Project 5: Designing an Emulsion, and Making the One We Can Source

Purpose

To design an emulsion you cannot make, and then make one you can.

Four projects have handed you formulas. This one asks you to write one: to state a goal, choose which levers to move, predict what each will do, and only then weigh anything. The design half is where the advanced chemistry lives — double-jet addition, sulfur-plus-gold digestion, an organic stabiliser — and it stays on paper, for reasons the page gives in the open rather than leaving you to wonder.

The make half is a plain-silver orthochromatic emulsion: the same silver charge and almost the same halide as Project 3, with three or four drops of a red dye added to the salted gelatin before precipitation. Those drops are the most visible change in the whole part, and not because of what they do to the negative. They change the light you are allowed to work under.

!This project was narrowed, and the reason is evidence rather than caution

As originally designed it was a double-jet, sulfur-plus-gold, dye-sensitised, stabilised iodobromide make. Nobody has published one at a scale and hazard level a home laboratory can run, and the course narrowed the project rather than assembling the missing parts from four different sources.

Specifically: no source at any tier gives a home double-jet procedure; no published home formula performs a measured thiosulfate digestion; the only tested gold dose sits inside an ammoniacal make this part excludes; and no source at any tier gives a dose for a tetraazaindene or for benzotriazole in an emulsion. Rule 6 does not permit a formula to be built out of pieces that were never published together, so those four features are designed and not performed, and the bromide restrainer, which does have two published doses, takes the stabiliser’s place.

Learning objectives

By the end of the two sessions you should be able to:

• write an emulsion specification: goal, levers, predicted effect of each, and a version code, before any weighing;
• describe double-jet precipitation, say what controlling pAg buys in size distribution, and state what a constant-rate single jet cannot control;
• quote the published ranges for sulfur and for gold with their source, convert them to a dose for your own batch, and explain why the page prints ranges rather than a recipe;
• make up a 2 per cent dye stock in one operation and dose it in drops;
• state what is and is not known about erythrosin’s hazard classification, and say why the course does not reach for the word “harmless” and does not stop at the word “banned”;
• design and run a spectral test with coloured gels that measures your own material’s response;
• complete a five-emulsion comparison matrix from your own five batch records.

Prerequisites

Project 4, and both theory pages this project performs: precipitation, nucleation and growth for the jets and pAg, and washing, digestion and sensitisation for the sensitiser quantities and the dye. Part IV’s spectral sensitivity owns the mechanism by which an adsorbed dye passes energy into a crystal, and this page does not re-derive it.

You also need four batch records — Projects 1 to 4 — or the matrix at the end has nothing to fill.

Safety classification

Level B, on the same basis as Projects 1 to 4, with one addition ruled on specifically.

The dye is permitted at Level B, in this project only. The conditions are the course owner’s and are not negotiable at the bench: bought in the smallest available quantity, made up as a single 2 per cent stock in one operation so that the powder is weighed once or never, and thereafter dosed in drops.

What is not a hazard here, and why. There is no ammonia. Ammonium bromide is a salt and this is a plain-silver make: nothing here forms the silver ammine complex that an ammoniacal emulsion depends on, and no ammonia is evolved at the pH and temperature of the make. One incompatibility follows from that and is worth stating: ammonium salts and strong alkali together give ammonia, so caustic soda does not come near this bench. There is no gold in the make — the gold is designed and not weighed — no cadmium, no mercury, no chromium in any oxidation state, and no acid. The chemical exposure that is genuinely new is about one gram of a red powder, once, and the control for it is dust discipline.

Level D, named and not performed. Panchromatic sensitisation. The course teaches it and gives no procedure, for two reasons: the corpus holds no hazard record for pinacyanol chloride or any of its relatives, and the work demands total darkness or a green safelight, which is a control failure waiting to happen with hot gelatin and silver nitrate on the bench, and an accessibility barrier as well.

Hazards

Silver nitrate, as before: Danger, GHS03, GHS05, GHS09 — H272, H314, H400, H410. The controls are Part II’s.

Erythrosin powder, for the one minute the jar is open. The honest description of what is known is below, under the dye. The controls do not depend on resolving it: weigh over a tray in still air with no fan or extractor draught running, nitrile gloves and eye protection while the powder is open, and a disposable FFP2 or FFP3 mask as a precaution against an unquantified dust rather than compliance with a limit, because no limit exists.

