Fizeau's gilding solution

Fizeau did not publish a toner. He published a way of making a picture stay on a plate, under a heading that says so — Note sur un moyen de fixer les images photographiques — and the two problems he names in his first sentence are fixing the proofs and giving the lights of the picture more intensity. The bath that came out of it is the first gold toner in photography, the ancestor of every gold formula in this book, and the reason most surviving daguerreotypes can be looked at rather than only stored. It is also, as nobody knew until Hardwich worked it out fifteen years later, two reagents rather than one.

Dissolution de chlorure d'or — the gold solution — the gold, as gold(III), and the solution that is poured

Ingredient	Quantity	Form the source specifies
Chloroauric acid	1 g	Printed as "un gramme de chlorure d'or". Which gold chloride is not stated and cannot be recovered; the course reads the photographic literature's "gold chloride" as chloroauric acid, and the provenance note and The mechanism both set out what turns on the choice.
Water	to make 500 mL	"un demi-litre d'eau pure" — a make-up volume, and Fizeau specifies pure water. At 1 g in 500 mL the stock is 0.2 per cent w/v, which is the figure Ware reads independently out of Gaudin's 1844 report of the same method.
Fizeau states no temperature, no vessel and no keeping time for either stock. Both are half a litre, so the mixed bath is a litre, and every strength quoted on this page is computed against that litre.

Dissolution d'hyposulfite de soude — the hyposulfite solution — the reducing agent and the ligand, and the solution that receives

Ingredient	Quantity	Form the source specifies
Sodium thiosulfate pentahydrate	3 g	Printed as "trois grammes d'hyposulfite de soude". The crystals are what was sold, and the course reads the nineteenth-century "hyposulphite of soda" as the pentahydrate throughout; all the molar arithmetic on this page uses 248.19 g/mol and says so.
Water	to make 500 mL	"un demi-litre d'eau pure". At 3 g in 500 mL the stock is 0.6 per cent w/v, which is again Ware's figure from Gaudin.
This is the vessel the gold goes into, and the order is not a convenience. Thiosulfate has to be in local excess at every moment of the addition, and the whole of the mixing instruction exists to keep it so.

Solution d'hyposulfite pour le lavage — the fixing bath, renewed for every plate — the bath the plate must pass through before it can be gilded

Ingredient	Quantity	Form the source specifies
Sodium thiosulfate pentahydrate	100 g	Fizeau gives a proportion and not a weight — "environ une partie de sel pour quinze d'eau", about one part of salt to fifteen of water. The 100 g and the 1500 mL below are the course's scaling of that proportion, chosen because the schema needs numbers; Fizeau scaled nothing.
Water	1500 mL, added	Fifteen parts of water to one of salt. This is a proportion of water rather than a make-up volume, so no final volume and no per cent is stated for it anywhere on this page. Fizeau's own condition on this bath is not its strength but its freshness.
Recorded because Fizeau publishes it as part of the same note and because the gilding does not work without it, not as a bath for anyone to make. Its modern descendant is the plain hypo fixing bath, at about six times the strength.

Mixed in this order — making the gilding bath

1. Start with the whole of it, in the receiving vessel of Dissolution d'hyposulfite de soude — the hyposulfite solution — The thiosulfate solution is what receives. Reversing this is the one change that destroys the formula.
2. Then add the whole of it, poured in little by little of Dissolution de chlorure d'or — the gold solution — "peu à peu et en agitant" — little by little, and agitating. Hunt's 1854 translation renders it "a little by little, agitating between each addition", which is the same instruction with the pause made explicit.

On verse alors la dissolution d'or dans celle de soude, peu à peu et en agitant; la liqueur mixte, d'abord légèrement jaunâtre, ne tarde pas à devenir parfaitement limpide. — The gold solution is then poured into the soda solution, little by little and with agitation; the mixed liquor, at first slightly yellowish, soon becomes perfectly limpid.

Equal volumes, because both stocks are half a litre, so the bath is one litre carrying 1 g of gold chloride and 3 g of hypo. Fizeau states the direction of addition and gives the colour change as its confirmation; he does not state a ratio, and the page does not invent one.

Used in this order — the washing Fizeau requires before the plate is gilded

1. Water — A few drops of alcohol are put on the still-iodised surface first, and when the alcohol has wetted the whole face the plate is plunged into a basin of water. Fizeau states the reason and it is a wetting one — the alcohol makes the water adhere over the whole plate and stops it withdrawing at the edges during each immersion, which would infallibly produce stains.
2. Solution d'hyposulfite pour le lavage — the fixing bath, renewed for every plate — no time is published; the bath is renewed for every plate — The ordinary hyposulfite fixing of the daguerreotype, at about one part of salt to fifteen of water. Fizeau's condition is freshness rather than duration.
3. Water — "le reste du lavage s'effectue comme d'ordinaire" — the rest of the washing is done as usual, with the wash water as free from dust as possible. Hunt's 1854 plate shows the apparatus a narrow trough of distilled water warmed by a spirit-lamp, the plate drawn out slowly and the water blown off.

L'épreuve étant encore toute iodée, mais exempte de poussière et de corps gras sur les deux surfaces et les épaisseurs, l'on verse quelques gouttes d'alcool sur la surface iodée; quand l'alcool a humecté toute la surface, on plonge la plaque dans la bassine d'eau, puis de là dans la solution d'hyposulfite. Cette solution doit être renouvelée à chaque épreuve, et contenir environ une partie de sel pour quinze d'eau.

Recorded because it is what the source publishes, and not as something to follow. It is set down for two reasons. It is the condition Fizeau puts on his own formula — a plate carrying grease, dust or a tired fixing bath does not gild — and it is the earliest published statement in this formulary that a fresh fixing bath and an exhausted one are different reagents, which is the whole of Hardwich's later argument about sulphiding and the reason every fixer page in this book has a capacity section.

☠Level D: this page explains a bath and gives no procedure

Neither solution above contains a Level D substance. Chloroauric acid is Level B in this course’s chemical encyclopaedia and sodium thiosulfate is Level A. The classification is not about the bath. It is about the object the bath is poured over.

