Washing, Digestion and Sensitisation

At the end of ripening you have crystals of about the right size suspended in gelatin, and a beaker of dissolved salts you never wanted. You also have, on Duffin’s own numbers, an emulsion whose speed is somewhere around one seventieth of what the same crystals could reach. Both problems are solved in this lesson, and they have to be solved in that order: you cannot sensitise an emulsion you have not washed.

What washing removes, and what each thing costs if it stays

Duffin lists what is in the liquid at the end of ripening: the by-products of the double decomposition, whatever acid or ammonia was used, and the excess halide ions. Four separate problems follow.

Crystallisation on a non-absorbent support. The nitrate salt is soluble and stays soluble right up to the moment the coating dries, at which point it crystallises in the layer. Duffin: it “would be impossible for an emulsion of this composition to be coated on film because the salts, etc. still present would crystallize at the drying stage and render the physical condition of the coated layer unsuitable or, in some cases, prevent complete drying.” Ross gives the same fact from the maker’s side and draws the line the projects in this part are built on: it is rarely a problem for paper coating, because most papers absorb the potassium nitrate, but on glass or film the by-products dry and crystallise out on the surface and ruin the plate.

That is why Project 1’s chloride paper is not washed and Project 3’s iodobromide plate must be. Wall states the division as period practice — bromide papers “are always washed emulsions, the gaslight being usually unwashed” — and it is a property of the support, not a shortcut.

Digestion is retarded by excess halide. This is the load-bearing one. Duffin: the great majority of modern emulsions are given a subsequent chemical sensitising treatment, “and this process is retarded by the presence of excess halide.” The bromide that was your ripening agent is a restrainer once ripening is finished; leaving it in means the sensitisation you are about to perform will not work properly.

pH. An emulsion ripened with ammonia is at high pH and “would be even less acceptable for digestion and coating”.

Keeping. A salt-laden emulsion is a worse-behaved emulsion in storage, and Duffin’s storage advice belongs here: hold emulsion below 5 °C, at which “a stable finished emulsion” keeps for several months, while “undigested material is usually kept for only a few days”.

!The one thing washing does not remove

Washing removes ions. It does not remove the sulfur compounds in the gelatin, which are part of the gelatin. Everything the gelatin lesson said about active and inert grades still applies after the wash, and a maker using an active gelatin is still on a shorter clock than one using an inert grade.

Noodle washing, step by step

Noodle washing is the oldest method and, until recently, was the commonest. Duffin’s description of it is four operations, and every home recipe in this part is a version of them.

Noodle washing, in six operations

1. 1Set — cool until the emulsion is a fairly soft jelly; more gelatin is often added first
2. 2Shred — small pieces expose surface area; larger than a pea and the wash time doubles
3. 3Wash — many changes of ice water rather than one long soak
4. 4Diffusion — ions leave the gelatin phase down a concentration gradient; each change restores the gradient
5. 5Endpoint — conductivity or pAg in industry; pH from 10 down to between 7 and 8 in the one published domestic test
6. 6Drain and hold — squeeze, pat dry, refrigerate in the dark; never freeze

Drawn as a sequence, not as apparatus. The one number that is a measurement rather than a practice is the pH endpoint, and it is one practitioner's, recorded twice.

Set. Cool the emulsion, usually after adding a further quantity of gelatin, until it sets to a fairly soft jelly.

Shred. Break it into small fragments. Duffin says “usually by shredding to produce noodles”; the domestic tool is a potato ricer, which Ross notes could pass for a small version of the machines factories used. The governing quantity is surface area, and she gives the rule as a threshold: pieces much bigger than a pea double the washing time.

Wash. Suspend the noodles in water and change the water frequently. Duffin’s mechanism: “during this washing treatment the ions diffuse, by osmosis, out of the gelatin phase into the aqueous phase and are then removed.” Part III’s diffusion and osmosis explains why many short changes beat one long soak — each change restores the concentration gradient that is doing the work.

The published schedules: six changes of three minutes in ice water, about twenty minutes in all for the domestic iodobromide plate emulsion; about twelve changes of five minutes for Baker’s Trumm bromide paper; and an hour and a half in twelve five-minute changes for the practitioner’s ammoniacal plate emulsion.

Endpoint. Duffin gives two industrial tests: total ionic strength by electrical conductivity, and the residual silver ion concentration, pAg, measured potentiometrically with a silver electrode against a standard calomel reference. He states which is better and why: two emulsions with the same excess halide but different total ionic strength will reach a given conductivity with different halide concentrations left, and since it is the halide concentration that critically affects the digestion, “the direct determination of pAg is much to be preferred.”

Neither instrument is in this course’s laboratory. The only published domestic endpoint the corpus holds is Ross’s, and it is worth having exactly: she checked pH after each five-minute change and found the lowering curves identical from 10 to 7.8 on two separate trials, concluding that you need to get the pH between 7 and 8 and that overwashing is to be avoided, but that checking is unnecessary if a careful and consistent wash cycle is maintained. That is one practitioner’s measurement, repeated once, not a standard, and this course publishes it as hers.

Drain and hold. Squeeze the noodles until little or no water comes off, pat them dry, and refrigerate in a lightproof container. Duffin’s temperature is below 5 °C; the practitioner’s limits are several days for washed noodles, and never freeze.

✕The trap in noodle washing, and why the gelatin goes in before the wash

Duffin names two “very grave disadvantages” of noodle washing. The first is that it is slow and needs handling. The second is the one that will bite you: noodle washing cannot concentrate an emulsion, and in fact dilutes it. “As the removal of the inorganic salts proceeds, the emulsion noodles tend to swell and absorb water and it is quite common for the emulsion concentration to have dropped to almost half its initial value by the time the halide-ion concentration is sufficiently low for digestion.”

