Davey and Knott's double-jet iodobromide emulsion

Solution No. 1 — the salted gelatin, and the vessel the emulsion is made in

Ingredient	Quantity	Form the source specifies
Gelatin	102.6 g	inert gelatin, which is the whole point of the choice; see Function of every ingredient
Potassium chloride	102.6 g	the figure the course's copy prints, and the same figure it prints for the gelatin on the line above; a transcription duplication cannot be excluded and the arithmetic both ways is under Function of every ingredient
Water	2873 mL, added	at 40 °C; Kodak states an amount of water and a temperature for each solution, and states a make-up only for the finished emulsion, which is 18 litres. Solution No. 1 is the one solution of the four that is not run in: it is what the others are run into.
Added to the working bath, not to the stock
Gelatin	1206 g	after Solution No. 4 and its twenty minutes of ripening, then dissolved by a further fifteen minutes at 45 degrees C — added dry
Potassium chloride	770 mL of a 10% solution	after the wash has taken every soluble bromide out, and before the make-up to 18 litres — printed as a 10 per cent solution by weight, and as "about" 770 c.c.
The patent gives no order of dissolution. The order recorded here is the one the chemistry requires and the one every other emulsion in this formulary uses: the salt goes into the water first and the gelatin is swollen in the salt solution, because gelatin dropped into water swells into lumps that then have to be melted through, and because the chloride has to be evenly distributed before any silver arrives. The two working-bath lines are attached to this solution because Solution No. 1 is the vessel that becomes the emulsion.

Solution No. 2 — the chloride jet

Ingredient	Quantity	Form the source specifies
Potassium chloride	513 g
Water	2668 mL, added	at 45 °C

Solution No. 3 — the silver jet

Ingredient	Quantity	Form the source specifies
Silver nitrate	1000 g
Water	2668 mL, added	at 45 °C

Solution No. 4 — the converting halide

Ingredient	Quantity	Form the source specifies
Potassium bromide	820.8 g
Potassium iodide	205.2 g
Water	2565 mL, added	at 45 °C

Mixed in this order — the whole make, in the order the patent gives it

1. Start with the whole of Solution No. 1 — the vessel, at 40 degrees C, holding the inert gelatin and its own chloride
2. Then add the whole of Solution No. 2 — one of the two jets, at 45 degrees C
3. Then add the whole of Solution No. 3 — the other jet, at 45 degrees C, run at the same time as Solution No. 2 and not after it
4. Then add the whole of Solution No. 4 — added after the ninety seconds and one minute of ripening

Run Solutions Nos. 2 and 3 simultaneously into Solution No. 1 in a vessel, taking 90 seconds to do this. Then ripen for 1 minute at 45 C. Next add Solution No. 4 then ripen for 20 minutes at 45 C. Next add 1206 gms. of inert gelatin (dry). Then ripen at 45 C. for 15 minutes during which time the gelatin dissolves. Set and shred the emulsion and then wash until free from all soluble bromide and then add about 770 ccs. of 10% solution of KCl (by weight), and then add water to make 18 litres.

The schema's three ways of combining solutions are a ratio, an order of addition and a sequence of baths, and none of them can say “simultaneously”, which is the one word this formula turns on. The order of the parts above is the order in which the solutions are introduced; Solutions No. 2 and No. 3 share a position rather than occupying two, and the instruction is what says so. That is a gap in the schema rather than a hedge about the source, and it is recorded in the notes on this page under Variants.

!Read this before the table above

This entry is filed under a planned name that promises more than any source delivers: double-jet iodobromide emulsion with sulfur-gold sensitisation. The formula printed above is a genuine double jet and a genuine iodobromide, published by Kodak with every weight stated. It is not sulfur-and-gold sensitised, and it is not an oversight that it is not: Davey and Knott built it so that chemical sensitisation would have nothing to work on. No published formula at any tier combines a double-jet iodobromide precipitation with a sulfur-and-gold digestion and prints the quantities for both. Sulfur-and-gold sensitisation is taught in full on this page, from the patents that disclose it, and is not smuggled into the table.

Purpose

This emulsion exists to put its image inside the grain instead of on the surface, and every strange thing about the formula follows from that one requirement.

Kodak Limited’s laboratory at Wealdstone wanted a negative emulsion for a reversal process — the one described by Knott and Stevens, which the patent identifies as its intended use — in which the useful latent image is buried where a surface developer cannot reach it. A normal negative emulsion is built to do the opposite. It is precipitated in an active gelatin that carries traces of sulphur compounds, ripened long and hot, washed, and then digested: reheated for an hour or more so that those sulphur compounds can build sensitivity specks on the crystal faces. Davey and Knott took that list of operations and removed the ones that make a surface sensitive. They used an inert gelatin from which the sulphur has been oxidised out. They used no ammonia. They held every solution and every stage below about 50 °C, against the 70 to 90 °C the patent itself gives as normal for a non-ammoniacal make. They ripened for twenty-one minutes in total rather than one or two hours. And after the wash they gave the emulsion no digestion at all, stating in the description that treatments after washing which increase sensitivity, such as heat treatment, are to be avoided.

What is left is a formula that teaches by subtraction. It is the only published double-jet iodobromide make in this corpus whose weights can be transcribed, so it is where the course shows what a double jet is with real numbers attached. It is the corpus’s one worked example of halide conversion, in which a silver chloride skeleton is turned into a silver iodobromide by nothing more than a change of the halide in the liquor around it. And because it is built to refuse chemical sensitisation, it is the clearest way in the whole formulary to say what sulfur and gold actually do — you can see their absence in the ingredient list, in the choice of gelatin, in the temperature ceiling and in the missing digestion.

Recommended uses

As the worked example of a double jet. Two jets, ninety seconds, a stated vessel, stated weights and a stated make-up. Every other statement of a double jet available to this course is a set of conditions without quantities. Read it beside Kodak’s own US 3,655,394 and Berry and Marino’s US 3,772,031, which give the conditions this one does not.