A daguerreotype cannot be made without mercury and iodine, both of which are Level D, and the Daguerre 1839 entry sets out why in full. This page therefore explains what the gilding bath is, what it does and why, and gives no procedure: no apparatus, no sequence to follow, no flame under a plate. The historical operation heats a mercury-bearing plate under a standing film of liquid with a spirit-lamp until, in Hardwich’s version, the liquid boils. There is no domestic control regime that answers that, and none is offered.

What a reader can do instead is one click away. The same chemistry — gold(I) displacing silver, in a thiosulfate ligand — is what the gold thiocyanate toner and the gold-acetate and borax baths do to a salted paper or albumen print, at Level B, with everything published.

The process atlas entry owns the plate and its conservation; the Daguerre entry owns the manual it came out of. This page owns the bath: what Fizeau printed, what the four later printings of it say instead, what is actually in the beaker after the two solutions meet, and why a formula whose reagents fight each other stayed in use for a quarter of a century.

Purpose

To bind the image physically to the plate, and to make the lights brighter while doing it.

Both halves of that are Fizeau’s, in that order, and the first one is the one modern accounts forget. A finished ungilded daguerreotype is a pattern of loose mercury-silver amalgam globules lying on a mirror. The AIC’s conservation page states the consequence in one sentence: ungilded plates are very prone to abrasion and the image can easily be wiped off. Cassell’s 1911 account says the same thing from the other end — the lights are an adhering and very delicate film of mercury which, if fingered in any way, would be wiped off. Hunt’s 1841 report of Fizeau frames the whole operation as producing “a more perfect adhesion of the mercury and silver”, and his test of a successful gilding is that the picture will bear rubbing with the finger.

That is why this entry is filed under fixing in two of its four nineteenth-century English sources and under toning in the other two. It does both, and the reason it does both is a single fact of the chemistry: gold deposits onto the amalgam and onto the bare silver at different rates and to different visible effect.

The second half of the purpose is the one that made it commercial. Eder’s judgement is that the advance was generally adopted and largely increased the public demand for daguerreotypes; the Britannica’s is that nearly all the daguerreotypes extant have been treated in this manner and that their permanence is in great measure due to the operation. A process that had been a curiosity of 1839 became an industry partly because of a litre of liquid containing one gram of gold.

There is a third purpose that is the course’s rather than Fizeau’s. This is where gold toning begins. Every alkaline gold bath in this formulary — the acetate, the borax, the thiocyanate, Kodak’s GP-1 — exists because this one worked and then, on paper, failed. Understanding why it failed is most of what a reader needs in order to understand why the modern ones are built the way they are.

Recommended uses

None. This is a historical-study entry and the honest answer to “what is this for” is reading and understanding.

What the sources recommend it for is a short and consistent list, and it is worth having because it marks the boundary of the formula’s competence.

• Daguerreotype plates, gilded once, after fixing and washing. Fizeau, Hunt in both books, Hardwich, Cassell and the Britannica all describe the same single application.
• Old plates as well as new ones. Fizeau states it — fût-elle fort ancienne, even a very old proof — and the Académie saw it demonstrated in the same volume, when Arago showed a plate of Hubert’s made more than a year earlier and gilded by this method. Very few toners in this book work on an image that has been in a drawer for a year.
• Paper prints, from 1847 to about 1855, and then not. Reilly traces the borrowing precisely: the idea of gold toning paper prints came from the daguerreotypists, P. F. Mathieu applied it in 1847, Le Gray popularised it after 1850, and for a few years sel d’or was the most widely used method for both albumen and salted papers. It was abandoned for reasons this page’s mechanism section explains.

Ware adds one live use: the method is still used to good effect by latterday daguerreotypists. The course records that and does not join them, for the reason in the callout above.

When another formula is preferable

Always, for anything a reader intends to do.

For gold-toning a salted paper or albumen print, the gold-acetate toner is the direct historical replacement and the gold-borax toner the faster one. Both separate the gold from the thiosulfate entirely, which is the single change that fixed everything wrong with this formula.

For gold-toning a silver gelatin print, Kodak’s GP-1 at 1 g of gold chloride in ten millilitres of a one per cent stock, with 15.2 mL of a thiocyanate whose strength Kodak never states, is the modern engineering of the same idea at a tenth of Fizeau’s gold; its own page carries both the comparison and the missing quantity.

For a warm image colour without gold at all, the sulphide toner does deliberately, and under control, the thing this bath does accidentally.

For the fixing half of the job, any bath in the fixer section is better than a roughly six per cent hyposulfite solution used once. The plain hypo fixing bath is the same chemistry at about six times the strength, and the printing-out plain thiosulfate fixer is what a salted-paper worker uses now.

For the history without the plate, the Daguerre 1839 entry and the cyanide fixing and toning entry are the two neighbours of this one, and neither is a procedure either.

Mixing

Fizeau publishes four sentences of mixing instruction and every clause in them is load-bearing.

Two separate solutions, each in pure water, each half a litre. One gram of gold chloride in one; three grams of hyposulfite of soda in the other. He specifies eau pure for both and says nothing about temperature, vessel or filtration.

The gold goes into the hypo, and not the other way round. On verse alors la dissolution d’or dans celle de soude — the gold solution is poured into the soda solution. Hunt’s 1841 account has it, Hunt’s 1854 translation has it, Ware’s reading of Gaudin has it. Four independent printings and not one of them reverses it.

Little by little, with agitation. Peu à peu et en agitant. Hunt makes the pause explicit in 1854: “a little by little, agitating between each addition.”

The colour tells you it worked. La liqueur mixte, d’abord légèrement jaunâtre, ne tarde pas à devenir parfaitement limpide — the mixed liquor, at first slightly yellowish, soon becomes perfectly limpid. Ware’s phrasing from Gaudin is that the slightly yellowish liquid rapidly became quite colourless.

⚗Why the order of addition is the formula

Gold(III) in a thiosulfate solution has two fates and they compete. If thiosulfate is in excess where the two meet, the gold is reduced to gold(I) and immediately captured by two thiosulfate ligands as the soluble complex — the sel d’or Fizeau correctly identified as “un hyposulfite double de soude et d’or”. If gold(III) is in local excess instead, there is not enough thiosulfate to complete either the reduction or the capture, and what comes down is a precipitate rather than a solution.

Pouring the gold into the hypo guarantees the first case at every instant. Every drop of gold arrives in a large volume of thiosulfate and never sees anything else. Pouring the hypo into the gold guarantees the second case for the whole first half of the addition. The direction of addition is not a detail of technique; it decides which compound you make.