The dye stock is about half ethanol by volume. A 50 mL bottle of roughly 50 per cent ethanol on a shelf near a water bath is a flammable liquid, labelled and stored as one.

Staining, treated as a signal rather than a nuisance. Erythrosin marks skin, clothing, worktops and trays. Unlike most hazards it tells you exactly where the powder went, so a stain is evidence that a control failed and is worth acting on rather than scrubbing off and forgetting.

A deep red safelight is dimmer than the one you have been using. More sessions go wrong from knocking something over in the dark than from the chemistry. Lay the bench out in the light first.

Required PPE

Splash goggles, nitrile gloves, apron and closed shoes throughout, with eyewash and spill provision within reach. Add the disposable mask for the single dye weighing, and change gloves the moment they are stained.

Ventilation

Ordinary room ventilation, and again the control that matters is still air rather than moving air. Nothing here evolves a vapour: the make runs at 55 °C, the only volatile ingredient is ethanol added in two 5 mL portions at the end, and the dye is a solid dispersed in a solvent. The dye weighing happens with the extractor off, because a draught across an open jar of fine powder is the mechanism you are trying to prevent.

Materials

Everything Project 3 used — the ricer, the flask, the bag, the ice — plus:

• a deep red safelight filter of the class the manufacturers specify for orthochromatic material, and a 15 W bulb;
• an amber dropper bottle of about 60 mL for the dye stock, and a beaker about twice the final volume;
• coloured gels: a yellow, a green and a red, for the spectral test. Lighting gel swatch books are cheap and carry all three;
• a tray to weigh over, and a plastic spoon that will be stained red for ever.

Chemicals

Chemical	Quantity	Form
Silver nitrate	5 g	crystals, weighed dry
Ammonium bromide	4.1 g	ground fine
Potassium iodide	1 mL of 10 % w/v	0.1 g of the salt
Gelatin	2 g, then 5 g	photographic, inert
Erythrosin, disodium salt	3 to 4 drops of a 2 % w/v stock	about 3 to 4 mg; buy on CAS 16423-68-0
Ethanol	25 mL of stock solvent, then 5 mL + 5 mL	drinking-grade, 95 %
Potassium bromide	2 drops of 10 % w/v, optional	the restrainer, about 10 mg
Distilled water	25 mL of stock solvent, plus 45 + 25 + 45 mL

🧮The make, in numbers, and why it is not a one-variable comparison with Project 3

Silver. 5 g ÷ 169.87 = 0.0294 mol, identical to Projects 1, 2 and 3.

Halide. 4.1 g of ammonium bromide ÷ 97.94 = 0.0419 mol; the iodide is 0.1 g ÷ 166.00 = 0.00060 mol. Total 0.0425 mol, so the excess is 44 mol per cent against Project 3’s 45, and the iodide is 2.0 mol per cent of the silver in both. Those two were matched deliberately: the halide is not the variable here.

But four other things are. The cation is ammonium rather than potassium; the wash is six changes of five minutes rather than three, thirty minutes rather than twenty; the ripening is sixty minutes rather than forty-five; and there is a second 5 mL of ethanol at 34 to 35 °C that Project 3 has not got. Add the dye and that is five differences. So the comparison against Project 3 tells you what these two emulsions do, and only the dye’s effect is isolated — because the spectral test compares the same sheet through different filters, which is a genuine one-variable measurement.

The dye. 3 to 4 drops of a 2 per cent solution, at 20 drops to the millilitre, is 0.15 to 0.20 mL, which is 3.0 to 4.0 mg. Per mole of silver that is 102 to 136 mg. Baker’s 1941 figure, from a different formula on a different continent seventy-two years earlier, is 25 cc of the same 2 per cent strength for 500 g of silver nitrate, which is 0.5 g for 2.943 mol, or 170 mg per mole. The two agree within a factor of 1.7, which for a dye dose is close corroboration.

The restrainer, if you use it. 2 drops of 10 per cent potassium bromide is about 10 mg, so 340 mg per mole. Baker’s finals carry 50 cc of a 1 per cent ammonium bromide solution in ten litres holding 500 g of silver nitrate, which is 170 mg per mole. Same order, from a 1941 manufacturing formula and a 2013 domestic one.

Equipment

Project 3’s, plus the deep red safelight and the gels. The balance is used once for the dye and needs to read to 0.01 g; everything else is a syringe or a dropper.