That matters because a coatable emulsion needs body. Duffin: using gelatins of average viscosity and strength, “the gelatin concentration cannot be permitted to drop appreciably below 5%”. So the second gelatin goes in before the emulsion is set and shredded, not after washing — which is exactly what every recipe in this part instructs, and now you know why.

He also gives the industrial trick for limiting the swelling, which the course records without recommending: wash in a solution of an inorganic salt carrying multi-charged ions instead of in water, magnesium sulfate being the common example, which greatly reduces the water take-up and is not itself taken up to any great extent. No source in this course’s corpus performs that at domestic scale.

The alternatives: coagulation, and what the course cannot describe

Every modern washing method works the other way round: instead of letting the salts diffuse out of a solid gelatin, they make the gelatin come out of the solution. Duffin: “Modern methods of emulsion washing are all based on the principle of causing the gelatin to coagulate and, therefore, separate from the warm emulsion solution when the silver halide is carried with it.” The density of the silver halide makes the mixture settle rapidly as a curd, and the liquid is decanted.

Four routes, all Tier 1 and all described rather than instructed:

• Solvent precipitation. Add a water-miscible organic solvent in which gelatin is not soluble; industrial ethanol or methylated spirit is the classic, and probably the earliest coagulation method used in production. Duffin’s objections are practical: the emulsion comes down “in rather a slimy form which is difficult to handle”, separation is poor, and it takes around a litre of spirit per gram mole of silver.
• Salt precipitation. Add large quantities of a salt carrying at least one doubly-charged ion, or smaller quantities of triply or quadruply charged ions, and the gelatin comes out of the aqueous phase. Ammonium sulfate, borrowed from protein chemistry, is one of the commonest.
• Precipitation by sulfonic acids (Duffin’s own heading spells it “sulphonic”), and acylated gelatin — a gelatin whose amino groups have been chemically modified, most notably by reaction with phthalic anhydride, so that it loses solubility when the pH is dropped to about 4 and can be redissolved by dropping the pH further to about 3. The emulsion is made in the modified gelatin, then acid is added and the whole coagulates.
• Redispersion, which is not a method but the step all of them need. Duffin: without efficient redispersion the grains clump, giving aggregates of developed silver and a coarser result; it needs intensive stirring and the pH for it is frequently critical.

⚗Ultrafiltration and dialysis: named, and not described

Part V’s manifest lists ultrafiltration and dialysis alongside flocculation as alternatives to noodle washing. No source this course has read describes either as a photographic emulsion washing method, gives conditions for one, or reports a result from one. Both are real membrane separations and both are plausible on the chemistry — the thing to be removed is small ions, and the thing to be kept is a colloid and a suspension, which is exactly the separation a membrane makes — but plausible is not published, and Rule 1 stops the course there.

What the corpus supports is the sentence above it: the modern industrial answer is coagulation, in the four forms Duffin lists. If you want to try a membrane method, treat it as your own experiment, record it as one, and do not expect this course to have told you what to expect.

Digestion: a second heat treatment, and what it is for

The washed emulsion has crystals of the right size and is, in Duffin’s words, “still in a relatively insensitive condition”. Digestion is the treatment that fixes that: “a treatment, often after the addition of a chemical sensitizer, at an elevated temperature, from 40° to 70 °C, which may vary in duration from a few minutes up to several hours, when speed and fog are increased.”

Note the last four words. Speed and fog. The whole of digestion is a race between them.

Speed and fog during digestion, after Duffin's Figure 5.1

• Relative log speed
• Fog, a well-behaved emulsion
• Fog, a foggy emulsion

Show the numbers behind this plot

Three curves against digestion time from 5 to 50 minutes for an iodobromide emulsion. Relative log speed rises steeply from about 0.35 at 5 minutes, through 1.0 at about 13 minutes and 1.5 at about 22 minutes, to a maximum near 1.75 at about 37 minutes, then falls back to about 1.3 by 50 minutes. Fog stays almost flat at about 0.05 until roughly 27 minutes, then rises with increasing steepness to about 0.55 at 50 minutes. A third, dashed curve shows the fog of a badly behaved emulsion, which leaves the baseline at about 20 minutes and reaches 0.9 by 50 minutes, so that its fog has overtaken the speed gain long before the speed maximum would have been reached. The useful window is the region where speed is still climbing and fog has not yet left the baseline, and the practical endpoint is the point where fog begins to rise rather than the point of maximum speed.

Series	Digestion time (minutes)	Fog, and related log speed
Relative log speed	5.00	0.35
Relative log speed	10.00	0.80
Relative log speed	15.00	1.15
Relative log speed	20.00	1.42
Relative log speed	25.00	1.60
Relative log speed	30.00	1.70
Relative log speed	37.00	1.75
Relative log speed	43.00	1.68
Relative log speed	50.00	1.32
Fog, a well-behaved emulsion	5.00	0.04
Fog, a well-behaved emulsion	15.00	0.04
Fog, a well-behaved emulsion	25.00	0.05
Fog, a well-behaved emulsion	30.00	0.08
Fog, a well-behaved emulsion	35.00	0.16
Fog, a well-behaved emulsion	42.00	0.32
Fog, a well-behaved emulsion	50.00	0.55
Fog, a foggy emulsion	5.00	0.04
Fog, a foggy emulsion	18.00	0.05
Fog, a foggy emulsion	24.00	0.12
Fog, a foggy emulsion	30.00	0.25
Fog, a foggy emulsion	38.00	0.48
Fog, a foggy emulsion	44.00	0.68
Fog, a foggy emulsion	50.00	0.90

Published data: the axes, the shapes and the two fog cases are Duffin's Figure 5.1 for an iodobromide emulsion; the individual points were read off the printed curves by this course, so treat the values as approximate and the shape as the teaching.