The yellow that clears is the reagent reporting on itself. Gold(III) chloride solutions are yellow; the bisthiosulfatoaurate(I) complex is colourless. Fizeau’s limpide is the end of the reduction, and it is the only analytical instrument in the procedure.

What the arithmetic says about how much room there is. Reilly’s description of sel d’or is that it is formed by mixing gold chloride into an excess of sodium thiosulfate, and the interesting question is how much excess Fizeau actually left. Taking the gold as chloroauric acid trihydrate at 393.83 g/mol and the hypo as the pentahydrate at 248.19 g/mol:

1 g ÷ 393.83 g/mol = 2.54 mmol of gold; 3 g ÷ 248.19 g/mol = 12.09 mmol of thiosulfate; ratio = 4.76 to 1

Fizeau's molar ratio

The balanced reaction below needs four thiosulfate per gold — two oxidised to supply the two electrons, two retained as ligands. Fizeau’s bath carries a 19 per cent excess over the stoichiometric minimum, and that is all. The conversions and the ratio are the course’s arithmetic on Fizeau’s published weights, not Fizeau’s; he gives no molar reasoning and the printed extract explicitly withholds the chemical discussion, ending with the words that the memoir closes with considerations on the reactions taking place at the different stages.

That 19 per cent is the number that explains the whole page. It explains why the order of addition is absolute; it explains why the later English printings all moved to a larger excess; and, as the next callout shows, it very nearly settles which gold salt Fizeau was buying.

🧮Which gold chloride? Do the arithmetic and one candidate falls out

“Chloride of gold” named more than one thing in 1840, and Ware’s footnote says the results differ. Fizeau’s 3 g of hypo is 12.09 mmol of thiosulfate. Run 1 g of each candidate against the 4-to-1 requirement. The molar masses are the encyclopaedia’s where it has a page, and the course’s addition of the water of crystallisation where it has not:

Candidate	Molar mass	1 g is	Thiosulfate needed	Fizeau supplies
Chloroauric acid trihydrate, HAuCl₄·3H₂O	393.83	2.54 mmol	10.16 mmol	12.09 mmol — 19 % spare
Sodium tetrachloroaurate dihydrate, NaAuCl₄·2H₂O	397.80	2.51 mmol	10.06 mmol	12.09 mmol — 20 % spare
Chloroauric acid, anhydrous, HAuCl₄	339.80	2.94 mmol	11.77 mmol	12.09 mmol — 3 % spare
Gold(III) chloride, AuCl₃	303.32	3.30 mmol	13.19 mmol	12.09 mmol — 9 % short

The three chloroaurates all work, the two hydrated ones with room to spare and the anhydrous acid by a margin of about three per cent, which is inside a weighing error. True anhydrous gold(III) chloride does not work at all: at one gram to three, there is not enough thiosulfate in the beaker to reduce and complex all of the gold, and the bath would be running in exactly the condition the order of addition exists to avoid.

The arithmetic is the course’s and the conclusion is an inference, not a finding. What it establishes is that Fizeau’s ratio suits a chloroaurate and not the simple trichloride — and the chloroaurates are what was actually sold. It does not choose between the acid and the sodium salt, because those two differ by one per cent. Fizeau’s own remark that the liquor contains sel marin alongside the double hyposulfite does not decide it either: the sodium can come from the excess hyposulfite whichever salt went in. The chloroauric acid page carries the assay problem in its modern form — the two solids are about 50 and 64.93 per cent gold, so a gram of the wrong one is roughly thirty per cent more gold than the formula intended.

The wash is part of the mixing instruction, in the sense that matters. Fizeau spends more words on preparing the plate than on preparing the bath, and the reason is chemical rather than fussy. The surface must be free of dust and above all of grease; the alcohol is there to make the water wet the whole face so that it does not retreat from the edges and leave stains; and the hyposulfite bath must be renewed for each plate. That last condition is the earliest statement in this formulary of something the fixer pages spend chapters on: a fresh thiosulfate bath and a tired one are not the same reagent.

Behaviour

The bath announces its own readiness and then says nothing more. The yellow-to-limpid change on mixing is the only signal the formula gives. After that there is no colour, no smell and no visible end point until the plate is in it.

On the plate it works fast and it is watched rather than timed. Fizeau gives one or two minutes with a strong lamp under the plate, and an end point that is an appearance — the proof is seen to clear and take on great vigour, and when the effect is produced the liquid is poured off. Hunt’s 1854 translation gives the same. Hardwich, using the pure salt at one part in 500, gives the only physical end point in the literature: the flame is applied until the liquid begins to boil.

Heat is not optional and no source treats it as such. Every account — Fizeau, Hunt twice, Hardwich, Cassell, the Britannica, Ware’s reading of Gaudin — has a lamp under a level plate. That is unusual in this formulary. Nearly every other bath here works at or below room temperature, and the reason this one does not is in the mechanism: the displacement of silver by gold(I) is a slow reaction on a metal surface with no development chemistry to accelerate it.

It behaves differently on the two parts of the picture, and that is the whole effect. Hardwich is the clearest:

This Silver, which appears of a dark colour and forms the shadows of the image, is rendered still darker; a very delicate crust of metallic Gold gradually forms upon it, whereas in the case of the image itself the whitening effect is immediate and striking.

He then does something better than describe it: he uses it as evidence. A plain silver surface, such as a collodion image, is darkened by hyposulphite of gold. The daguerreotype’s highlights are whitened. Therefore the highlights are not plain silver, and metallic mercury must enter into their composition. That is a chemical argument for the structure of the daguerreotype image, made from the behaviour of a toning bath, in 1864.

On paper it behaved badly, and the reason was inconstancy rather than weakness. Hardwich records what the trade found: prints coloured as fast in an old bath as in a newly mixed one, but the tints were more fugitive from the old one. The new bath toned mostly by depositing gold; the old bath toned entirely by sulphuration. A bath that gives the same result today and a different one next week, with no way of telling which you have, is a bath the profession abandons, and it did.

Image characteristics

It does not make the plate look gold. This is the most persistent error about the process and it has a named correction: Barger’s remark, quoted by Ware, that the reports that gilded daguerreotypes appear gold coloured are part of the myth of the process. Ware’s own evidence for the correction is the exception that proves it — the two 1845 plates by Philipp Graff found in the exiled Kaiser’s archive in 1996, which really are specularly golden, are described as quite unlike gold-toned daguerreotypes, and are suspected of having been gold-plated by galvanisation rather than toned.