Estimated cost

Cost band £££. The silver is the same 5 g as every other make. The additions are a dye bought in the smallest quantity offered, a safelight filter, and a gel swatch. This page quotes no prices; dated figures live in the laboratory planner. Note one planner consequence of the design half: if you ever decide to perform the gold, chloroauric acid is sold in gram quantities at a price that makes the milligram you need an expensive milligram.

Estimated consumables cost

The same five grams of silver as every make in this part, plus the one substance that is unique to it: a few milligrams of erythrosin, bought in the smallest jar sold and effectively never exhausted at 3 to 4 mg a batch.

Consumed	This session	Sourced price	Cost this session
Silver nitrate	5 g	£59.95–£112.90 per one jar: 25 g at the lower figure, 10 g at the higher (£2.40–£11.29 a g)	£11.99–£56.45
Ammonium bromide	4.1 g	None. A named price gap: silver-halide salts other than potassium bromide	—
Potassium iodide	0.1 g, as 1 mL of 10 % w/v	None. A named price gap: silver-halide salts other than potassium bromide	—
Gelatin, photographic inert	7 g (2 g then 5 g)	£17.45 per 100 g, inert photographic gelatin (£0.17 a g)	£1.22
Erythrosin, disodium salt	3–4 mg, as 3–4 drops of a 2 % w/v stock	None. A named price gap: erythrosin, disodium salt	—
Ethanol, 95 %	35 mL, including the dye stock solvent	None. ethanol carries a cost band and no dated figure	—
Potassium bromide	about 10 mg, optional restrainer	£23.00 per 250 g (£0.09 a g)	£0.00
Hot-press watercolour paper	at least 4 sheets	None. A named price gap: hot-press cotton watercolour paper	—
Coloured gels: yellow, green, red	one swatch of each, for the spectral test	Swatch books are often free	—
Distilled water	about 140 mL	None. distilled-water carries a cost band and no dated figure	—

The priced rows come to £13.21 to £57.67 for one run of this session, at the retail ranges read on 5 September 2026 and recorded in the laboratory planner. That is a floor, not a total: 6 of the 10 rows carry no dated price, so they are counted as nothing here and are certainly not free. A priced entry is a dated range to plan against, never a quotation.

The deep red safelight filter, the amber dropper bottle and the dedicated spoon are equipment. The dye is the row a reader should look at twice: it is a named price gap, it is bought once, and the batch dose is milligrams.

Waste streams

Project 3’s, unchanged in kind: eight to twelve litres of wash water carrying nitrate, bromide and iodide with dissolved silver; make residues and failed coats; used fixer and first wash.

The dye adds milligrams to a stream that already exists. It leaves in the wash water and in the processing, in quantities of a few milligrams per batch, and it does not get its own disposal route: it joins the silver-bearing stream, with the same jurisdictional caveat.

Alternative route

Without the dye, run the whole project and make an undyed emulsion, then do the spectral test on it anyway. A colour-blind material through a green gel gives a real and instructive result: almost nothing. You lose the lever and keep the design exercise, the make and the matrix, and the page says plainly that the omission is then your decision rather than an absence of evidence.

Without silver nitrate, the design half stands alone and is the more valuable half. Write the specification, cost it, hazard-assess it, and defend it in the record.

Preparation

Session one, part one: the design, written before anything is weighed

This is the half of the project that carries the 300 minutes. Write it out, in this order.

1. The goal. One sentence, naming the property you are trying to move and the material it is for. “A plate two stops faster than BrI-01 at the same contrast” is a goal. “A better emulsion” is not.

2. The levers, and what each is worth. Five are available in the literature, and only some of them are available to you.

Lever	What the published record says it buys	Available to you?
Addition rate and jets	Grain count and size distribution; a double jet at controlled pAg gives regular cubes or octahedra	Design only — no home procedure published
Halide excess and ripening time	Mean grain size, and the width of the size distribution	Yes, and Project 3 measured it
Sulfur digestion	Speed, several-fold; fog rises with it	Design, with an experiment in Project 3
Gold, on top of sulfur	About 0.33 log units, one stop, over sulfur alone	Design only — the tested dose is inside an excluded make
Spectral dye	Green sensitivity, and a different safelight	Yes. This is the lever you actually move

3. The prediction. For every lever you move, one sentence saying which direction and by how much, and what would falsify it.

4. The version code, assigned now: Or-01a. A code assigned before the make is a commitment; one assigned afterwards is a label.