Duffin’s warning about the dashed curve is the one to remember: if the conditions are not optimal in terms of pH, the nature of the gelatin or the quantities added, “the speed increase might be much smaller or the onset of fog might take place sooner”. Two batches with the same schedule and different gelatin can sit on those two different fog curves.

The vocabulary is worth pinning down because the literature uses four words for it. Digestion, chemical ripening and after-ripening all name the same operation, and Duffin’s preferred term, which says what is actually happening, is chemical sensitisation.

Sulfur sensitisation, and the arithmetic of a few milligrams

Sulfur sensitisation is, Duffin says, “the most widely-used method of conferring speed and contrast on a silver-halide emulsion”. The finding behind it is the one the gelatin lesson tells as history: fast gelatins contain substances which form small quantities of silver sulfide when emulsions made in them are digested.

2 Ag⁺ + S²⁻ → Ag₂S

The composition of a sensitivity speck. This is what the speck is made of; the route by which a thiosulfate or a thiourea delivers that sulfide at the crystal surface is not given by any source this course has read.

Modern practice, rather than relying on the gelatin, adds a known quantity of a sulfur compound before digestion. Two are named in the corpus: thiosulfate, and thiourea and its relatives, of which Sheppard’s own example was thiosinamine — allyl thiourea.

Duffin’s published range, and the arithmetic the course needs

This is the only sulfur quantity in the corpus that is stated as a dose, and it is worth setting out exactly as it stands.

The active sulphur quantity present in gelatin varies from 1 to 2 parts per million for the inert types to 100 parts per million for active gelatins. Bekunov gives quantities … of between 25 and 150 parts per million of thiosulphate in a gelatin added to sensitize an iodobromide emulsion containing 3.1 molar per cent of iodide and mean grain size 1.2–1.4 µ² surface area.

That is a concentration in the gelatin, not per mole of silver, so it has to be converted before it means anything to a maker. Duffin supplies the conversion factor in the next sentence: assume 200 g of gelatin per gram mole of silver.

🧮From parts per million in gelatin to milligrams per mole of silver

At the bottom of Bekunov’s range, 25 ppm of the gelatin:

25 × 10⁻⁶ × 200 g = 0.005 g = 5 mg of thiosulfate per mole of silver.

At the top, 150 ppm:

150 × 10⁻⁶ × 200 g = 0.030 g = 30 mg per mole of silver.

So the published range is 5 to 30 mg of thiosulfate per mole of silver. Now scale it to a domestic make. The plain-silver iodobromide recipe in this part carries 5 g of silver nitrate, which is 5 ÷ 169.87 = 0.0294 mol, so the whole of Bekunov’s range is:

5 mg/mol × 0.0294 mol = 0.15 mg, up to 30 × 0.0294 = 0.88 mg.

A tenth of a milligram to nine tenths of a milligram, for the entire batch. That is far below what a domestic balance weighs, so it can only be delivered from a dilute stock — the percentage-solution method from the precipitation lesson, taken one or two dilutions further. Designing that stock and its drop count is the course’s own work, not something it found published, and any page that uses it must say so.

One figure this course cannot reproduce. Duffin adds that Bekunov’s numbers mean “about 2 × 10¹⁹ sulphur atoms per gram mole of silver will be available for sensitization”. Working back from 5 to 30 mg does not land on 2 × 10¹⁹ for any single obvious choice of thiosulfate salt and of how many of its two sulfur atoms count as available, and the passage does not say which he assumed. The course therefore quotes his atom count as his and publishes only the milligram range as arithmetic it has checked.

The three things the curves say

Speed against digestion time at four thiosulfate doses, after Duffin's Figure 5.8

• 25 ppm in the gelatin
• 40 ppm
• 80 ppm
• 150 ppm

Show the numbers behind this plot

Four curves of relative log speed against digestion time in hours, one for each of four quantities of thiosulfate in the gelatin: 25, 40, 80 and 150 parts per million. Every curve rises to a maximum and then falls. The largest dose, 150 parts per million, rises first and peaks earliest and lowest, at about 0.42 after roughly one hour. The 80 parts per million curve peaks at about 0.52 after about one and three quarter hours. The 40 parts per million curve peaks at about 0.57 after about two hours. The smallest dose, 25 parts per million, peaks last and highest, at about 0.68 after nearly four hours, and is still above 0.6 at five hours. So a smaller quantity of sensitiser takes longer to reach its maximum and reaches a higher one, and all four maxima lie between about 0.4 and 0.7 log units, which is between one and a third and two and a third stops.

Series	Digestion time (hours)	Relative log speed
25 ppm in the gelatin	0.50	0.22
25 ppm in the gelatin	1.50	0.38
25 ppm in the gelatin	2.50	0.55
25 ppm in the gelatin	3.70	0.68
25 ppm in the gelatin	5.00	0.63
40 ppm	0.90	0.26
40 ppm	1.40	0.44
40 ppm	2.00	0.57
40 ppm	2.50	0.50
80 ppm	0.80	0.30
80 ppm	1.20	0.44
80 ppm	1.70	0.52
80 ppm	2.10	0.47
150 ppm	0.50	0.20
150 ppm	0.80	0.36
150 ppm	1.05	0.42
150 ppm	1.40	0.30

Published data: the axes, the four doses and the shapes are Duffin's Figure 5.8, from Bekunov; the individual points were read off the printed curves by this course, so the values are approximate and the ordering of the four maxima is the teaching.