What it does is add force at both ends of the scale. Fizeau’s account is the mechanism and the description at once: the silver of the shadows is browned by the thin layer of gold that covers it, so the blacks are strengthened; the mercury of the lights gains solidity and brilliance from its amalgam with the gold, so the highlights are raised. Cassell’s summary is that it brought out the details with greater force and brilliancy; the Britannica’s is that the picture is better seen in most lights.

Colour, where it changes, is not entirely gold’s doing. Ware records Gaudin noticing that the pure crystalline salt gives “less warm” colours than Fizeau’s mixed bath. That difference is the tetrathionate: the mixed bath carries a sulphiding agent that the pure salt does not, so part of the warmth of a Fizeau-gilded plate is silver sulfide rather than gold. It is the earliest reported observation of the impurity, seventeen years before Hardwich explained it.

Permanence and handling change more than appearance does. Hunt’s 1841 test is that a well-gilded picture will bear rubbing with the finger; the AIC’s modern statement is the same fact from the conservation side, that ungilded plates are very prone to abrasion and the image can easily be wiped off, and that gold toning changes the composition of the image layer. It is not a varnish over the picture. It is an alloy in it.

What the course cannot tell you. No source in this corpus gives a density measurement, a reflectance curve or a quantified change in the tonal scale of a gilded plate against an ungilded one, and the page does not supply one. Every characteristic above is a period observation or a conservation observation, and each is attributed.

The mechanism

Four reactions, in the order they happen, and only the first two were understood at the time.

The bath makes itself

Gold arrives as gold(III) and has to become gold(I) before it can tone anything. Thiosulfate does both jobs — it supplies the electrons and then supplies the ligands.

2 S₂O₃²⁻ → S₄O₆²⁻ + 2 e⁻

Thiosulfate is the reducing agent, and it becomes tetrathionate

Ware’s general statement is that oxidation of thiosulfate by almost any oxidising agent first produces tetrathionate by removing two electrons, and Hardwich’s vocabulary entry names gold chloride as one of the three reagents that do it to hyposulphite of soda. Reducing gold(III) to gold(I) takes two electrons, so one gold consumes two thiosulfate as fuel — and then keeps two more as ligands:

AuCl₄⁻ + 4 S₂O₃²⁻ → [Au(S₂O₃)₂]³⁻ + S₄O₆²⁻ + 4 Cl⁻

Sel d'or, with the by-product nobody wanted

That anion is sel d’or: Fordos and Gélis isolated it as colourless crystals in 1843, called it hyposulphite of gold and sodium, and Ware gives the salt as Na₃[Au(S₂O₃)₂]·2H₂O — trisodium bisthiosulfatoaurate(I) dihydrate in the modern name nobody uses. Fizeau identified it correctly in 1840 without isolating it, from the colour change and from the fact that the liquor also contained sea salt, three years before the chemists confirmed it.

!The coefficients here are the course's, and one printing does not balance

Ware prints this reaction with two gold and six thiosulfate on the left. In the course’s copy of Argyronomicon that version balances neither chlorine nor charge, and it cannot balance the electrons: two gold(III) becoming two gold(I) requires four electrons, and one tetrathionate supplies only two. The equation above is the same chemistry with the coefficients recomputed from the electron count, and the recomputation is the course’s. Whether the printed version is the book or the scan, the course cannot tell from the copy it holds, and says so rather than reproducing an equation it cannot verify.

The bath tones by displacement

Gold(I) sitting in a thiosulfate cage is a mild oxidant looking for an electron, and metallic silver has one to give. The complex changes its metal:

[Au(S₂O₃)₂]³⁻ + Ag → [Ag(S₂O₃)₂]³⁻ + Au

Gold out, silver in: the toning reaction

Read that against Fizeau’s own sentence and they are the same statement in two languages, 179 years apart: dans cette opération, de l’argent s’est dissous, et de l’or s’est précipité sur l’argent et sur le mercure — in this operation silver has dissolved and gold has precipitated onto the silver and onto the mercury. The silver leaves as the same bis(thiosulfato)argentate complex that every fixer in this book makes, which is why a gold bath built on thiosulfate is also, unavoidably, a fixing bath. Reilly makes that point exactly: sel d’or is what would now be called a toning-fixing bath, because it is formed by mixing gold chloride into an excess of thiosulfate and so retains its power to dissolve silver chloride.

Where the reaction happens decides what it looks like. On the bare polished silver of the shadows the deposit builds slowly into what Hardwich calls a very delicate crust of metallic gold, and the surface darkens. On the amalgam globules of the highlights the gold alloys into the mercury-silver particle rather than plating over it, and the particle becomes harder, brighter and mechanically anchored. The AIC’s phrase for the whole is that gold toning changes the composition of the image layer, which is the right level of precision: the image is not coated, it is alloyed.

The by-product tones as well, and nobody asked it to

The tetrathionate from the first reaction does not sit there. Hardwich’s own words are the clearest statement of the problem in the nineteenth-century literature:

When a solution of Chloride of Gold is added to Hyposulphite of Soda, we have in the liquid not only the double Hyposulphite of Gold and Soda known by the name of Sel d’Or, but also a portion of the unstable Tetrathionate of Soda, prone to liberate Sulphur. The action of the Bath is therefore complex from the very first, but becomes more so on keeping the solution for a time, since spontaneous decomposition ensues.

And tetrathionate is a sulphide toner of metallic silver:

4 Ag + S₄O₆²⁻ + H₂O → 2 Ag₂S + HSO₃⁻ + HSO₄⁻

The unwanted half: tetrathionate sulphides the silver

So a plate gilded in Fizeau’s bath is part gold-toned and part sulphide-toned, in a proportion nobody controls. Ware states it as the reason the pure crystalline salt was worth its considerable price, and Gaudin’s observation that the pure salt runs cooler is the visible evidence. The course has no encyclopaedia page for sodium tetrathionate, and no page can be linked for it here; what the reader needs to know about it is on this page and on the sulphide toner.