As the reference for halide conversion. The step in which Solution No. 4 turns 5.887 moles of silver chloride into silver bromide and silver iodide is one of the most instructive single operations in emulsion chemistry, and this is the corpus’s only quantified instance of it.

As the control in any discussion of chemical sensitisation. An emulsion that has been built to be insensitive at the surface is the natural comparison for one that has been built to be as sensitive as possible there. Read it against the sulfur-and-gold doses under The mechanism.

As a record of what an industrial iodide content looks like. At 21.0 mole per cent of the silver, this emulsion carries between three and twenty times the iodide of an ordinary negative emulsion. The reason is under Interactions.

It is not a make to perform. Nothing in the formula is exotic, and its safety classification is the same Level B that every silver nitrate procedure in this course carries, but the scale is eighteen litres and a kilogram of silver nitrate, the product has no use outside the reversal process it was designed for, and it would need an internal developer this course does not publish to show an image at all. The course prints it as a document, not as an instruction.

Part V’s own emulsion sequence is a ladder of makes a student can actually run, and this is not one of its rungs. It belongs to the reading beside them: the industrial answer to the question the paper emulsions raise, which is what a manufacturer does that a person at a bench cannot.

When another formula is preferable

To make a negative emulsion at a domestic scale, nothing on this page helps. The make Part V performs is a plain-silver iodobromide dry-plate emulsion two orders of magnitude smaller than this one, made single-jet, and it is discussed in the lessons rather than here.

To make a printing material, the three paper emulsions in this formulary are the ones that have been made: the silver chloride contact emulsion for a first make, Wall’s soft-working gaslight paper for a chlorobromide, and Trumm’s bromide paper emulsion for a washed bromide enlarging paper.

To sensitise an emulsion with gold, the preparation is the Steigmann gold sensitiser stock, which carries the one gold dose in this corpus that somebody has tested. To sensitise one to green, it is the erythrosin stock.

To harden a coated emulsion, see the chrome alum hardening bath.

For a normal, fast, surface-sensitive iodobromide negative emulsion, this course has no formula at any scale, and says so rather than assembling one. The conditions are published — Illingsworth’s pH of 4.0 or less and pAg of 8.6 to 9.2, Berry and Marino’s pH 6.0 and pAg 9.0 at 70 °C, Fuji’s pH 6.5 and pAg 9.3 at 60 °C — and the sensitisation is published, and no document read for this course prints them together with weights.

Mixing

The order below is Kodak’s, quoted in the combination beneath the table. What follows is what each step is doing and why it cannot be reordered.

Four solutions, all below 50 °C. Solution No. 1 is made up at 40 °C and the other three at 45 °C. The ceiling matters: the description states that during all of the emulsion-making procedure the temperature of the solutions and emulsions should not be above about 50 °C, and in the same document Kodak gives 70 to 90 °C as the normal precipitation temperature for a neutral emulsion. Twenty-five degrees of restraint is a deliberate brake on ripening, which is a solution-and- redeposition process and slows sharply as the temperature falls.

The double jet: ninety seconds. Solutions No. 2 and No. 3 are run into Solution No. 1 at the same time. A single jet — the way every other emulsion in this formulary is made — pours silver into a standing bath of halide, so the halide starts in enormous excess and falls steadily as the silver goes in, and the conditions the crystals grow under change continuously from the first drop to the last. A double jet holds them still. Silver and halide arrive together, react on contact, and the concentration of free halide in the vessel stays where the operator set it. Ninety seconds is fast for the volumes involved: the two jets carry 5,336 c.c. of water between them into a vessel made up with 2,873 c.c., so something like three and a half litres a minute goes in through the two jets together.

⚗What the vessel is doing during those ninety seconds

Silver nitrate and potassium chloride meet at the jets and precipitate silver chloride immediately.

Ag⁺ + Cl⁻ → AgCl

The first precipitation

Because the two streams arrive together, the ratio of silver to chloride in the vessel is fixed by the ratio of the two flow rates rather than by how far through the addition you are. The chloride already in Solution No. 1 sets where that ratio sits: it is a reservoir of halide that the silver never exhausts, so free chloride is present from the first instant to the last and no part of the batch ever sees silver in excess. Silver in excess, even locally and briefly, is where fog nuclei and coarse grains come from.

One minute’s ripening at 45 °C. Barely any. Enough for the smallest crystals to redissolve and redeposit on the larger ones, and not enough for that process to run far.

Solution No. 4, then twenty minutes. This is the conversion. Bromide and iodide arrive in the vessel where the silver is already locked up as chloride, and the chloride is displaced. The twenty minutes is the time the conversion and the recrystallisation that follows it are given.

1206 g of dry gelatin, then fifteen minutes at 45 °C while it dissolves. The bulk gelatin comes in after the crystals exist, which is the standard division of labour in every emulsion in this formulary: a little gelatin during precipitation to keep the crystals apart, a lot of gelatin afterwards to carry them.

Set, shred, wash until free from all soluble bromide. Setting and shredding turn the emulsion into noodles with an enormous surface area, and the wash then takes the potassium nitrate and the excess halides out by diffusion. Kodak’s instruction here is unusually absolute — free from all soluble bromide — because the next step is to put a measured amount of soluble halide back, and that is only a measurement if the starting point is zero.

About 770 c.c. of 10 per cent potassium chloride, then water to 18 litres. The soluble halide goes back in as chloride, not as bromide, and the amount is specified in the description: the finished emulsion should carry above 0.03 g of potassium bromide per litre, preferably 0.03 to 0.3 g per litre, or the equivalent of any other bromide or chloride, equivalent meaning the same suppressing action on the silver ion.

🧮Why Kodak had to define equivalent by action and not by weight

The course’s arithmetic, from the solubility products in OpenStax’s Appendix J at 25 °C: silver chloride 1.6 × 10⁻¹⁰, silver bromide 5.0 × 10⁻¹³, silver iodide 1.5 × 10⁻¹⁶.