The double jet: what it is, and why you are not going to do it

The double-jet apparatus, drawn as a specification and not as a build

1. 1Silver jet — metered, below the surface, on one side of the impeller
2. 2Halide jet — metered to match, on the opposite side; its rate is what the controller varies
3. 3Silver electrode and reference — reads pAg continuously; a change in pAg alone decides cubes against octahedra
4. 4Jacketed, stirred vessel — gelatin solution at a held temperature, agitated fast enough to disperse both streams before either can supersaturate locally

Drawn as a specification. Nothing on this page asks you to build it: no source at any tier publishes a home double-jet make, and this drawing exists so that the design half has something concrete to argue about.

What the double jet buys. Both solutions arrive at once, so the ratio in the vessel does not change through the run — and if the halide rate is trimmed against a silver electrode, the silver ion activity stays where you put it. That matters because pAg decides crystal habit: the patent literature states, citing Berry and Skillman’s work on twinned silver bromide crystals, that a change in pAg alone lets a double-run precipitation give regular cubes or octahedra. A worked industrial example runs at 70 °C with pH held at 2.0 and pAg at 9.0 for thirty-five minutes and yields cubic-regular silver bromoiodide of about 0.2 µm.

What your single jet does instead. Silver goes into a large halide excess at a fixed 5 mL a minute, so the excess falls continuously through the addition and pAg drifts with it. The crystals grown at the start are not grown under the same conditions as those grown at the end, and the size distribution is correspondingly broad. Duffin’s own comment on a bromide-ripened emulsion is that the broad spread is characteristic of emulsions ripened in a large excess of bromide, and that the low contrast it gives suits negative work.

!Why the course does not ask you to perform it, in one paragraph

No source at any tier that this course could reach publishes a home double-jet make, and no source quantifies what error a constant-rate addition without a silver electrode introduces — which is precisely the question the project’s own must-verify list asks. Rule 6 does not allow the course to assemble a schedule from an industrial patent and a practitioner’s syringe note and present the result as practice.

One lead was identified and not followed. A page titled for the making of silver halide emulsions with a silver-ion-specific electrode was seen in the navigation of a site the course used and was not obtained. If it carries a tested domestic pAg-controlled procedure, this ruling changes. It is recorded here so that the gap is a known one rather than a silent one.

Sulfur and gold, as design

Both are published as ranges, and a range is what the page prints.

Sulfur. Bekunov’s figures, as Duffin gives them, are 25 to 150 parts per million of thiosulfate in the gelatin, at an assumed 200 g of gelatin per mole of silver — which is 5 to 30 mg per mole, or 0.15 to 0.88 mg for a batch carrying 5 g of silver nitrate. Project 3’s page designs the stock that delivers it, and labels that design as the course’s own.

Gold. Duffin gives 1 to 30 mg per gram mole of silver, as chloroauric acid or a more complex gold salt, with larger quantities needed if the gold goes in at the completion of digestion rather than during it. His Figure 5.9, for a sodium aurothiocyanate iodobromide emulsion, puts the optimum near 3 × 10⁻⁶ moles of gold and the maximum gain at about 0.33 log units over sulfur alone, which is one stop.

🧮An ambiguity in the published gold figures that a design has to resolve out loud

Duffin’s two statements do not obviously agree, and the corpus does not reconcile them.

Reading one: the range is milligrams of gold. Then his optimum of 3 × 10⁻⁶ mol, at gold’s 197.0 g/mol, is 3 × 10⁻⁶ × 197.0 = 0.59 mg of gold per mole of silver — just below the bottom of his own stated 1 to 30 mg range.

Reading two: the range is milligrams of the salt. Chloroauric acid trihydrate is 393.8 g/mol, so 3 × 10⁻⁶ mol of it is 1.18 mg — right at the bottom of the range.

Reading two makes the two figures consistent, which is a reason to prefer it and not a reason to assert it. Your specification must say which it means and why, because a factor of two on a milligram matters when you are making up a stock. This is the kind of thing a design exercise is for: the literature is not always self-consistent, and a designer who does not notice inherits the inconsistency.

Gold needs the sulfur. It is not an alternative treatment but one applied on top of a sulfur-sensitised emulsion, which is why the practice is called sulfur-plus-gold. And the fog test is what says stop: fog rises throughout digestion while speed rises to a maximum and falls, so the useful endpoint is the last moment before fog begins to move, not the speed peak.