Hardwich adds the kinetics, and they are worse than the equation suggests. A tetrathionate solution goes milky within days from deposited sulphur on its own; add hyposulphite crystals to it and it begins to deposit sulphur shortly, continues for days, and turns acid to test-paper. The two things in this bath that a formulator would most want to keep apart are the two things the formula is made of.

Why heat

The displacement reaction runs at a metal surface, and its rate is set by how fast the complex reaches the silver and how fast the electron transfer goes when it gets there. Both improve with temperature, and neither has any other lever in this formula: there is no accelerator, no alkali and no catalyst. Fizeau’s forte lampe and Hardwich’s until the liquid begins to boil are the whole of the rate control, and one to two minutes at that temperature is what they bought. The alkaline gold baths that replaced this one solve the same problem a different way, by raising the pH so that gold(III) is reduced to gold(I) before the print goes in — which is why the gold-acetate toner needs twenty-four hours of ripening and then works cold.

Function of every ingredient

Chloroauric acid, 1 g in 500 mL of pure water, so 0.2 per cent w/v in the stock and 0.1 per cent w/v in the mixed litre. The gold, and the only substance on this page that ends up in the finished picture. Fizeau prints it as chlorure d’or and the course reads that name in the photographic literature as this compound rather than gold(III) chloride proper; the maths callout above shows that the anhydrous trichloride would leave the bath short of thiosulfate, which is corroborating evidence for the reading rather than proof of it. Its job is to arrive as gold(III), be reduced in the beaker to gold(I), and be carried to the plate as an anion that will trade its metal for silver. More gold makes a faster bath and, on paper at least, one Reilly warns deposits gold superficially so that the layer is largely removed in the fixer; on a metal plate the more immediate consequence is arithmetic — past about 1.2 g in this litre there is no longer enough thiosulfate to reduce and complex it, and the formula stops being sel d’or. Less gold simply tones more slowly, and Hardwich’s pure-salt variant at an equivalent 0.2 per cent shows the working range is not narrow. At 0.1 per cent w/v of gold chloride the mixed bath is two to ten times the gold of Reilly’s albumen and salted-paper toners, which run at 0.1 to 0.5 g of gold chloride per litre. That is what a single one-to-two-minute application at boiling point costs. Its own page carries the corrosive classification, the storage rule that gold stocks live in the dark because light reduces them, and the assay warning.

Sodium thiosulfate pentahydrate, 3 g in 500 mL of pure water, so 0.6 per cent w/v in the stock and 0.3 per cent w/v in the mixed litre. The hardest-working ingredient in the formulary, because it does three different jobs in the same beaker and the third one is a fault.

• Reducing agent. Two thiosulfate ions give up two electrons to become one tetrathionate, and those two electrons are what turn gold(III) into gold(I). Nothing else in the bath can do it.
• Ligand. Two more thiosulfate ions wrap the gold(I) into the [Au(S₂O₃)₂]³⁻ anion, which is soluble, colourless and stable enough to keep. Without them the gold(I) would not stay in solution at all.
• Silver solvent. Whatever is left over after those two duties is ordinary hypo, and it dissolves silver. On a daguerreotype that costs the picture nothing and may help it, because the plate has already been fixed and the loose halide has already gone; on a paper print it is the reason sel d’or is a toning-fixing bath rather than a toner.

More thiosulfate is the direction every later printing moved: Hunt in 1854 and Cassell in 1911 both give roughly a 1-to-4 weight ratio against Fizeau’s 1-to-3, which raises the molar excess from 19 per cent to 59 and 70 per cent respectively. That buys safety margin on the order of addition and more silver solvency, at the cost of more tetrathionate and so more sulphiding. Less thiosulfate runs the formula below its stoichiometric floor, and below about 2.5 g in this litre the bath cannot reduce and complex a gram of gold at all.

The pentahydrate is what “hyposulphite of soda” meant in the trade and what every molar figure on this page assumes, at 248.19 g/mol. The anhydrous salt has a separate encyclopaedia page and a molar mass of 158.11, so a gram of it carries over half as much again the thiosulfate of a gram of the crystals; nothing in the sources suggests Fizeau used it, and the course flags the distinction because the formulary keeps hydrates apart everywhere else.

The same substance again, in the wash, at about one part to fifteen of water. This is a different ingredient in every sense that matters even though it is the same chemical: it is a fixing bath, it is used before the gilding rather than in it, and Fizeau’s condition on it is not a strength but a renewal — it must be changed for every plate. Its function is to take the unexposed silver iodide off the plate and, by being fresh, to take it off without leaving anything behind that will sulphide the image when it is heated in the gold bath a few minutes later. Hardwich’s later finding that an old hyposulphite fixing bath becomes a toning bath without any addition of gold is the retrospective justification for a rule Fizeau stated as a working observation. A stronger wash bath fixes faster and carries more risk of leaving residual thiosulfate in the plate’s edges and mounting; a weaker or a tired one leaves silver iodide behind, which the gold bath will then attack, and leaves polythionates behind, which will tone the picture a colour nobody chose.

Water, pure, one litre of it in the finished bath. Fizeau says eau pure three times in four sentences and returns to it for the washing, where he asks for water as free from dust as possible. Neither he nor any later source gives a temperature for making up either stock. The purity requirement is not fastidiousness: a bath running on a 19 per cent stoichiometric margin, poured over a mirror, has no tolerance for a chloride or a sulfide arriving from the tap.

No substance in this formula is unexplained by its source, but one is unnamed by it. The sodium chloride Fizeau reports in the mixed liquor — plus du sel marin, qui ne paraît jouer aucun rôle dans l’opération — is a product rather than an ingredient, and his judgement that it plays no part is correct: four chloride ions per gold are released when the gold is complexed, and they do nothing thereafter.

Interactions

Gold(III) and thiosulfate, in the wrong order. Covered under Mixing and worth repeating here as an interaction rather than a procedure: these two reagents give one product when thiosulfate is in excess and a different one when gold is, and the only thing standing between them is the direction of pouring. This is the clearest case in the formulary of an interaction that a quantity table cannot express, which is why the schema carries the order of addition as part of the formula.

Thiosulfate and its own oxidation product. Hardwich’s vocabulary entry is the important one: a tetrathionate solution is unstable on its own, and adding hyposulphite makes it decompose faster, not slower, depositing sulphur for days and turning the liquid acid. Every batch of this bath contains both, in a fixed ratio set by the stoichiometry, from the moment it clears.