A soluble halide suppresses the silver-ion concentration by the common-ion effect, and how well it does so depends on the solubility product of the silver salt it forms:

[Ag⁺] = K_sp / [X⁻]

The silver-ion ceiling set by a soluble halide

Take the 10 per cent potassium chloride as 10 g in 100 mL, so 770 mL carries about 77 g. In 18 litres that is 0.0574 mol/L of chloride, and the silver-ion ceiling is 1.6 × 10⁻¹⁰ / 0.0574 = 2.8 × 10⁻⁹ mol/L, a pAg of 8.55.

Now do it with bromide. Kodak’s floor of 0.03 g of potassium bromide per litre is 2.52 × 10⁻⁴ mol/L, and the ceiling it sets is 5.0 × 10⁻¹³ / 2.52 × 10⁻⁴ = 2.0 × 10⁻⁹ mol/L, a pAg of 8.70. The preferred upper end, 0.3 g/L, gives pAg 9.70.

So 77 grams of potassium chloride does roughly the work of 0.02 grams of potassium bromide per litre — the two land within about a tenth of a pAg unit of each other, and the chloride figure sits just outside the loose end of Kodak’s stated range. Whether it sits inside or just outside depends on which table of solubility products you use: published values for the silver halides differ, sometimes in the exponent, which is why this course names one table and uses it throughout. The point that survives every table is the ratio. Silver chloride is about 320 times more soluble than silver bromide, so it takes about 320 times as much chloride to hold the silver ion down to the same level, and a rule written in grams of one halide is meaningless for another. That is exactly why Kodak wrote equivalent and then defined it as the same suppressing action.

No digestion. The formula ends at the make-up. Everything a normal emulsion does next — remelt, hold at 50 to 60 °C for an hour, watch the speed climb — is what this one is written to avoid.

Behaviour

It is slow where you look for speed. The patent publishes no speed figures, no characteristic curve and no developing times, and this course has made no measurement, so nothing quantitative can be said. What the construction says qualitatively is unambiguous: an emulsion precipitated in inert gelatin, without ammonia, below 50 °C, ripened for twenty-one minutes and never digested has almost none of the surface sensitivity that the same silver halide would have after the ordinary treatment. Sheppard’s own patent measures that difference at his end: an emulsion made from a gelatin whose natural sulphur compound has been oxidised out is of impractically low light-sensitiveness, and adding the sulphur compound back raised it, in his cases, ten or over twenty times.

The image it does form is internal. The abridgment of the British specification states that the emulsions, after exposure, may be developed in an internal developer. A surface developer reaches only latent-image specks on the outside of a grain; an internal developer, or a surface bleach followed by development, reaches the ones inside. An emulsion of this kind, exposed and put into an ordinary developer such as D-19, gives very little, and that is the result the design is aiming at.

Keeping is not published. No shelf life, no capacity and no storage temperature appear in the patent, and none is offered here. What can be said is that the formula contains no bacteriocide of any kind — no phenol, no thymol — which for a warm nutrient jelly is a real limitation, and that the absence is Kodak’s, not this course’s editing.

Its restraint is built in rather than added. Most emulsions get their antifoggant in the finals as a small dose of a bromide or an organic stabiliser. This one gets it by having the wash removed every soluble halide and then a measured amount of potassium chloride put back. The effect is the same — the silver-ion concentration in the coated layer is held down, so silver has less opportunity to reduce itself without light — and the method is more exact, because it starts from a known zero.

Image characteristics

Almost nothing here is published, and the honest entry for most of the fields is that Kodak did not say. The patent is a process claim; its examples exist to show that the process can be carried out, not to characterise the product.

What is published is the halide. The emulsion is a silver iodobromide in which the iodide is 21.0 mole per cent of the silver, or 24.9 per cent by weight of the silver halide — the claims are drawn to at least approximately 6 per cent and the description prefers 10 to 20 per cent, and Example I sits at the top of that. That is a very high iodide content by any ordinary standard. Fuji’s 1984 specification, written for emulsions that were then in production, states that among silver bromoiodide emulsions those containing about 10 mole per cent or less of silver iodide are the most preferred; Trumm’s bromide paper emulsion carries one mole per cent. Example I sits at twice the top of Fuji’s preferred band and at twenty-one times Trumm’s. What high iodide buys, and what it costs, is under Interactions.

The rest — grain size, size distribution, habit, contrast, tonality, image colour — is unpublished and is recorded as unpublished in the data above rather than guessed at. A reader who wants a published grain size for a double-jet emulsion should go to Kodak’s own two: Illingsworth’s Example 1 gives about 0.2 µm for cubic-regular silver bromoiodide, and Berry and Marino’s Example 7 gives 0.15 to 0.20 µm edge length for a cubic silver bromide, both from double jets at a controlled pAg.

The mechanism

The double jet, and what pAg is doing

A silver halide crystal grows by silver ions and halide ions arriving at its surface faster than they leave. Which faces grow, and how fast, depends on the ratio of the two ions in the liquid, and that ratio is what pAg measures: the negative logarithm of the silver-ion concentration, exactly as pH measures the hydrogen-ion concentration. Because the solubility product ties the two together, fixing pAg fixes the halide concentration as well.

pAg = −log[Ag⁺], and the product of [Ag⁺] and [X⁻] is K_sp, so fixing one fixes the other

Why one number is enough

In a single jet, pAg starts very high — halide in gross excess, almost no free silver — and falls through the whole addition. Every crystal in the batch therefore grows through a different history, which is why a single-jet emulsion has a broad spread of sizes. Kodak’s own patent on the point, US 3,655,394, states it as a design rule: the solutions are run simultaneously into a rapidly agitated solution of a peptiser, and quantities of halide ion greatly in excess of silver ion are to be avoided if the grains are to come out regular. It sets the window at pH no more than 4.0 and pAg between 8.6 and 9.2, and cites Berry and Skillman for the finding that a change in pAg alone decides whether a double run gives regular cubes or octahedra.