Three findings come out of that figure and the text around it, and every one of them is counter-intuitive.

More sensitiser gives less speed. Duffin: “it is clear from many publications that quantities far larger than these produce no more speed and some of Bekunov’s data, given in Figure 5.8, show that larger quantities produce lower maximum speed.” A maker who reasons that if a little is good, more will be better, gets a slower emulsion.

The optimum is a time, and it moves. The smaller the dose, the longer to the peak and the higher the peak. So the dose and the digestion time are not two independent settings; they are one setting expressed two ways.

Almost none of the sensitiser has reacted at the peak. “Recent work has shown that, under widely differing circumstances, only a fraction, usually around 10%, of the sulphur sensitizer has been broken down by the time that the speed has reached its maximum.” And, strangely, adding only that fraction does not work: a correspondingly smaller quantity does not give as high a sensitivity after a long digestion, and fog goes up. Duffin reports the suggestion of Sutherns and Loening that what is also needed is sufficient surface coverage of adsorbed unreacted sensitiser. The stuff that has not reacted is doing something.

Duffin is also explicit that a lower limit has not been established, and that very small quantities probably bring on fog before useful sensitivity is reached, since the rate of digestion is controlled by the restrainer content as well as by the sensitiser.

🔬How little sulfide, in atoms

Duffin does the count and it is worth carrying, because it is the number that makes sensitisation believable. In an emulsion with 4 × 10¹⁴ grains per gram mole, if all the sulfur sensitiser were converted to silver sulfide there would be about 5 × 10⁴ sulfur atoms per grain. One crystal carries about 3 × 10⁶ silver ions on its surface, so only about 3 per cent of the surface silver would be associated with sulfur — and the sulfide tends to form in a small number of groups at particular places rather than spreading evenly.

Now remember that only about a tenth of the sensitiser actually breaks down by the time speed peaks, and that the latent image needs a handful of silver atoms at one of those places. A few specks, a few thousand atoms each, on a crystal of a billion ions, multiplying the speed of the material by a factor of five. That is chemical sensitisation.

⁇What the sensitivity speck actually is

Duffin sets out two positions and does not resolve them, and neither does this course.

Silver sulfide. James and Vanselow proposed that the speck is silver sulfide in a layer form — islands on the grain surface — and that silver sulfide alone explains the properties the speck shows.

Silver. Chibisov and colleagues held that the sensitivity speck is composed only of silver, that silver sulfide formation is a step towards it or a parallel and possibly unnecessary reaction, and that silver sulfide actually has an adverse effect. Mitchell’s evidence pointed the same way: stripping the gelatin film from reduction-sensitised silver bromide crystals removed the bulk of the surface sensitivity, which suggests a small group of silver atoms attached to but slightly separate from the grain.

Duffin’s own summing-up leans to a combination — that the sensitivity and fog specks “would not be pure silver sulphide or pure silver but would probably comprise both substances” — and he notes that gold’s behaviour is strong evidence against the pure-silver theory. A third position, from Sauvenier and Hautot, puts the sensitivity speck at the adsorbed silver–thiosulfate complex itself, with the thiosulfate’s solvent action etching the surface into the right condition.

Sheppard, who found the effect, wrote in his own 1924 patent only that the grains’ sensitiveness “has been shown to correspond to the presence in said grains of nuclei of unstated chemical composition”. Forty-two years later Duffin was still hedging: if silver sulfide formation must be accompanied by an even smaller quantity of something else, “then this occurs in a way which, at the moment, we cannot detect.”

Gold, and sulfur-plus-gold

Gold sensitisation was found by the Agfa workers in the 1930s and patented in the United States by Waller in 1946. Duffin: the gold compound is used in amounts comparable to the sulfur sensitiser and is one of two types, potassium aurothiocyanate or potassium aurochloride, the aurothiocyanate being the original and formed by treating gold chloride with an equivalent quantity of an alkali thiocyanate.

Three things are worth knowing before any number is quoted.

It needs the sulfur. Waller’s patent states that all normal gelatin contains small proportions of sulfur compounds such as thiosinamine or allyl isothiocyanate, that their presence is believed essential to the gold effect, and that gelatins known as inert benefit from a positive addition of one. Gold is not an alternative to sulfur sensitisation; it is a treatment applied on top of it, which is why the practice is called sulfur-plus-gold.

The doses. Duffin: “from 1 to 30 mg. of gold, either as chloroauric acid or a more complex gold salt, being used per gram mole of silver. If a gold sensitizer is added at the completion of digestion, then larger quantities are needed.” His Figure 5.9, for a sodium aurothiocyanate iodobromide emulsion, shows an optimum near 3 × 10⁻⁶ moles of gold and a maximum increase over sulfur sensitisation alone of about 0.33 log units, which is one stop. Waller adds the direction from the other side: too much gold loses the speed gain, and the quantity used is far too small to alter the tone of the developed image.

The preparation, and the tested dose. The practitioner recipe in the corpus uses six or seven drops of a Steigmann aurous ammonium thiocyanate solution, prepared by adding 6.0 ml of a 1 per cent gold chloride solution to 50 ml of a 1 per cent ammonium thiocyanate solution and letting it clear, stored in a dark dropper bottle. That preparation is quoted by her from the 1973 SPSE handbook, which this course has not read.

🧮Does the practitioner's dose sit inside Duffin's published range?

It does, and the arithmetic is worth showing because it is the strongest cross-tier agreement in this part — and because it depends on two assumptions the course has to declare rather than bury.