Tetrathionate and metallic silver. The sulphiding reaction above, running on the shadows of the picture at the same time as the gold is depositing on them. Nothing in the formula separates the two, and Ware’s summary is that the result is a mixed one, variable in its composition of silver, gold and sulphide.

Gold(I) and the amalgam against gold(I) and bare silver. The interaction that produces the entire visible effect. The same reagent whitens one part of the image and darkens the other, and Hardwich’s control experiment — a plain silver collodion image, which only darkens — is what turns that from an observation into an argument.

Thiosulfate and acid, which is the incompatibility that governs the whole fixer section of this formulary and is doubly relevant here because the bath acidifies itself as the tetrathionate breaks down. Thiosulfate in acid decomposes to sulfur and sulfur dioxide; the reaction is written out on the 1928 bisulphite fixing bath page.

Gold baths and fixer, generally. Reilly’s rule for every gold toner is that the solution is ruined by even a trace of fixer. This formula is the exception that defines the rule, because its fixer is built in — and the price of that convenience is exactly the inconstancy that killed it on paper.

Variants

Hardwich’s pure salt, 1864

The important variant, and the one that shows what the impurity was costing. Instead of making the bath in situ, the isolated crystalline sel d’or — precipitated from a concentrated solution with ethyl alcohol and sold by apothecaries at considerable cost — is dissolved on its own at one part in 500 parts of water, which is 0.2 per cent w/v of the salt. The plate goes on a levelling-stand, the solution is poured on, and the spirit-lamp is applied until the liquid begins to boil.

No gold chloride is added to any thiosulfate, so no tetrathionate is formed, so there is no sulphiding. Ware records the consequence from Gaudin’s side: the pure salt gives “less warm” colours than Fizeau’s mixed bath. Sutton’s 1855 handbook, in Ware’s account, describes the pure-salt route on paper as giving “the French violet tints”. The course lists this as a variant of Fizeau’s bath and not as a separate entry because it is the same anion, arriving by a different road.

The English printings, which are not the same formula

Four sources print quantities and only two of them agree. The conversions below are the course’s arithmetic at 1 grain = 64.79891 mg and 1 imperial fluid ounce = 28.4131 mL, on the same basis the albumen solution entry uses for its period measures.

Printing	Gold stock	Hypo stock	Gold-to-hypo by weight	Working bath, gold chloride	Thiosulfate to gold, molar
Fizeau, 1840	1 g in 500 mL	3 g in 500 mL	1 to 3	0.10 % w/v	4.76 to 1
Hunt, 1841	15 gr in 1 pint	45 gr in 1 pint	1 to 3	0.086 % w/v	4.76 to 1
Hunt, 1854	8 gr in 16 fl oz	32 gr in 4 fl oz	1 to 4	0.091 % w/v	6.35 to 1
Cassell, 1911	7 gr in 10 fl oz	30 gr in 4 fl oz	1 to 4.3	0.114 % w/v	6.80 to 1
Eder, 1945	1 part	300 parts in 1,000 of water	1 to 300	0.10 % w/v	476 to 1

Three things fall out of that table and each is worth a sentence.

Hunt’s 1841 account is Fizeau’s formula in English clothes. Fifteen grains is 0.972 g, forty-five is 2.916 g, and a pint is standing in for half a litre. The ratio is identical and the working bath is 14 per cent weaker only because an imperial pint is 568 mL rather than 500.

The two later English printings moved to a bigger thiosulfate excess and kept the gold. Hunt in 1854 and Cassell in 1911 both land the working bath within a fifth of Fizeau’s gold concentration while raising the molar excess of thiosulfate from 19 per cent over the minimum to 59 and 70 per cent. The course reads that as practitioners discovering that Fizeau’s margin was uncomfortably thin — an inference, not a finding, since neither source says why it prints what it prints. Note also that Hunt in 1854 translates the rest of Fizeau’s note almost word for word, including Fizeau’s own “one part of salt to fifteen of water” for the wash, and still prints different stock weights.

Eder’s figures are not a variant of this formula. Three hundred parts of hyposulphite of soda in a thousand of water is a 30 per cent w/v solution — a fixing bath, at ordinary fixing strength — with one part of gold chloride in it, and it carries nearly five hundred thiosulfate for every gold. That is a combined fixing-and-toning bath rather than the gilding bath Fizeau published, and the course does not follow it. Eder’s gold figure of 0.1 per cent w/v is nonetheless exactly Fizeau’s working strength, which suggests the discrepancy lives in one number rather than in the whole recipe. The page records that and does not resolve it: Eder is the historian of record for the period and his statement is reported as his.

What replaced it

Hardwich states the outcome plainly: In the Gold toning process, as now employed, all danger of Sulphuretting the prints is avoided … we no longer mix the two solutions. His own papers on the tetrathionates and their reaction with hyposulphite are dated September and October 1854, and Ware dates his published warning to the first edition of 1855. Everything in gold toning after the mid-1850s is built on that decision.

• Gold-acetate toner and gold-borax toner replace the thiosulfate with an alkaline buffer, and let gold(III) reduce itself by oxidising water. Slower to make, cold to use, no sulphide.
• Gold thiocyanate toner and Kodak’s GP-1 keep the sulfur ligand but swap thiosulfate for thiocyanate, which has no tetrathionate to give.
• Sulphide toning does deliberately and controllably what this bath did by accident.

None of those is a variant of this formula in the schema’s sense — each is its own entry with its own provenance — but all four descend from it, and the line of descent is the argument this page exists to make.

Contested history

⁇Who first gilded a daguerreotype

The course’s statement is that Fizeau published the method and that the priority is not settled.

Fizeau, 1840 is the usual attribution and is the position of Eder, Reilly, Vogel, Cassell, the Britannica and Hardwich, all of whom name him and no one else. Ware also gives Fizeau as “usually credited”, records Arago’s announcement to the Académie on 23 March 1840, and cites Fizeau’s own publication later that year — which is the document this page is written from and which is uncontested as a publication of the method.

Himly, 1839 is Ware’s own qualification. He records evidence, from Baier’s 1964 documentary history and Hoerner’s 1982 article in History of Photography, that August Friedrich Karl Himly, professor of chemistry at Göttingen, had already put the idea into use by 19 October 1839 — five months before Arago’s announcement. The course has not read Baier or Hoerner and reports Ware’s summary as Ware’s.