Davey and Knott’s double jet has no electrode in it. Their control is the composition of the vessel and the two flow rates: potassium chloride already present in Solution No. 1, a chloride jet carrying a 17 per cent molar excess over the silver jet, and a ninety-second run. That is a coarse form of the same idea — hold the halide in excess everywhere, all the time — and it is what a 1947 laboratory could do without a silver ion-selective electrode.

🧮The excess, in moles

The course’s arithmetic from the weights in the table. Silver nitrate 1000 g at 169.87 g/mol is 5.887 mol of silver. Solution No. 2 carries 513 g of potassium chloride at 74.55 g/mol, which is 6.881 mol — a 16.9 per cent excess over the silver in the jets alone. Solution No. 1’s own chloride, read as 102.6 g, adds 1.376 mol, bringing the total chloride to 8.258 mol and the excess over silver to 40.3 per cent. If that figure is a transcription duplication and the true weight is 10.26 g, the excess is 19.2 per cent instead. Either way the chloride is in excess throughout, which is the condition the process requires; the two readings differ in how far from the silver-ion end of the scale the precipitation sits, and therefore in grain size, which the patent does not report.

The conversion: why you would make the wrong halide first

The step that makes this emulsion an iodobromide is not a precipitation at all. The silver is already a solid — silver chloride — and Solution No. 4 changes what it is made of without ever dissolving it back to free silver ions in bulk.

AgCl + Br⁻ → AgBr + Cl⁻

Conversion of the chloride to the bromide

AgCl + I⁻ → AgI + Cl⁻

Conversion of the chloride to the iodide

Both reactions go essentially to completion, and the reason is the solubility products. The equilibrium constant for the first is the ratio of the two: 1.6 × 10⁻¹⁰ divided by 5.0 × 10⁻¹³, which is 320. For the second it is 1.6 × 10⁻¹⁰ divided by 1.5 × 10⁻¹⁶, which is about 1.1 million. Iodide beats bromide for the same silver by a further factor of about 3,300, so when Solution No. 4 arrives carrying both, the iodide is taken up first and essentially completely.

That is the chemistry. The photographic consequence is the reason for the whole route. A crystal grown directly from bromide and iodide is one thing; a crystal that was silver chloride and has been converted from the outside inwards is another. The conversion begins at the surface and works in, and the interior of the grain retains a different history from the shell — which is precisely the structure an internal-image emulsion needs, because the traps that hold photoelectrons deep in the crystal are the ones a surface developer cannot reach.

🔬What is left in the liquor afterwards

Solution No. 4 brings 6.897 mol of bromide and 1.236 mol of iodide against 5.887 mol of silver. If the iodide is taken up completely, the emulsion holds 1.236 mol of silver iodide and 4.651 mol of silver bromide, and 2.247 mol of bromide are left over in the liquor along with all 8.258 mol of the chloride that has been displaced. That is why Kodak’s next instruction is a wash, and why it is a wash to exhaustion.

Sulfur sensitisation, which this formula is written to prevent

Samuel Sheppard’s patent of 1926 is the origin of the practice and states the mechanism in the words of the man who found it. The sensitiveness of emulsion grains, he writes, corresponds to the presence in them of nuclei of unstated chemical composition; he attributes the action of his compounds to their forming in the grains small, mostly ultramicroscopic nuclei of silver combined with sulphur, selenium or tellurium — such as silver sulfid.

2 Ag⁺ + S²⁻ → Ag₂S

The sensitivity speck, in its simplest form

The equation above is the simplest formulation of what such a nucleus might be and is not Sheppard’s own: he is careful to say that the composition of the nuclei is unstated, and identifies them only as silver combined with sulphur. The reading that has since become standard — and it is a reading, not a quotation — is that a speck of silver sulfide on a crystal face acts as a trap: a photoelectron freed somewhere in the crystal is caught there rather than wandering until it recombines, and the silver atoms that follow it build the latent image at that one spot instead of scattering over the whole surface. Concentration is what buys speed, because the same number of absorbed photons then produces fewer, larger and more developable specks.

Sheppard’s own quantities show how little of it is needed. He gives 2 to 3 grains of a sensitising compound such as thiosinamine to 100 pounds of dry gelatin, which is 2.9 to 4.3 parts per million of the gelatin, and states the same limit a second way as 1 part of thiosinamine to 300,000 parts of emulsion computed dry, which is 3.3 parts per million. He also states the failure mode: too high a concentration causes fog, and sometimes reversal.

Now read Davey and Knott’s first ingredient again. Inert gelatin is gelatin from which that sulphur compound has been removed — Sheppard’s own patent describes making it by oxidising the natural sensitiser out with a peroxide or by introducing the oxidation into the liming stage of gelatin manufacture. Choosing inert gelatin is not a neutral choice of raw material; it is the removal of the sensitiser at source. The British abridgment says so directly: the colloid binder contains little or no sulphur sensitizers, e.g. inert gelatin.

Gold sensitisation, and why it needs the sulfur first

Waller, Collins and Dodd’s Ilford patent of 1946 is the disclosure the whole industry cites for gold. Its finding is that adding a very small quantity of a soluble gold salt to a gelatino-silver halide emulsion before or during digestion, with the emulsion held on the acid side of neutrality and below about pH 8, gives a very considerable increase in sensitivity. The claimed quantity, for an emulsion of normal silver halide content — Waller’s own example is 2 to 3 per cent calculated as metallic silver — is an amount of gold salt per litre equal to that in 1 to 100 cc of a 1 in 10,000 solution. That is 0.1 to 10 milligrams of the salt in a litre carrying 20 to 30 grams of silver, so between roughly 3 and 500 parts of gold salt per million parts of silver.

Three things in that patent matter more than the number. The first is that Waller states the gelatin’s own sulphur compounds are believed essential to the effect, so that inert gelatins benefit from a positive addition of one: gold is applied on top of sulfur sensitisation, not instead of it. The second is that including a small quantity of an alkali or ammonium thiocyanate with the gold improves the result further — the thiocyanate holds the gold in solution as a complex so that what the emulsion meets is a gold(I) species rather than a gold(III) one, which is the chemistry the course’s Steigmann gold sensitiser stock is built on. The third is the pH condition, which is a condition on fog: too alkaline a digestion and the gold makes fog faster than it makes speed.