Why the course does not perform the gold. The only tested dose in the corpus — six or seven drops of a Steigmann aurous ammonium thiocyanate solution, prepared from 6.0 mL of 1 per cent gold chloride in 50 mL of 1 per cent ammonium thiocyanate — sits inside an ammoniacal make, which this part excludes. Transplanting a step out of one formula and into another is invention, not adaptation. Note also the hazard: chloroauric acid is Danger, with GHS05, GHS07, GHS08 and GHS09.

The stabiliser that is named and never dosed

The specification asks for a stabiliser, and the two the manifest names are 4-hydroxy-6-methyl-1,3,3a,7-tetraazaindene and 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 — adenine at 30 to 50 ppm in hide gelatins against 4 to 6 in ossein — which is excellent mechanism and is not a dose.

So the emulsion you make uses the bromide restrainer, which has a published dose in two independent sources: about 170 mg of bromide per mole of silver from a 1941 manufacturing formula, about 340 from a 2013 domestic one. Worth noting in passing, because it would bear on the design even if a dose existed: benzotriazole’s notified classification is GHS09 with H411, toxic to aquatic life with long lasting effects, so a stabilised emulsion would add a constituent to the waste stream as well as to the finals. Name that distinction in your specification: the organic stabilisers are omitted because no quantity has been published, not because they are unsuitable, and those are different sentences.

Procedure

Session one — make and set

Make the dye stock first, in daylight, before the session. One operation, one weighing, and then never again. Stage the solvent — 25 mL of distilled water and 25 mL of drinking-grade ethanol — an amber dropper bottle, a beaker about twice the final volume, and a plastic spoon. Wear gloves and eye protection, put a tray under everything, and turn the extractor off. Buy the smallest quantity offered and make up the whole of it, rinsing the container into the beaker in stages so that no powder is left to be weighed a second time; for a 50 mL stock at 2 per cent that is 1 g. Label the bottle: substance, CAS 16423-68-0, 2 % w/v, solvent half water and half ethanol, the date, and flammable, contains ethanol.

The solvent is not a detail. Both published sources specify equal parts alcohol and water, seventy-two years apart, and the practitioner’s reason for drinking-grade ethanol is that laboratory and hardware grades carry added ingredients whose effect on the dye she could not vouch for.

Stage 1 (45 minutes). Dissolve 4.1 g of ammonium bromide in 45 mL of distilled water, sprinkle in 2 g of gelatin, cover and bloom fifteen minutes, then hold thirty minutes in the bath at 55 °C. Bloom the second gelatin — 5 g in 25 mL — alongside.

Stage 2, the dye goes in (2 minutes). Under the deep red safelight, stir 1 mL of 10 per cent potassium iodide into the salted gelatin, then 3 to 4 drops of the 2 per cent erythrosin stock. From this moment the material is orthochromatic and the earlier safelight is no longer adequate for it.

⚗Where the dye goes in, and why the three published methods are not interchangeable

Three sources, three different operations, and the page keeps them apart.

• Into the salted gelatin, before precipitation — the practitioner’s method, and the one you are performing. The dye is present while the crystals form.
• Into the silver solution, just before mixing — Baker’s 1941 method. Also “in the make”, but not the same operation: the dye meets the crystal surface as it is created rather than being present in the medium the crystals grow in.
• Bathing the finished, dried plate in an aqueous solution of 1 part in 5,000 — Wall’s 1924 method, which the course reports and does not adapt, because his bath contains ammonia and he does not restate it without.

Wall gives the Tier 1 statement of the trade-off, which is worth having: bathed plates as a rule have a higher colour sensitivity than plates coated with an emulsion the dye was added to during mixing, but they do not keep as well. He adds two constraints that read as superstition and are not: only perfectly clean glass dishes, because metals tend to reduce the dyes and cause fog, and home-bathed plates should be used soon.

The dye works by adsorbing onto the crystal surface, absorbing a photon the crystal itself cannot, and passing the energy or the electron into the crystal. Part IV owns that mechanism. What matters here is that adsorption is a surface phenomenon, which is why the dose scales with silver rather than with volume, and why there is an optimum rather than a “more is better”.

Stages 3 to 6 are Project 3’s, unchanged: temper the silver solution warm, add it at 5 mL per minute over ten minutes with steady stirring, hold ten minutes unstirred, stir in the second gelatin, bag it and refrigerate.