The date of general adoption is separately muddled. Eder gives the invention as 1840 in one chapter and says the use of sel d’or for daguerreotypes was “suggested by Fizeau (1841)” in another; the AIC’s conservation page dates the introduction of gold toning to 1841; Ware’s Argyronomicon says gold toning was first recommended by Fizeau “as early as 1841”. Fizeau’s note is printed in the volume of Comptes rendus covering July to December 1840, so 1840 is the publication date and 1841 is most likely the date of general adoption, which is a distinction several of these sources do not draw. The course uses 1840 for the publication and says “by 1841” for the practice.

Safety

Level D, and the classification does not come from the bath. This is worth saying plainly because the page would otherwise look mis-classified. Take the two solutions on their own and they are unremarkable: chloroauric acid is Level B, corrosive, with H314 in every one of the 200 aggregated ECHA reports behind its entry, and sodium thiosulfate is Level A. Nothing about a litre of 0.1 per cent gold chloride and 0.3 per cent hypo demands a laboratory.

What demands one is the plate. This formula has exactly one substrate, and it cannot exist without mercury and iodine, both Level D in this course’s encyclopaedia. Mercury carries the harmonised classification under Regulation (EC) No 1272/2008 with H330, fatal if inhaled, together with H360D, H372, H400 and H410, and HSE’s EH40 gives mercury and its divalent inorganic compounds a long-term workplace exposure limit of 0.02 mg/m³ with no short-term figure and a biological monitoring value on top. The Daguerre entry and the Level D policy make the argument in full. A page cannot classify a procedure below the classification of the object it is performed on, and the rubric is explicit that Level D is historical study with no actionable home procedure.

The heat is the specific aggravation, and the course cannot quantify it. The historical operation applies a spirit-lamp flame beneath a plate whose highlights are a silver-mercury amalgam, under a standing film of aqueous liquid, until — in Hardwich’s version — it boils. Mercury has a measurable vapour pressure at ordinary room temperature; NIOSH gives 0.0012 mmHg, and the mercury page calls that the most important number on it precisely because it is not zero. No source in this course’s corpus measures what a heated gilded plate releases, and the page does not estimate it. The classification does not depend on the estimate: it follows from the substrate, and the absence of a measurement is itself a reason not to write a procedure.

A second hazard belongs to the bath and is real. Hardwich records that the tetrathionate in this mixture decomposes in the presence of hyposulphite, deposits sulphur, and turns the liquid acid to test-paper. Thiosulfate in acid gives sulfur and sulfur dioxide, as the 1928 bisulphite fixing bath page sets out. A thiosulfate bath that is acidifying itself and being boiled in an open dish is a foreseeable source of sulfur dioxide, and no nineteenth-century account mentions ventilation of any kind.

What is not a hazard here, and why saying so matters. Making up the two stocks — weighing a gram of a corrosive gold salt and three grams of hypo, and dissolving each in half a litre of water — is Level B work of a kind this course teaches on four other pages. There is no cyanide anywhere in this formula, which distinguishes it sharply from the cyanide fixing and toning entry it sits beside in the historical section; there is no chromium(VI), no lead and no uranium. The Level D line runs between the beaker and the plate, and that is an unusually clean example of how a hazard assessment actually works: the substances, the quantities and the operation are assessed separately, and the highest of them governs.

One caution that applies to reading rather than doing: EH40 states that the absence of a substance from its list does not indicate that it is safe.

Storage

Fizeau publishes no keeping time for anything on this page, and the chemistry says the mixed bath should not have one. Hardwich, describing the same mixture in use on paper, states that spontaneous decomposition follows on keeping and that an old bath tones entirely by sulphuration, and his vocabulary entry gives the reason: a tetrathionate solution deposits sulphur within days on its own, and hyposulphite accelerates it. A bath whose active salt and whose principal impurity destroy one another is mixed and used, and every period account describes it being made for the work in hand.

The two stocks are a different question and the sources are silent on it. By the general rules the encyclopaedia gives, a 0.2 per cent gold chloride stock keeps in the dark, because chloroauric acid is reduced by light and by organic matter — Reilly’s storage instruction for every gold stock is a dark bottle — and a 0.6 per cent hypo solution keeps reasonably in a stoppered bottle away from acid. Neither statement comes from Fizeau, and both are marked as the encyclopaedia’s rather than the formula’s.

The pure crystalline salt is the storable form of this formula, and that is why it existed. Ware records it precipitated with ethyl alcohol, sold by apothecaries as a colourless crystalline solid, and priced high even for a gold compound because of the manipulations needed to make it. A solid can be bottled, weighed and kept; the solution it comes from cannot.

The plate afterwards. Gilding is itself a storage measure, which is the point of the whole entry: it converts an image that can be wiped off with a finger into one that Hunt says will bear rubbing with it. It does not remove the need for the cover glass and the sealed case, and the conservation of surviving plates is a specialist field with its own literature.

Incompatibilities

Acid, with the thiosulfate half of this bath. Universal in the fixer section and sharper here, because the bath acidifies itself as its tetrathionate breaks down. Sulfur and sulfur dioxide are the products, and a boiling open dish is the worst arrangement for both.

Any oxidant, with the thiosulfate. Ware’s general statement is that almost any oxidising agent takes two electrons off thiosulfate and makes tetrathionate. Hardwich names iron(III) chloride, copper(II) chloride and gold chloride specifically, and iodine as well. In this formula that reaction is the intended one; anywhere near a gold stock bottle it is contamination.

Light and organic matter, with the gold(III) stock. The chloroauric acid page carries both: organic matter reduces gold(III) to metallic gold, which is why the salts are sometimes supplied in sealed ampoules, and light reduces dissolved gold, which is why the stock lives in the dark.

Fixer, with every other gold toner in this book — and not with this one. Reilly’s rule is that a gold toning solution is ruined by even a trace of fixer. Fizeau’s bath is immune to that particular accident because its fixer is a constituent, and the immunity is bought at the price of everything else on this page. It is a good example of a specification and a defect being the same fact seen twice.

Grease and dust, with the plate. Fizeau’s own incompatibility, stated before any of the chemical ones: the surface must be perfectly free of foreign bodies and above all of fatty matter, and the alcohol step exists solely so that water will not retreat from the edges and leave stains. On a mirror that is developed by reflection, a stain is not cosmetic.

