, with the historical method described for understanding only. The single photographic use has one: for a toned cyanotype, the tannic acid toner and the bleach-and-tone chemistry of the toned cyanotype process page, both of which the course teaches and both of which give a larger colour change than the shift under discussion.

A home darkroom cannot supply Level C’s controls, and the course therefore gives no procedure. It is worth saying that Ware, who published the method, reaches the same conclusion by the same route: he describes the colour and then declines to recommend the practice to the general public. This entry is not correcting its source. It is explaining, at greater length than a toning chapter has room for, what the refusal 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). Thallium(I) is an element in its stable oxidation state, and every treatment that exists moves it from a solution into a solid. HSDB records what the treatments actually achieve for soluble thallium compounds, and the numbers are not encouraging: granular activated carbon after precipitation, 84 per cent removal; lime, 54 per cent; ferric chloride, 30; alum, 31. It also states that thallium sulfate is a poor candidate for incineration, which closes the other obvious route.

In Great Britain the classification runs through the List of Waste’s own vocabulary. Its legal definition of “heavy metal” is “any compound of antimony, arsenic, cadmium, chromium (VI), copper, lead, mercury, nickel, selenium, tellurium, thallium and tin, as well as these materials in metallic form, as far as these are classified as hazardous substances” — thallium is named, and this compound is classified. The assessment then has the harmonised entry to work from, with the asterisk on Acute Tox. 2 obliging the assessor to look for data that would make it worse rather than take the Annex VI figure as final.

The environmental classification is where the sources disagree most and it is worth being exact. The harmonised entry gives H411, Aquatic Chronic 2 — toxic to aquatic life with long-lasting effects. Notifiers add H400, very toxic to aquatic life, at 21.1 per cent. Japan’s NITE-CMC puts it a whole category higher again, at Aquatic Acute 1 and Aquatic Chronic 1 with H400 and H410. The ILO-WHO card does not hedge at all: the substance is very toxic to aquatic organisms and may cause long-term effects, and “Do NOT let this chemical enter the environment.” The ecotoxicity data behind that are avian rather than aquatic — oral LD50 values of 23.7 mg/kg in the ring-necked pheasant and 36.7 mg/kg in the mallard, which are the numbers a rodenticide’s registration generates.

In the United States the discarded commercial chemical is EPA hazardous waste P115, and PubChem’s gloss on that number is worth quoting because it is a category and not just a code: “An acute hazardous waste when a discarded commercial chemical product or manufacturing chemical intermediate or an off-specification commercial chemical product or a manufacturing chemical intermediate.” Lead(II) acetate is U144 on the same list of lists, without that word. Thallium(I) sulfate is also a CERCLA hazardous substance with a reportable quantity of 100 lb, and an extremely hazardous substance with a threshold planning quantity of 100 or 10,000 lb. The federal drinking-water standard for thallium is 2 micrograms per litre.

🧮One gram, read by two regimes

CAMEO states that “the mean lethal dose in an adult is probably about 1 gm of thallium sulfate”, and Ware independently says that “a dose of a gram or so is usually fatal”. Take that gram and ask the other question about it — what it would do to water.

The salt is Tl₂SO₄ at a relative molecular mass of 504.83 (PubChem). Wall’s 1912 dictionary and Cassell’s 1911 Cyclopaedia both give thallium an atomic weight of 204, which is the figure Lamy published and Crookes was arguing about.

(2 × 204) ÷ 504.83 = 408 ÷ 504.83 = 0.808, that is 80.8 %

Thallium by mass

So one gram of the salt carries about 808 milligrams — 808,000 micrograms — of thallium. The US federal drinking-water standard for thallium is 2 micrograms per litre.

808,000 µg ÷ 2 µg/L = 404,000 L ≈ 400 m³

Water lifted above the standard by one gram of the salt

Four hundred cubic metres is four hundred tonnes of water. The same gram is, on one reading, a lethal adult dose, and on the other, enough to put a small reservoir over a regulatory limit. Neither number is a hazard statement on its own — a dose is not a concentration and a limit is not a threshold of harm — but together they say why the ILO-WHO card’s storage instruction is a location rather than a container: store it in an area without drain or sewer access.

None of that is a domestic route, because there is no domestic route. GOV.UK points householders at council hazardous waste services and states on the page itself that these exist in England and Wales only. Check your local regulations; they govern.

History

Thallium has the shortest photographic history of any substance in this encyclopaedia, and it starts later than photography does — later, in fact, than the process it tones. When Herschel published the cyanotype the element did not yet exist as a known thing, and nothing in the record below is earlier than 1861.

• 1861. William Crookes finds a green line in the spectrum of material recovered from the sulphur of a sulphuric-acid works. “When I discovered thallium two years ago”, he writes in the paper the Royal Society received on 5 February and read on 19 February 1863.
• 1862. M. Lamy, in France, isolates the metal and publishes its properties, including the description of the sulphate that Crookes says he has little to add to, and an atomic weight of 204 in the Comptes Rendus of 8 December.
• 1863. Crookes’s On Thallium appears in the Philosophical Transactions, volume 153, pages 173 to 192 — twenty pages that are, among other things, an argument with Lamy about the atomic weight. Crookes weighs the sulphate and converts it four ways, to the iodide, to sulphate of baryta, to the chloride and to the platinochloride, and gets values from 201.85 to 203.55; he complains that Lamy published 204 without giving a method, a number of experiments or a spread. The same volume carries Bunsen and Roscoe’s Photo-chemical Researches Part V.
• 1904. Dr G. Krebs patents a flash powder in which thallium is one of a dozen metals whose salts are to be added so that colours reproduce correctly on orthochromatic and panchromatic plates.
• 1905–07. Sheppard and Mees try a thallium salt on a plate, along with fourteen other metals, and record a negative result.
• 1911–12. Cassell’s Cyclopaedia and Wall’s Dictionary both give the green line as a wavelength standard, one for a monochromatic source and an interpolation chart, the other for setting a prism at minimum deviation.
• The rodenticide decades. The element’s dominant use has nothing to do with photography. HSDB lists rat poison, ant bait and mammal control, at 0.5 to 2.0 per cent by weight in a paste or on grain, under a shelf of trade names — Zelio, Ratox, Eccothal, Rattengiftkonserve. Between 1935 and 1955, 778 people were reported poisoned by thallium-containing insecticides, rodenticides and therapeutic chemicals, with 46 deaths. HSDB records a cluster of six poisonings in one district between June and October 1976, traced by epidemiological investigation to a 3 per cent thallium sulfate solution used in homes as a rodenticide; every case the summary describes individually is a child between three and five years old, and one of them was operated on for a suspected brain tumour and died.
• 1972 onward. Thallium sulfate stops being used as a rat poison in the United States; the labelling rule becomes “for use by government agencies only” and sale is prohibited in some localities. It is not an approved EU pesticide.
• About 2000, and 2016 and 2020. Ware’s Cyanomicon sets the toner down in print, between the lead toner and the nickel toner, with the colour named and the practice refused — which is where a reader of this course will meet it, and the reason for this page.

The register that plans this encyclopaedia had missed the substance, and the miss was the ordinary one: 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. It was added on 6 September 2026, the same day as lead(II) acetate and for the same reason — the planned Level D formulary entry lead, thallium and nickel cyanotype toners names three metals, and the encyclopaedia could explain none of them. It can now explain two; no nickel salt is registered at all, and that gap is on the record here rather than left to be discovered by whoever writes the toners entry.

Like the lead page, 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.