6 ml of a 1 per cent gold chloride solution is 60 mg of the salt, made up in about 56 ml of stock, so about 1.07 mg of gold chloride per ml. Six to seven drops, assuming 20 drops to the millilitre, is 0.30 to 0.35 ml, so 0.32 to 0.37 mg of gold chloride. Assuming the salt is chloroauric acid trihydrate, HAuCl₄·3H₂O at 393.8 g/mol against gold’s 197.0, the salt is 50 per cent gold by mass, so 0.16 to 0.19 mg of gold. Her batch carries 10.8 g of silver nitrate, which is 0.0636 mol, so the dose is about 2.5 to 3.0 mg of gold per mole of silver.

And it is classified Danger. PubChem’s aggregation of the ECHA C&L Inventory gives chloroauric acid signal word Danger with GHS05, GHS07, GHS08 and GHS09 — corrosive, harmful, a health hazard and an environmental one. That is why the gold step is designed on paper in the fifth project rather than run here: the evidence for what gold does is unusually good, and the evidence for handling it at this level is not the obstacle either. The obstacle is that the published dose lives in an ammoniacal make and Part V keeps ammonia out.

Duffin’s published range is 1 to 30 mg per mole. Her figure lands in the lower part of it, near the optimum his Figure 5.9 shows. A Tier 1 range from an industrial chemist and a Tier 2 dose from a working practitioner, agreeing across half a century and two continents, is about as good as the evidence in this subject gets.

But the course still does not perform it, and the reason is not the arithmetic. The only tested gold dose in the corpus sits inside an ammoniacal make, which Part V excludes at Level B, and transplanting a step out of one formula into another is exactly the invention Rule 6 forbids. Gold is taught here, and designed with in the fifth project, and not done.

Two further routes are named and not used. Reduction sensitisation produces silver specks rather than sulfide ones, is prone to fog, and — Duffin notes — can confer considerable sensitivity without fog, which shows that the silver of the sensitivity speck must differ in some way from that of the latent image or it would catalyse development on its own. And sensitisation by other metals: Duffin records references to cadmium salts and to mercury, iron, palladium and platinum in the patent literature, with palladium raised at a 1960 London conference on maximum sensitivity. Cadmium and mercury are excluded from this course at any level, and are taught only as history in Part XXVI.

Spectral sensitisation in practice

Part IV’s spectral sensitivity owns the mechanism — how an adsorbed dye absorbs a photon the crystal cannot and passes the energy or the electron into it. This page owns the doses, the addition point and the consequence.

Where the dye goes in, and how much

Three published methods, and they are three different operations that the page keeps apart.

Source	Method	Dose as published
Baker 1941, page 108	into the silver solution, just before mixing	25 cc of 2 per cent erythrosin in equal parts alcohol and water, for 500 g of silver nitrate
Ross, AmBr with Variations	into the salted gelatin, before precipitation	3 to 4 drops of a 2 per cent solution, for 5 g of silver nitrate
Wall 1924, pages 24–25	bathing the finished plate	a 1:5000 aqueous solution with 0.5 per cent ammonia

Reduced to a common basis, the first two agree closely. Baker’s 25 cc of 2 per cent is 0.5 g of dye for 500 g of silver nitrate, which is 2.943 mol, so about 170 mg per mole of silver. Ross’s 3 to 4 drops of 2 per cent, at 20 drops to the millilitre, is 3.0 to 4.0 mg for 0.0294 mol, so 102 to 136 mg per mole of silver. A Tier 1 manufacturing formula and a Tier 2 domestic recipe, seventy-two years apart, within a factor of 1.7 of each other. Both specify the same solvent: half water, half alcohol.

Wall’s figure is not comparable and must not be set beside them: 1:5000 is a bath concentration, 200 mg per litre, not a quantity per mole. His bath also contains ammonia, which Part V excludes, and he does not restate it in borax, so the course reports it and does not adapt it.

Wall does give the Tier 1 statement of the trade-off between the two methods, which is worth having: plates orthochromatised by bathing “as a rule … have a higher colour sensitivity than those coated with an emulsion to which the dye is added during the mixing, but they do not keep quite as well”. He adds two practical constraints — only perfectly clean glass dishes, because metals tend to reduce the dyes and cause fog, and that home-bathed plates should be used soon.

!There is an optimum, and no published figure is it

Duffin’s page 125 is the sentence that turns the dye from a constant into a variable: sensitivity rises with the quantity of dye adsorbed to a maximum, and “it will be observed that a rapid loss of useful speed occurs if too much dye is added. The exact position of the maximum point X is a function of the particular dye and the nature of the emulsion.” He adds that for many commercial purposes the balance of other properties makes a smaller quantity than the speed maximum desirable.

So the two agreeing doses above are where two makers landed for their emulsions, not the optimum for yours. Finding your own is a measurement, and it is what makes the fifth project a design exercise rather than a recipe.

Erythrosin: what to say about it, and what not to

Erythrosin is the disodium salt of tetraiodofluorescein, and it is also the food colour E127 and FD&C Red No. 3. The course’s ruling permits it at Level B in the fifth project only, and requires the page to be exact about the evidence.

Its GHS status is “unclassified on a thin evidence base”, which is not the same thing as “harmless” and the course does not let one stand in for the other. The ECHA notifications aggregated under EC 240-046-0 report that 37 of 39 companies say it does not meet GHS hazard criteria — but PubChem’s own note records that only about 5.1 per cent of companies provided GHS information at all, and two of the thirty-nine did lodge hazard statement codes that are not itemised at that heading. There is no British workplace exposure limit, and EH40 states that absence from its list does not indicate that a substance is safe; under COSHH that means exposure is reduced as far as reasonably practicable with no figure to work to. Dust discipline therefore applies regardless of the absent classification.