The industrial afterlife of that patent is worth stating, because it is unusual. Kodak’s Example 1 in US 3,655,394 says its cubic-regular silver bromoiodide is sulfur- and gold-sensitized as described in Waller, U.S. Pat. No. 2,399,083. Thirty-eight years after Waller, Fuji’s US 4,469,784 says of its own double-jet emulsions that each was sulfur- and gold-sensitized as described in U.S. Pat. No. 2,399,083. One 1946 Ilford patent is still the reference the industry points at, and that is why this page cites it rather than paraphrasing a textbook.

⚗Three published sulfur-and-gold doses, side by side

None of these is a dose for the formula on this page. They are what the literature prints for the operation this formula omits, and they are given here so that the omission has a size.

Berry and Marino, Eastman Kodak, 1973. A cubic silver bromide emulsion made by double jet into bone gelatin at pH 6.0 and pAg 9.0, kettle at 70 °C, run time 42 minutes, grains 0.15 to 0.20 µm. Sensitised with 10 mg of sodium thiosulfate per silver mole for 90 minutes at 55 °C, or with 10 mg of sodium thiosulfate and 5 mg of potassium chloroaurate per silver mole for 60 minutes at 55 °C.

Heki and Mifune, Fuji, 1984. Post-ripening at 60 °C and pAg 8.0 with 5 × 10⁻⁶ mol of sodium thiosulfate, 3 × 10⁻⁶ mol of chloroauric acid and 4 × 10⁻⁴ mol of potassium thiocyanate per mole of silver. Converting with the course’s own arithmetic: about 1.24 mg of sodium thiosulfate pentahydrate (0.79 mg if the anhydrous salt is meant, which the patent does not say), 1.02 mg of chloroauric acid and 38.9 mg of potassium thiocyanate per mole of silver.

Waller, Collins and Dodd, Ilford, 1946. 1 to 100 cc of a 1 in 10,000 solution of the gold salt per litre of emulsion, digested and coated below about pH 8.

The two thiosulfate doses do not agree, and the disagreement is informative: Kodak’s 10 mg per mole against Fuji’s 0.8 to 1.2 mg per mole is an order of magnitude, and Kodak’s is sulfur on its own while Fuji’s is sulfur alongside a thiocyanate at forty times the weight of either sensitiser, which changes the whole balance of the digestion. The gold doses are closer — 5 mg of potassium chloroaurate against 1 mg of chloroauric acid per mole, the same order — and both are, as Waller says, far too little to change the colour of the image. What the three have in common is the scale: a sensitiser is measured in milligrams against a mole of silver, that is, against something like a hundred grams of silver halide.

The stabiliser this course could not otherwise price

One number in Illingsworth’s patent is worth recording here because the course’s own research pass concluded that no source it had read gave it. Example 6 of US 3,655,394 adds, to a sulfur- and gold-sensitised cubic-grained emulsion, 4 g of 4-hydroxy-6-methyl-1,3,3a,7-tetraazaindene per silver mole, together with 2.5 mL of 6-methoxyquinoline per silver mole and an infrared sensitiser at 400 mg per silver mole. The course’s tetraazaindene page records that no dose for a hand-made emulsion exists at any tier; this is a dose for an industrial one, and it is four grams per mole, which is a thousand times the gold and four hundred times the sulfur. A stabiliser is not a trace additive.

Fog is a condition of the make, not only of the additives

Illingsworth’s Example 5 is the sharpest demonstration in the literature the course has read of how completely the precipitation decides what the sensitisation will do. Two cubic emulsions, identically sulfur- and gold-sensitised and identically spectrally sensitised, differing only in the pH at which their grains were precipitated: the ones made at pH 2.0 come out at a fog of 0.04, and the ones made at pH 5.6 at 0.79 to 0.97. Twenty times the fog, from a decision taken before any sensitiser was added. Illingsworth’s reading is that the higher pH allows reduction sensitisation during precipitation — specks of metallic silver formed without light — and that sulfur and gold then build on those specks as readily as on the ones exposure makes.

Function of every ingredient

Gelatin, 102.6 g in Solution No. 1 and 1206 g added dry afterwards. Gelatin does three separate jobs in this formula and the two doses do two of them. The 102.6 g in the vessel is the peptiser: it adsorbs on the crystal faces as they form and keeps the particles from coalescing into a useless precipitate. It is deliberately small — about one thirteenth of the total — because gelatin also retards crystal growth, and a heavy dose during precipitation gives finer grains than the process wants. The 1206 g added after the conversion is the binder: it gives the emulsion enough body to set into a firm jelly, survive shredding, hold together through the wash, and coat. More gelatin here makes a tougher, slower-washing noodle and a thicker coating per unit of silver; less makes an emulsion that will not set cleanly and whose noodles disintegrate in the wash water. The third job — supplying sulphur — is the one that has been designed out. Both doses are inert gelatin, chosen because the sulphur compounds that ordinary photographic gelatin carries would sensitise the grains, and this emulsion must not be sensitised.

Potassium chloride, 102.6 g in Solution No. 1, 513 g in Solution No. 2, and about 770 mL of a 10 per cent solution at the end. Three appearances, three different jobs.

In Solution No. 1 it is the halide already in the vessel when the jets start, and it fixes the silver-ion concentration at the instant of nucleation, when the number of crystals in the finished emulsion is decided. More of it means a lower silver-ion concentration, more nuclei and a finer grain; less means the reverse. This is the one quantity on the page the course cannot vouch for absolutely: the Google Patents text prints 102.6 g here, which is the same figure it prints for the gelatin on the line above, and a transcription that duplicated the preceding line cannot be ruled out. The British abridgment of the same specification confirms that the vessel does carry potassium chloride of its own, but prints no weight. At 102.6 g the total chloride runs 40.3 per cent in excess of the silver; at 10.26 g it runs 19.2 per cent. Both are working excesses and the process does not fail at either, so the doubt affects the grain size rather than the chemistry. The course prints what its copy prints and says this.