AgNO₃ + NH₄Br → AgBr(s) + NH₄NO₃

The precipitation. Ammonium bromide is a salt: this is a plain-silver make and no ammine complex forms.

Session two — wash, ripen, finish, coat

Stage 7, wash (35 minutes). As Project 3, but six changes of five minutes, thirty minutes in all, rather than six of three. That is the source’s own figure for this emulsion and it is a real difference; record it as one.

Stage 8, ripen (60 minutes). Stir in 5 mL of ethanol and hold at 55 °C for sixty minutes without stirring — a quarter longer than Project 3.

Stage 9, finals (5 minutes). Optionally 2 drops of 10 per cent potassium bromide as the restrainer. Then, at 34 to 35 °C, a further 5 mL of ethanol. No hardener: the default stands, and the wet layer is tender and handled by the corners.

Stage 10, coat (40 minutes). Coat paper and plate from the same melt, at the coating weights Projects 3 and 4 established, so that both comparisons are available. About 15 mL per 8 × 10 inch sheet for roughly 100 mg/dm², and 15 mL per 4 × 5 inch plate for the heavier plate coat.

Stage 11, the tests. Two, and they answer different questions.

1. The step wedge, at your fixed lamp and distance, against Project 3’s strip.
2. The spectral test, which is the one that matters: four identical exposures of the step wedge, unfiltered, through yellow, through green and through red, on four pieces of the same sheet.

Expected observations

• The salted gelatin goes pink when the dye enters, and the finished emulsion is a deeper, warmer yellow than Project 3’s by transmitted light.
• The deep red safelight is noticeably dimmer and reds in the room go nearly black.
• The unfiltered and yellow-filtered strips should be close, and the green-filtered strip should record something. On an undyed emulsion the green strip records almost nothing at all.
• The red-filtered strip should be blank, or nearly so, on both the dyed and the undyed material. That blank is the same fact as the safelight, and seeing them as one thing is the point of the exercise.
• A dyed emulsion fogs more readily than an undyed one held the same way. Record base density with every batch.

What the dye adds, what the safelight avoids, and where the three gels sit

• Intrinsic: what silver bromo-iodide already sees (350–500 nm) — ultraviolet, violet and blue — Project 3, and every undyed emulsion in this part
• What erythrosin adds (500–600 nm) — strong yellow, orange and green, with a marked want of sensitiveness in the blue-green — the eosine signature
• Deep red safelight, and the red gel (600–700 nm) — the 906 class filter, 15 W, not less than 1.2 m; the red-filtered test strip is blank for the same reason

Not measured. The two sensitivity bands render written statements — the intrinsic response of an undyed silver bromide emulsion, and a 1924 formulary's description of erythrosin's strong yellow, orange and green sensitivity with its want of sensitiveness in the blue-green. The safelight band is drawn from manufacturers' filter recommendations rather than a transmission curve. Your four filtered strips are what turn this drawing into a measurement. The coloured strip approximates where the visible spectrum falls and is a reading aid only; the wavelengths in the labels carry the information. The bands and curves are drawn to show the relationship, not measured.

What is happening chemically

The dye is on the surface, not in the crystal. A spectral sensitiser works by adsorbing to the crystal face, absorbing a photon in a region where silver bromide is transparent, and delivering the energy or the electron into the conduction band, where the ordinary latent-image chain takes over. That is why the dose is set per mole of silver — which is to say per unit of crystal surface — and not per litre.

And that is why there is an optimum rather than a maximum. Duffin’s page 125 is the sentence that turns the dye from a constant into a variable: sensitivity rises with the quantity adsorbed to a maximum, and a rapid loss of useful speed occurs if too much dye is added, with the position of that maximum being a function of the particular dye and the particular emulsion. So neither published dose is the optimum for your emulsion, and finding yours is a measurement rather than a lookup.

Why more dye is not more speed: the shape Duffin describes

• Useful green speed against dye adsorbed — drawn, not measured

Show the numbers behind this plot

A single curve of useful green speed against the quantity of dye adsorbed on the crystal surface, drawn to render a written statement rather than any measured data. From zero the curve rises steeply, flattens through a maximum labelled as point X, and then falls away noticeably faster than it rose. A vertical guide marks point X and is annotated as the maximum, with a note that it is reached before the crystal surface is fully covered by dye. Two shaded regions divide the axis. To the left of point X the region is labelled as under-dyed, where adding dye adds green sensitivity. To the right it is labelled as over-dyed, where adding more dye costs useful speed rapidly, and a second note records that commercial practice often deliberately sits below point X because the balance of other properties makes a smaller quantity desirable. Two markers sit near the left of the axis without numerical positions on it, one labelled as the practitioner's published dose of about 102 to 136 milligrams per mole of silver and the other as the 1941 manufacturing figure of about 170 milligrams per mole, and the caption states that neither is known to be at point X for any emulsion other than its own. The teaching the drawing carries is that the dose is a variable the student measures rather than a constant to copy, and that the penalty for overshooting is steeper than the penalty for undershooting.