The FDA revoked its food authorisation in January 2025, and the reason matters. The revocation was compelled “as a matter of law, based on the Delaney Clause”, after two studies found cancer in laboratory male rats. In the same notice the agency states that “the way that FD&C Red No. 3 causes cancer in male rats does not occur in humans”, that relevant human exposures are typically much lower, and that claims of human risk are not supported by the available scientific information. Both halves belong on the page: a reader who meets the headline elsewhere and not here would be right to distrust the course, and a reader who meets it without the mechanism would draw the wrong conclusion. Note also what the notice is about — eating the dye — and what you are doing, which is dissolving about a gram of it once and thereafter dispensing drops.

The practical controls are the ones any unclassified fine powder gets: buy the smallest quantity, make the whole of it up into one 2 per cent stock in a single operation so the powder is weighed once or never, work over a tray in still air with no fan running, nitrile gloves and eye protection while the powder is open, a disposable FFP2 or FFP3 mask as a precaution against an unquantified dust rather than as compliance with a limit, and a labelled stock — substance, CAS 16423-68-0, strength, solvent, date, and the fact that it is half ethanol and therefore flammable. Treat staining as the visible sign that a control has failed: erythrosin marks skin, clothing and worktops, and unlike most hazards it tells you exactly where the powder went.

The consequence that is the real point

An orthochromatic emulsion cannot be handled under the safelight the earlier projects use. That is not a nuisance, it is the lesson.

Three manufacturer sources agree on what it needs, and none of them says darkness. ILFORD’s filter table gives 906, dark red, for orthochromatic materials, and 915 light red for orthochromatic graphic-arts materials, against 904 dark brown for fast blue-sensitive materials. ILFORD’s current ORTHO Plus sheet states that blue and green sensitivity “enables the film to be handled in deep red safelight”, with the 906 filter, a 15 W bulb and not less than 1.2 m. Kodak’s darkroom illumination guide says blue-sensitive and orthochromatic black-and-white films can be handled under red safelights.

So the cost of the dye is a filter change and a fresh safelight fog test — both of which this course already owns, from Station 4 of the darkroom session in Part IV and the safelight work in Part XVI. Change the dye, and the light you may work under changes. Nothing else in the course lets a student feel spectral sensitivity in their own room.

Panchromatic sensitisation is a different ruling and the answer is no. Pinacyanol chloride and its relatives push sensitivity to about 720 nm and require total darkness or a very dim red headlamp; the corpus holds no hazard record for any of them; and working in complete darkness with hot gelatin, a 55 °C bath and silver nitrate is both a control failure waiting to happen and an accessibility barrier. It is taught, from Wall’s account — the class ceiling of not more than 1 in 75,000 in the bath, the named dyes, the observation that the violet dyes give the best all-round results — and it is not performed.

Finals, and the ones the course cannot give you

“Finals” is the trade word for everything added after digestion and before coating. Four categories, and the course can source three of them.

Restrainer. A small quantity of soluble bromide, added to hold back fog. Two independent doses, and they agree in order of magnitude. Baker’s page 108 finals carry 50 cc of a 1 per cent ammonium bromide solution — 0.5 g — in ten litres carrying 500 g of silver nitrate, which is about 170 mg per mole of silver. Ross adds 2 drops of 10 per cent potassium bromide, about 10 mg, to a batch of 0.0294 mol, which is about 340 mg per mole. Same order, from a 1941 manufacturing formula and a 2013 domestic one; she calls hers “added insurance against excessive base fog”.

Organic stabilisers: named, and not dosed. The manifest asks for the identity and dose of 4-hydroxy-6-methyl-1,3,3a,7-tetraazaindene and of benzotriazole. No source at any tier in this course’s corpus gives a dose for either in an emulsion. Duffin’s “Restrainers” section is about substances naturally present in gelatin, not about added stabilisers: he reports Gordon and Swann’s finding that hide gelatins contain on average 30 to 50 ppm, sometimes 200, of adenine — a heterocycle structurally related to many stabilisers — while ossein gelatins carry only 4 to 6 ppm. That is excellent mechanism and it is not a dose. So this course substitutes the bromide restrainer, which has two published doses, and teaches the organic stabilisers as chemistry the student reads about. The reason is the absence of a published quantity, not a judgement that they are unsuitable, and those are different sentences.

Bacteriocide. Duffin’s list is phenol, thymol and p-chloro-m-cresol, with the halogenophenols effective at 0.1 per cent of the weight of the gelatin and the others needing up to 5 per cent. Baker’s Trumm bromide paper carries 10 g of phenol per 16 kg, dissolved in part of the spirit. This course omits the phenol and says so, because it is not a domestic reagent and Rule 5 keeps it out of a Level B procedure. The cost is stated rather than absorbed: the emulsion’s resistance to bacteria and mould over weeks of storage, which is replaced by small batches, refrigeration below 5 °C and using the material rather than keeping it.

Wetting agent, alcohol and hardener. Wall instructs that with all paper emulsions saponin or quillaia tincture should be added; Baker notes that an alcoholic solution of saponin or extract of quillaia bark is frequently added to give better coatings. The modern equivalent is a commercial photographic wetting agent, and the dose for it is one of the gaps this part carries openly: the free tutorial that supplies Project 1’s formula names the wetting agent in its tools list and gives no quantity anywhere. Alcohol appears in almost every make at several per cent, for coating and setting. The hardener decision is made once, on the coating page, and the default is none.