In Solution No. 2 it is the halide of the first precipitation, 513 g against 1000 g of silver nitrate, a 16.9 per cent molar excess. Chloride is used rather than bromide because the whole route depends on starting from a silver salt more soluble than silver bromide, so that the later conversion has somewhere to go. Substituting bromide here would not be a variant of this formula; it would abolish it.

At the end, 770 mL of a 10 per cent solution is the soluble halide put back after the wash. Its job is restraint: it caps the silver-ion concentration in the finished, coated emulsion, so that silver has less chance to reduce itself to metal without light. Kodak specifies how much in terms of an equivalent weight of potassium bromide, 0.03 to 0.3 g per litre, and the arithmetic that reconciles the two is under Mixing. More of it means a cleaner, slower, longer-keeping emulsion; less means fog. Take it out altogether and the patent’s own condition is broken.

Silver nitrate, 1000 g in Solution No. 3. The silver: 5.887 moles of it, which at the finished 18 litres is 35.3 g of silver, or 55.6 g of the nitrate, per litre. Nothing else in the formula makes an image, and every other ingredient is there to control what happens to this one. It is delivered dissolved and hot enough not to crystallise, through one of two jets, over ninety seconds. Its rate is half of the pair of rates that sets the silver-ion concentration in the vessel: run it faster than the chloride and the excess collapses, the vessel goes silver-rich, and fog nuclei and coarse crystals follow. Its total decides the strength of the emulsion; more silver in the same volume of gelatin gives a heavier coating weight and a denser maximum black but a poorer ratio of gelatin to halide, and this formula’s ratio is already 1.12 grams of gelatin per gram of silver halide.

Potassium bromide, 820.8 g in Solution No. 4. The bromide that converts the silver chloride to silver bromide, and the bulk of the light-sensitive halide in the finished emulsion. 6.897 moles against the 4.651 moles of silver left after the iodide has taken its share, so an excess of 2.247 moles — 48 per cent more bromide than there is silver for it to find. The excess is there to drive the conversion to completion rather than to leave a partly converted crystal, and it is the reason Kodak’s wash instruction is to wash until free from all soluble bromide. Reduce it and conversion is incomplete, leaving silver chloride in the core; increase it and you have more to wash out and a longer wash.

Potassium iodide, 205.2 g in Solution No. 4. 1.236 moles, which is 21.0 mole per cent of the silver and 24.9 per cent by weight of the finished silver halide. This is the ingredient the patent’s claims are actually drawn to: at least approximately 6 per cent silver iodide, preferably 10 to 20 per cent. Iodide does several things at once and they are set out under Interactions; the one that matters here is that silver iodide is by far the least soluble of the three halides, so iodide added to a chloride emulsion is taken up first, completely, and irreversibly. Less iodide and the emulsion fails the claim and, on the patent’s account, loses the internal-image behaviour it is built for; more and the emulsion becomes harder to develop and harder to fix, because silver iodide is the halide that thiosulfate dissolves most reluctantly.

Interactions

Chloride against bromide against iodide. The three soluble halides in this formula are not interchangeable and the differences are all one number, the solubility product. Chloride makes the most soluble silver salt, which is why it is used for the skeleton that will later be converted. Bromide makes a salt 320 times less soluble, which is why bromide displaces chloride completely. Iodide makes one about a million times less soluble than the chloride, which is why iodide goes first and why no later operation will take it out again. Every design decision in the formula is a consequence of that ordering.

Iodide, and what twenty-one per cent of it costs. That a small proportion of iodide in a bromide emulsion raises its speed is the reason iodobromide emulsions exist at all, and it is why every negative emulsion in this formulary carries some. The course has not read a document at the standard its sources policy requires that sets out why, so no mechanism is asserted here; the patent’s own reference list points at Huse and Muelendyke on the spectral sensitivity of mixtures of silver iodide and silver bromide, in The Photographic Journal for June 1926, which this corpus does not hold. What can be said, and it follows directly from the solubility products already quoted, is what twenty-one per cent costs at the other end of the process. Silver iodide has a solubility product about three thousand times smaller than silver bromide’s, so it is much the hardest of the three halides for a fixer to take into solution, and an emulsion of this composition will fix slowly and can be left with residual halide if it is timed like an ordinary film. The patent quantifies none of this, and neither does this page.

Inert gelatin and everything that would have sensitised it. Ordinary photographic gelatin carries traces of naturally occurring sulphur compounds, and Sheppard’s discovery was that those traces, not the protein, are what makes an emulsion fast. Waller’s patent restates it from the gold side: the presence of such sulphur compounds is believed essential if the best results are to be obtained from gold, so inert gelatins benefit by an addition of a small quantity of such a compound. Davey and Knott’s choice of inert gelatin therefore disables sulfur sensitisation and gold sensitisation at the same stroke, because gold has nothing to build on.

The absence of ammonia. Ammonia is the classic ripening agent: it complexes silver ion, increases the solubility of silver bromide, and lets small crystals dissolve and redeposit on large ones far faster than heat alone. The patent’s own framing is that ammonia is ordinarily used during emulsification, and that a neutral emulsion uses none and is precipitated at 70 to 90 °C instead. This formula uses neither route: no ammonia and no high temperature. Ripening is therefore minimal, which keeps the grain small and the sensitivity low, and both are wanted.

Temperature and time as one control. Twenty-one minutes of ripening at 45 °C is a very small thermal budget. Trumm’s bromide paper emulsion gets two hours at 60 °C; Wall’s gaslight paper gets seventy-five minutes at 50 °C. Raising the temperature or extending the time here would coarsen the grain and raise the fog, and would begin to give the emulsion the surface sensitivity the design excludes. That is why the description sets a ceiling — not above about 50 °C — rather than a target.

pH. The British abridgment permits 4.5 to 7.0 for the emulsion. That is on the acid side of neutral and is consistent with everything else: acid conditions during precipitation suppress the reduction sensitisation that Illingsworth’s Example 5 shows to be the source of fog in a sulfur- and gold-sensitised emulsion, and they slow the ripening that this formula is trying to limit.