Waste

Nothing on this page is a waste stream a reader will generate, because nothing on this page is a procedure. What follows is what the substances are, for anyone who meets them in another context.

A spent gilding bath is both gold-bearing and silver-bearing. That follows directly from the toning reaction: every gold atom that lands on the plate puts a silver atom into solution as the thiosulfate complex, and any gold that did not land is still in the liquid as sel d’or. Gold is worth recovering on its own account and silver-bearing liquids belong to a silver-recovery route, exactly as the modern fixers in this formulary do.

The pre-gilding wash bath is an ordinary spent fixer carrying dissolved silver iodide as its thiosulfate complex, and it goes the same way.

Thiosulfate itself is an oxygen demand. A thiosulfate solution discharged to water consumes dissolved oxygen as it oxidises, which is the argument the fixer pages make and which does not change because the bath also contains gold.

Mercury waste is not a domestic waste stream in any jurisdiction the course has looked at. Elemental mercury and mercury-contaminated materials are controlled wastes handled by licensed contractors, and mercury’s classification carries H400 and H410. A damaged or dismantled daguerreotype plate is mercury-contaminated material.

Local regulation governs all of it, everywhere. This course states chemistry and general practice and never a jurisdiction-specific instruction; check your local regulations.

Troubleshooting

Every row below is a diagnosis rather than a remedy, which is what a Level D page can carry. The attribution column is not decoration: three of these come from the man who invented the process, three from the chemist who explained it, and the rest from conservation.

What is seen	What causes it	Whose observation
The mixed liquor stays yellow instead of clearing	The reduction of gold(III) has not completed; on Fizeau’s ratio there is only a 19 per cent excess of thiosulfate, so an under-weighed hypo or an over-weighed or anhydrous gold salt leaves gold(III) unreduced	The colour change is Fizeau’s own end point; the arithmetic behind this reading is the course’s
A precipitate on mixing instead of a clear liquid	The hypo was poured into the gold, so gold(III) was in local excess for the first half of the addition	The direction of addition is stated by Fizeau, Hunt twice and Ware; the failure mode is the course’s reading of it
Stains on the plate after washing	Water withdrew from the edges during immersion; the alcohol step exists to stop it, and Fizeau says such stains are produced “infallibly” without it	Fizeau, 1840
The gilding does not take	Grease or dust on the surface. Fizeau requires the plate to be perfectly free of foreign bodies and above all of fatty matter, on both faces and the edges	Fizeau, 1840
The picture warmer than expected, or browner	Sulphiding by the tetrathionate alongside the gold deposition, which is a constituent behaviour of the mixed bath and not a fault of the operator	Hardwich, 1864, with Gaudin’s comparison against the pure salt reported by Ware
Results that change from week to week with no visible cause	The bath was kept. On keeping, the mixture tones increasingly by sulphuration and decreasingly by gold, and the print colours in both cases look similar	Hardwich, 1864
The bath turns milky, deposits sulphur, or turns acid to test-paper	Tetrathionate decomposing, accelerated by the hyposulphite that is in the bath by design	Hardwich, 1864
A picture that tones brown in a fixing bath containing no gold at all	An exhausted hyposulphite bath. Hardwich’s finding is that an old fixing bath becomes a toning bath by itself; it is the reason Fizeau’s wash must be renewed for every plate	Hardwich, 1864
The image wiped off by a touch	The plate was never gilded. This is the failure the whole formula exists to prevent	AIC Photographic Materials Group
A plate that really does look gold	Almost certainly not gold-toned. Barger’s statement is that gilded plates appearing gold coloured is part of the myth of the process; the known specularly golden plates are suspected of being galvanised	Barger and White, quoted by Ware

Experiments

None of these needs gold, mercury, iodine or a daguerreotype. Each takes one claim off this page and puts it where a reader can see it.

Make an old hypo bath tone a print, and watch a fixer turn into a toner. This is Hardwich’s central observation and the cheapest experiment in the section. Fix one print in a fresh plain hypo bath and one in a bath that has been worked hard and stood, wash both identically, and compare the image colour by transmitted and reflected light. The prediction from this page is that the tired bath gives the warmer, browner image and that the difference is silver sulfide rather than gold. Keep the two prints and look again in a year: Hardwich’s second claim is that the sulphided tints are the more fugitive, and that one takes time rather than equipment.

Do the stoichiometry for all five printings and predict which ones fail. Take the table under Variants, compute the thiosulfate-to-gold molar ratio for each, and mark the ones that fall below 4 to 1. Then repeat the calculation substituting anhydrous gold(III) chloride for the trihydrate throughout and see how many printings survive. The exercise is the point: a formula written as weights hides a stoichiometry, and the stoichiometry is what decides whether it works. The arithmetic needs the molar masses on the two chemical pages and nothing else.

Gold-tone a salt print the modern way, and taste the difference the tetrathionate makes. Print and tone a salted paper or albumen sheet in the gold-acetate toner, which is Fizeau’s chemistry with the thiosulfate taken out. Record the image colour before and after. Then read Gaudin’s remark, reported by Ware, that the pure salt gives “less warm” colours than the mixed bath, and decide whether your alkaline-toned print is on the cool side of what the sel d’or literature describes. That is a one-sided comparison and the page says so — you cannot make the other half without gilding a plate — but the direction of the difference is testable.

Test Barger’s correction in a museum. Find gilded daguerreotypes in a public collection, in person or in the Library of Congress and Met online collections, and decide before you look at the catalogue whether the plate reads as gold-coloured. Then write down what you think the highlights and the shadows are made of and check it against Image characteristics above. The claim under test is a specific and falsifiable one: gilding changes the composition of the image layer and does not make the picture look gold.

Watch an order of addition change a product. The claim that pouring A into B and B into A give different results is easy to assert and easy to see. Using the silver and chloride chemistry of the salted paper sensitiser at Level B, add a dilute silver nitrate solution slowly to a dilute chloride solution, and in a second tube do the reverse, matching final concentrations exactly. Compare the two precipitates for how fine they are and how readily they settle. The mechanism is not the same as this page’s — one is peptisation, the other a redox stoichiometry — but the lesson is: what is transiently in excess decides what you get, and only the order of addition controls it.