Order and timing matter, and Kodak’s mixing note for the acid hardening fixers is the general lesson in a specific case: dissolve the ingredients in the proper order or the hypo decomposes and the alum precipitates; add the hardener stock to the hypo solution slowly with vigorous stirring, with both solutions cold. The same discipline governs finals. Duffin’s chrome alum failure mode is local over-concentration causing coagulation — the emulsion has to be moving, and the addition has to be slow.

Reading a published formula as a sequence of decisions

A historical emulsion formula looks like a shopping list and is not one. Every line is a decision someone made, and once you can name the decision you can change it deliberately instead of by accident. Take Baker’s orthochromatic formula, and read it as choices rather than quantities.

• Ammonium bromide 350 g, ammonium iodide 8 g — the halide and its excess, which sets ripening rate and habit; the cation is ammonium, which affects solubility and the pH left behind. The iodide is about 2 per cent of the halide: enough for speed and clean highlights, well below the 10 per cent above which sensitivity falls.
• Soft gelatin 150 g in the make, 650 g added afterwards — a low gelatin during precipitation, so growth is not over-restrained, and a large addition afterwards to bring the coating solution up to a workable 5 per cent or more before the wash dilutes it.
• Silver at 21 °C into salts at 32 °C, “flopped” in without stirring — an addition-rate decision, at the extreme fast end of Duffin’s table, and therefore a decision about grain count.
• Erythrosin into the silver just before mixing — a spectral decision, and an addition-point decision: in the make, not a bath.
• Set quickly in ice water, thoroughly washed after about five hours — a decision to stop ripening, and a decision that this emulsion goes on glass.
• Chrome alum, phenol, ammonium bromide, made to ten litres — hardener, bacteriocide, restrainer, and the coating weight, which is what “made up to ten litres” actually sets.

Read that way, a formula is a set of nine or ten answers, and your own make is a set of answers to the same questions. That is what the batch record is recording, and it is why a formula you have understood can be varied one line at a time and a formula you have merely followed cannot.

Recap

• Washing removes nitrate, excess halide and pH, and the second of those is the one that matters most: digestion is retarded by excess halide, so an unwashed emulsion cannot be properly sensitised.
• Paper absorbs the by-products and glass does not, which is why gaslight papers were unwashed and plate emulsions never were.
• Noodle washing is set, shred, wash in many changes, test, drain, hold cold. Six changes of three minutes domestically; about twelve of five minutes commercially.
• It dilutes as it washes — concentration can halve — so the second gelatin goes in before setting, and a coating solution needs about 5 per cent gelatin.
• The industrial endpoints are conductivity and pAg, with pAg preferred; the only published domestic endpoint is one practitioner’s pH curve, from 10 down to between 7 and 8.
• Coagulation washing replaced it industrially, by solvent, salt, sulfonic acid or acylated gelatin, and always needs efficient redispersion. Ultrafiltration and dialysis are named in the manifest and described by no source this course has read.
• Digestion is 40 to 70 °C for minutes to hours, and raises speed and fog together; pull where fog starts to move, not at the speed peak.
• Bekunov’s range is 25 to 150 ppm of thiosulfate in the gelatin, which at 200 g of gelatin per mole is 5 to 30 mg per mole of silver, and 0.15 to 0.88 mg for a 5 g silver nitrate batch.
• Larger doses give a lower maximum speed and reach it sooner, and only about 10 per cent of the sensitiser has decomposed at peak speed.
• Gold is 1 to 30 mg per mole, worth about 0.33 log units over sulfur alone, needs the sulfur to work, and is not performed in this course because the only tested dose is inside an ammoniacal make.
• Erythrosin is 100 to 170 mg per mole across two agreeing sources, has no optimum you can copy, and its real cost is a change of safelight to deep red.
• The stabiliser dose does not exist in this corpus. The bromide restrainer, which has two published doses, takes its place, and the page says that is an absence of evidence rather than a judgement.

Check your understanding

Question 1. A maker coats an unwashed iodobromide emulsion on glass and on watercolour paper from the same batch. The paper prints acceptably; the plates dry with a crystalline bloom and are unusable. What happened, and what second consequence would the unwashed emulsion have had even on paper?

• The plates were not clean; the emulsion would have been fine on both supports
• The soluble nitrate by-product crystallised at the drying stage on the non-absorbent glass while the paper absorbed it; separately, the excess halide left in the emulsion would have retarded any chemical sensitisation
• The glass was not subbed, so the emulsion frilled; there is no second consequence
• The emulsion was over-ripened, which only shows on glass

Show the answer and why

Answer: The soluble nitrate by-product crystallised at the drying stage on the non-absorbent glass while the paper absorbed it; separately, the excess halide left in the emulsion would have retarded any chemical sensitisation

Duffin gives the first half directly: the salts still present crystallise at the drying stage and render the coated layer unsuitable, or prevent complete drying altogether. Paper absorbs them, which is exactly why the historical division exists — bromide papers were always washed emulsions and gaslight papers usually unwashed. The second consequence is the one that decides the whole sequence of a make: digestion is retarded by excess halide, so the bromide that was your ripening agent becomes a restrainer that blocks sensitisation. You cannot sensitise an emulsion you have not washed.

Question 2. Duffin quotes Bekunov's range as 25 to 150 parts per million of thiosulfate in the gelatin. Convert that to a dose for a batch carrying 5 g of silver nitrate, and say what the answer forces the course to do.