Variants

The same patent’s Example II is the same idea run in stages. It keeps Solutions No. 2 and No. 3 and the ninety-second double jet, but converts in two steps — a small mixed bromide-chloride-iodide addition over two minutes, then a large bromide-and-iodide addition over twenty-five minutes — and adds ammonia before the bulk gelatin. This course does not print it: its water figure for one solution reads improbably in the copy available, one temperature is lost in the scan, and it introduces ammonia, so it would carry three separate caveats and add nothing the reader cannot get from Example I.

The same patent’s Examples III and IV, which the course knows only through the British abridgment, replace the chloride skeleton with something else again: in one, silver thiocyanate is formed from potassium thiocyanate and silver nitrate, and in the other sodium carbonate is used. Both are then converted with bromide and iodide in the same way. The description names silver thiocyanate alongside silver chloride as a salt more soluble in water than silver bromide, so the principle is identical and only the skeleton changes.

The cubic-regular route, Kodak 1972. Illingsworth’s US 3,655,394 is what a double jet becomes once an electrode is available. Instead of setting the excess by weighing chloride into the vessel, the pAg is measured continuously and held between 8.6 and 9.2 by adjusting the relative flow rates, with the pH kept at or below 4.0 with sulphuric acid. The result is grains bounded by cube faces — Kodak’s cubic-regular — and, crucially, free of reduction sensitisation, so that sulfur and gold can be applied afterwards without the fog that a pH 5.6 make would give. This is the emulsion the register’s name describes, and Kodak published its conditions and not its weights.

The pAg-controlled bromide route, Kodak 1973. Berry and Marino’s US 3,772,031 runs the same operation on silver bromide at pH 6.0 and pAg 9.0, 70 °C, 42 minutes, and gives the sulfur-and-gold doses that go with it. It is the closest thing the course has read to a fully specified modern negative-emulsion make.

No scaled-down version is offered. Scaling this formula to a bench is arithmetic anyone can do, and the result would be a formula nobody has made, for a process this course does not teach, requiring an internal developer it does not publish. Rule 6 forbids presenting that as a variant of Kodak’s, and there would be no point in it.

📓What the schema could not hold, and what the register asked for

Two shortfalls are recorded here rather than absorbed. The schema’s ways of combining solutions are a ratio, an order of addition and a sequence of baths, and none of them can express simultaneously — which is the single word that distinguishes a double jet from a single one. Solutions No. 2 and No. 3 are recorded as consecutive parts of an order of addition, and the instruction beneath the table is what tells the reader the truth. And the register’s name for this entry promises sulfur-and-gold sensitisation, which this formula does not have and is built not to have; the reasons are in the provenance note above and the chemistry is under The mechanism.

Safety

The classification is Level B, the same as every other silver nitrate procedure in this course, and it is a classification of the chemistry rather than of the scale. Nothing in the formula is acutely toxic, nothing evolves a hazardous gas, no step is carried out above about 50 °C, and there is no ammonia — which removes the one respiratory hazard that most historical negative emulsions carry. What raises it above Level A is silver nitrate and the fact that the work is done in the dark.

Silver nitrate is the hazard on the page. PubChem’s aggregated ECHA classification carries the signal word Danger with four pictograms: oxidising solid, corrosive, health hazard and hazardous to the aquatic environment. It causes severe skin burns and eye damage, it is an oxidiser and must be kept away from combustible material, and solutions of it stain skin and everything else black on exposure to light, and the stain is metallic silver held in the keratin, so it wears off as the skin does rather than washing off. Weigh it with the room light on, in a dedicated vessel, wearing nitrile gloves and eye protection, and never near paper towel or a solvent. One kilogram is a great deal of it; a spill is an expensive, staining, oxidising mess.

The halides are the least of it. Potassium chloride is classified by the large majority of ECHA notifiers as not meeting GHS hazard criteria; potassium bromide and potassium iodide carry only irritant classifications in the aggregated notifications. They are still laboratory chemicals and are still weighed with gloves on, because the point of a dedicated set of utensils is that nothing crosses between the darkroom and the kitchen.

Gelatin is not a chemical hazard. It is a biological one in the mundane sense: a warm gelatin solution is a nutrient broth, this formula contains no bacteriocide at all, and eighteen litres of it left at room temperature will spoil. Handle it as you would handle food that you are not going to eat.

Working in the dark is the hazard the ingredient list does not show. Hot solutions, a kilogram of silver nitrate and two simultaneous jets, under a safelight the emulsion will tolerate, is a combination where the ordinary precautions — a clear bench, everything at a known place, a container that cannot be knocked over, no glass on the floor — stop being good practice and become the control.

The sensitisers discussed on this page are not in the formula and are not innocuous. Chloroauric acid carries the signal word Danger with four pictograms. Thiosinamine, the sulfur sensitiser Sheppard and Waller both name, has no entry in this course’s chemical encyclopaedia and no hazard classification is offered for it here. Nobody should read the doses in the callouts above as a licence to perform a digestion; they are cited so that the absent operation has a published size.

Storage

The patent publishes nothing about keeping, and none is invented here.

What the formula itself implies is worth stating. An emulsion with no bacteriocide has to be kept cold, in the dark and covered, and used or coated promptly; Wall’s and Baker’s commercial formulas both carry a preservative for exactly this reason, and this one does not. Set emulsion keeps better than melted emulsion because the gel restricts diffusion and because it is not being held at a temperature bacteria enjoy. Every remelt costs something — a little more ripening, a little more fog — so an emulsion is remelted as few times as the work allows.

Of the ingredients rather than the product: silver nitrate is kept in a tightly closed amber or opaque bottle away from light, organics and reducing agents; potassium iodide slowly yellows in air and light as iodide oxidises to iodine, and a yellowed sample is no longer the weight of iodide the label says; gelatin is kept dry and cool, and takes up moisture from the air, which is why a formula’s gelatin weights are for the dry material.