Series	Quantity of dye adsorbed (arbitrary units)	Useful speed to green light — shape only
Useful green speed against dye adsorbed — drawn, not measured	0.00	0.02
Useful green speed against dye adsorbed — drawn, not measured	1.00	0.34
Useful green speed against dye adsorbed — drawn, not measured	2.00	0.62
Useful green speed against dye adsorbed — drawn, not measured	3.00	0.83
Useful green speed against dye adsorbed — drawn, not measured	4.00	0.95
Useful green speed against dye adsorbed — drawn, not measured	5.00	1.00
Useful green speed against dye adsorbed — drawn, not measured	6.00	0.93
Useful green speed against dye adsorbed — drawn, not measured	7.00	0.78
Useful green speed against dye adsorbed — drawn, not measured	8.00	0.56
Useful green speed against dye adsorbed — drawn, not measured	9.00	0.32
Useful green speed against dye adsorbed — drawn, not measured	10.00	0.14

Not measured. The shape renders Duffin's written statement that sensitivity rises with the quantity of dye adsorbed to a maximum reached before total coverage, and that a rapid loss of useful speed follows if too much is added. The two published doses sit somewhere on the left of this curve for their own emulsions; where point X lies for yours is what your own dye series would measure. The curve is drawn to show the shape, not measured from a real material. Your own materials will differ, and measuring them is what the sensitometry part of the course is for.

Data to record

Project 3’s fields, plus: the dye stock’s strength, solvent, date and lot; the number of drops and the computed milligrams per mole; the addition point (salted gelatin, before precipitation); the safelight filter, wattage and distance, and the date of the fog test re-run on this material; the four filtered step-wedge strips with their threshold and maximum-black steps; and the base density of an unexposed, developed strip.

And the design document itself: goal, levers, predictions, the gold-figure reading you chose, and what you would change next. That is a record of your reasoning, and it is the thing the capstone will want.

Analysis

1. The spectral test, which is the only one-variable measurement here

Read the four strips. The count between the unfiltered strip’s threshold and each filtered strip’s threshold is the filter factor for your material, in half-stop steps on a 0.15-increment wedge.

Then do the same four exposures on a Project 3 sheet. The difference between the two materials’ green results is the dye’s contribution, measured on your own bench, and it is the only number in this project that is not confounded by the four other differences between the two makes.

For scale: a current commercial orthochromatic film’s published filter factors are about 2.5 for yellow and 8 for tricolour green in daylight, and 1 and 4.5 in tungsten. Those are that manufacturer’s figures for their material, not a prediction for yours; they tell you the order of magnitude to expect and nothing more.

2. Speed and contrast against Project 3, with the confounds named

Lay Or-01a beside BrI-01a. Five things differ, so a difference in threshold is not attributable to any one of them. What the comparison is good for is a direction and a magnitude, and what it is not good for is a causal claim. If you want one, make Or-01b with the dye omitted and everything else held — one variable, and the cleanest experiment this project can produce.

3. The five-emulsion comparison matrix

This is where the part’s five batch records become one document. Fill every cell from your own notes; the row of column headings is the specification of what a batch record has to contain to be useful later.

	KCl-01	CB-01	BrI-01	Plate BrI-01b	Or-01
Halide, mol %
Halide excess, mol %
Gelatin, g per mole Ag
Washed? changes × minutes
Ripening: °C × minutes
Finals
Support and coating weight, mg/dm²
Safelight used, and fog test date
Threshold step, maximum-black step
Scale, in steps and stops
Image colour, and in which developer
Faults, and their diagnosis

Two rows are worth arguing with when it is full. Image colour — remember Wall’s controlled series: the colour is the developer’s business, so a colour difference across this row with one developer needs another explanation. And scale — the two Tier 1 sources this course holds give opposite directions for what bromide does to contrast, and your row is your own evidence on a question two manuals dispute.