• 5 to 30 mg for the batch, which a 0.01 g balance can weigh directly
• About 0.15 to 0.88 mg for the batch, which no domestic balance can weigh, so it must be delivered from a dilute stock the course designs and labels as its own method
• 25 to 150 mg for the batch, weighed directly
• The conversion cannot be made because the range is given per gram of gelatin

Show the answer and why

Answer: About 0.15 to 0.88 mg for the batch, which no domestic balance can weigh, so it must be delivered from a dilute stock the course designs and labels as its own method

Duffin supplies the conversion factor himself: assume 200 g of gelatin per gram mole of silver. So 25 ppm is 25 × 10⁻⁶ × 200 g = 5 mg per mole and 150 ppm is 30 mg per mole. A 5 g silver nitrate batch is 5 ÷ 169.87 = 0.0294 mol, so the whole published range is 0.15 to 0.88 mg. That is below any domestic balance, which is why the dose has to come from a serial dilution — and why this course is honest that designing that stock is its own work rather than something it found published. No formula in its corpus performs a measured thiosulfate digestion at domestic scale.

Question 3. You digest two identical batches, one with 25 ppm of thiosulfate in the gelatin and one with 150 ppm, and pull both at the same time. Which is likely to be faster, and what does Duffin's Figure 5.8 say about the general relationship?

• The 150 ppm batch, because more sensitiser always means more speed
• It depends entirely on when you pulled them: the 150 ppm batch peaks earlier and lower, the 25 ppm batch peaks later and higher, so larger doses give a lower maximum speed reached sooner
• They will be identical, because only the digestion time matters
• The 25 ppm batch, because sensitiser is a restrainer

Show the answer and why

Answer: It depends entirely on when you pulled them: the 150 ppm batch peaks earlier and lower, the 25 ppm batch peaks later and higher, so larger doses give a lower maximum speed reached sooner

The four curves in Figure 5.8 are the point of the figure: 150 ppm peaks around one hour at roughly 0.42 log units, while 25 ppm keeps climbing to nearly four hours and about 0.68. Duffin states the finding plainly — larger quantities produce lower maximum speed — and notes that quantities far larger than the range produce no more speed at all. Dose and digestion time are not independent settings; they are one setting expressed two ways, which is why a batch record must carry both. He also warns that if there is a lower limit it has not been established, and that very small quantities probably bring on fog before useful sensitivity is reached.

Question 4. Why does this course teach gold sensitisation with published quantities but not perform it, and what would performing it require the course to do?

• Gold is too expensive; performing it would require a larger budget
• The only tested gold dose in the corpus sits inside an ammoniacal make that Part V excludes at Level B; performing it would mean transplanting a step from one formula into another, which is inventing a procedure
• Gold sensitisation does not work at domestic scale
• Chloroauric acid has no GHS classification, so no hazard assessment is possible

Show the answer and why

Answer: The only tested gold dose in the corpus sits inside an ammoniacal make that Part V excludes at Level B; performing it would mean transplanting a step from one formula into another, which is inventing a procedure

The evidential position on gold is unusually good: Duffin gives 1 to 30 mg per gram mole of silver as a published range, his Figure 5.9 shows an optimum and a gain of about 0.33 log units — one stop — over sulfur alone, and a named practitioner's dose works out at about 2.5 to 3.0 mg per mole, comfortably inside it. What is missing is a tested plain-silver procedure: her gold step lives in an ammoniacal make, and Part V keeps ammonia out. Moving a step between formulas is exactly the kind of assembly Rule 6 forbids, so the gold goes into the design half of the fifth project. Chloroauric acid does carry a classification, and a serious one.

Question 5. A student adds erythrosin to their emulsion for the first time and asks what changes about the rest of their workflow. What is the honest answer?

• Nothing; the dye only affects the negative's response to green light
• The material must move to a deep red safelight — the manufacturers agree on a dark red filter rather than total darkness — and the safelight fog test must be re-run for the dyed material
• They must now work in total darkness
• They must switch to a hardening fixer

Show the answer and why

Answer: The material must move to a deep red safelight — the manufacturers agree on a dark red filter rather than total darkness — and the safelight fog test must be re-run for the dyed material

Three manufacturer sources agree and none of them says darkness: ILFORD's filter table gives 906 dark red for orthochromatic materials, its current ORTHO Plus sheet specifies deep red safelight with that filter, a 15 W bulb and not less than 1.2 m, and Kodak's guide says orthochromatic films can be handled under red safelights. So the cost is a filter change plus a fresh fog test, both of which the course already owns. Total darkness belongs to panchromatic sensitisation, which this course teaches and does not perform — the corpus holds no hazard record for pinacyanol chloride, and working in the dark with a 55 °C bath and silver nitrate is a control failure waiting to happen.

Question 6. Why does this course use a soluble bromide as its restrainer rather than benzotriazole or a tetraazaindene, and how should the page phrase that choice?

• Because the organic stabilisers are unsuitable for handmade emulsions; the page should say they do not work
• Because no source at any tier in the corpus gives a dose for either compound in an emulsion, while the bromide restrainer has two published doses that agree in order of magnitude; the page must say the reason is an absence of published evidence, not a judgement about the compounds
• Because the organic stabilisers are too hazardous to handle
• Because bromide is cheaper

Show the answer and why

Answer: Because no source at any tier in the corpus gives a dose for either compound in an emulsion, while the bromide restrainer has two published doses that agree in order of magnitude; the page must say the reason is an absence of published evidence, not a judgement about the compounds

The two doses that do exist are Baker's 1 per cent ammonium bromide in the finals, about 170 mg per mole of silver, and the practitioner's two drops of 10 per cent potassium bromide, about 340 mg per mole — the same order, seventy-two years apart. For benzotriazole and the tetraazaindenes there is nothing: Duffin's Restrainers section is about adenine and guanine naturally present in gelatin, which is mechanism and not a dose. Rule 7 makes the phrasing matter, because "we could not source this" and "we chose not to do this" are different sentences and the reader is entitled to know which applies. Here it is the first.

Check my answersClear