Incompatibilities

Silver nitrate and organic matter. It is an oxidiser. Paper, cloth, dust, alcohol and gloves are all fuel. This is the reason silver nitrate is weighed on a dedicated, non-absorbent surface.

Silver nitrate and ammonia, in any combination, unsupervised. This formula contains no ammonia and this course does not run ammoniacal makes, and the reason is not only the vapour. Silver solutions made strongly alkaline with ammonia and left to stand or to dry can deposit fulminating silver, which is a contact explosive. An ammoniacal silver solution is quenched, never stored and never allowed to dry out.

Silver nitrate and chloride, everywhere except in the vessel. Any chloride — tap water, sweat, a rinse that was not thorough — precipitates silver chloride where it lands. Distilled or deionised water throughout, and a rinse regime that keeps the chloride solutions and the silver solution apart until the jets, is part of the formula rather than good housekeeping.

Iodide and oxidising agents. Iodide is easily oxidised to iodine, which stains and which is not the ingredient the formula calls for. Keep the potassium iodide away from the silver nitrate container and from any peroxide or persulfate in the same cupboard.

Emulsion and metal. Stainless steel is fine; copper, brass and iron are not. Trace metal ions in an emulsion catalyse the reduction of silver halide and show up as fog.

Waste

Everything in this formula ends up in one of three streams and only one of them is difficult.

The wash water carries silver, in two forms: dissolved silver complexes and fine silver halide that escaped the noodles. It is not a rounding error: the whole purpose of the wash is to take soluble material out of an emulsion holding 635 grams of silver, and noodles shed fines while they are washed. Silver is toxic to aquatic life at very low concentrations, and the ECHA classification of silver nitrate carries the environmental pictogram for exactly that reason. Collect the wash water rather than pouring it away.

Spent or spoiled emulsion is silver-bearing waste in its most concentrated form. Eighteen litres of this formula holds about 635 grams of silver. It goes to a licensed waste contractor, or through a silver-recovery route, and never down a drain.

The soluble salts — potassium nitrate, potassium chloride, potassium bromide, potassium iodide — are the innocuous part chemically, and are still part of the silver-bearing stream while they are mixed with it.

Local regulation governs, always. What may be discharged, at what concentration, and under what consent, differs between jurisdictions and changes; nothing on this page is a discharge consent, and the only universal instruction is to find out what your own authority requires before you make anything at this scale.

Troubleshooting

The patent reports no faults, so what follows is reasoning from the formula and from the failure modes the sources named on this page do describe. It is marked as such rather than presented as Kodak’s experience.

A grey or brown cast in the finished emulsion, before any exposure. Silver reduced to metal without light. The usual causes at this stage are silver arriving faster than halide somewhere in the vessel, a trace of a reducing agent or a metal ion, or an over-warm ripening. In a double jet the first of those is a flow-rate problem, and the vessel’s own chloride is the buffer against it.

A coarse, sandy or gritty emulsion. Too few nuclei formed at the start, so what silver there was grew onto too small a number of crystals. Lower chloride in the vessel, a slower jet, or a warmer vessel would all push in that direction. This is the failure the doubt over Solution No. 1’s chloride weight bears on directly.

Incomplete conversion — a bluish or violet cast, or a material that fixes far faster than it should. Silver chloride left in the core because the bromide was insufficient or the twenty minutes was cut short. The excess bromide is there precisely to prevent this.

An emulsion that will not set, or noodles that disintegrate in the wash. Not enough gelatin, gelatin of too low a bloom strength, or a wash that was too warm. The formula’s 1206 g is added for this job and the wash is done cold for the same reason.

Fog that appears on keeping rather than immediately. The soluble halide put back after the wash is what holds this off. Too little chloride at the end, or a wash that did not in fact remove the bromide it was supposed to before the chloride was measured in, and the finished emulsion’s silver-ion concentration is not where Kodak specified it.

A spoiled, foul-smelling batch. There is no bacteriocide in this formula. Cold, dark, covered and used promptly is the whole of the defence.

No image at all under an ordinary developer. Expected, and not a fault. This is an internal-image emulsion, and a surface developer is not the tool for it.

Experiments

Recompute the conversion for a different iodide. Take the weights in the table, change the potassium iodide to 6 mole per cent of the silver — the bottom of the patent’s claim — and work out the new weight, the new free-bromide excess after conversion, and the new percentage by weight of silver iodide in the finished halide. Then ask what would have to change in the wash. The arithmetic under The mechanism is the worked version at 21 per cent.

Predict the effect of the disputed weight. Solution No. 1’s chloride is either 102.6 g or possibly 10.26 g. Using the solubility product of silver chloride, compute the silver-ion concentration in the vessel at the moment the jets start, under each reading, and say in a sentence which way the grain size would move. This is the same calculation the maths callout under Mixing performs for the finished emulsion, run at a different point in the process.

Compare the thermal budgets of the four emulsions in this formulary. Tabulate the ripening temperature and time for this formula, for Trumm’s bromide paper, Wall’s gaslight paper and the chloride contact emulsion. Rank them, and then rank them again by the surface sensitivity each is trying to achieve. The two orders should agree, and the disagreements are the interesting part.

Size the omitted digestion. Using the published doses in the callout under The mechanism, work out how much sodium thiosulfate and how much gold salt an eighteen-litre batch of this emulsion would take if it were sensitised the way Berry and Marino sensitised theirs. Express both as a fraction of the silver present. The answer is a good corrective to the intuition that a sensitiser is an ingredient like the others.

Read the two double jets against each other. Set Davey and Knott’s method of control — a weighed chloride excess in the vessel and two fixed flow rates — beside Illingsworth’s — a silver ion-selective electrode holding pAg between 8.6 and 9.2 while the flow rates are adjusted to suit. Write down what each method can and cannot hold constant, and what measurement would tell you which one you had used.