Troubleshooting

What you see	Likely cause	What to do
High base fog on the dyed emulsion and not on the undyed one	Dye plus a sixty-minute ripening, or a safelight that is no longer adequate	Re-run the fog test on this material; add the 2 drops of 10 per cent potassium bromide; shorten the ripening by fifteen minutes and compare
The green-filtered strip records nothing	The dye did not reach the crystals, or the stock has degraded	Check the stock’s date and that it is stored dark; confirm the drops went into the salted gelatin before the silver, not afterwards
The dye stock has gone cloudy or thrown a deposit	Erythrosin is only sparingly soluble in water and the stock is half water	Warm gently and shake; if it will not clear, make it again. Do not filter it and assume the strength is unchanged
Everything is stained red	A control failed, and the stain is telling you where	Trace it back: the weighing, the transfer, or a glove you kept wearing. Change the step that failed, not the cleaning routine
Speed no better than Project 3’s, or worse	Expected, and not a fault	The dye buys green sensitivity, not blue speed. Read the filtered strips, not the unfiltered one. Over-dyeing also costs speed rapidly — see the plot above
The plates fog where the paper did not	Plates carry three to four times the coating weight and sit in the same safelight	Test the safelight against a coated plate, not against a commercial paper

Clean-up

As Project 3, with two additions. The dye stock and its dropper go back to the dark, labelled store as soon as the drops are counted; leaving a stained dropper on the bench is how a red fingerprint reaches a plate. And anything stained is washed before it is put away, because dried dye on a spoon will redissolve into the next thing that spoon touches.

Storage

The dye stock, in an amber dropper bottle, labelled and dated, away from heat and from anything that could ignite half a bottle of ethanol.

Coated sheets and plates, as before, but stored separately from the undyed material and marked ortho, because they need a different safelight and the box is where that gets forgotten.

What nobody has established: whether a dyed emulsion made this way keeps as well as an undyed one. Wall’s statement is about bathed plates, not about emulsions dyed in the make, and no source this course has read compares the keeping of the two. Date two sheets and find out.

Disposal considerations

The wash water, residues, fixer and first wash of Project 3, unchanged, all collected and none to a drain. The dye adds a few milligrams to that same stream and does not get its own route. Local regulation governs what may be done with the collected waste, and it differs between authorities within one country; Part II’s chemical waste and silver waste owns the collecting arrangement.

Questions

1. Write a one-paragraph specification for an emulsion two stops faster than BrI-01 at the same contrast. Name the levers you would move, in order of expected effect, and say which of them you could actually perform and which you could only design.
2. Duffin gives a gold range of 1 to 30 mg per mole and an optimum near 3 × 10⁻⁶ mol. Compute that optimum both as milligrams of gold and as milligrams of chloroauric acid trihydrate, say which reading makes his two statements consistent, and explain why a designer has to state which one they used.
3. Your emulsion carries 102 to 136 mg of erythrosin per mole of silver. What would you expect from doubling it, and what published statement are you relying on? Design the series that would find your own optimum, and say how many sheets it needs.
4. Explain why the spectral test is a one-variable measurement while the comparison with Project 3 is not. List the five differences between the two makes.
5. The manifest asked for a tetraazaindene or benzotriazole stabiliser at a stated dose. State what the page does instead, and write the two sentences that distinguish “we could not source this” from “we chose not to do this”.
6. A reader tells you erythrosin was banned by the FDA and should not be in a course. Write the two-part answer: what was revoked and on what legal basis, and what the same notice says about the mechanism’s relevance to humans.

Further experiments

The dye series. Five batches at 1, 2, 4, 8 and 16 drops, everything else held, each coated at the same weight and read through the green gel. That locates point X for your emulsion, which no published figure can give you.

The undyed control, properly. Or-01b, identical but with the dye omitted. One variable, and the only clean answer to what the dye did.

Wall’s bathing route, adapted and labelled as yours. He claims bathed plates have higher colour sensitivity than emulsion-dyed ones but keep less well. His bath contains ammonia and the course does not adapt it — so if you want to test the claim you are designing the bath yourself, and everything you publish about it is your own measurement, clearly marked.

Recap

• The design half is where the advanced chemistry lives, and it stays on paper: the double jet because no home procedure is published; sulfur and gold because the published record gives ranges, and the only tested gold dose sits inside an excluded make; the organic stabilisers because no dose exists at any tier.
• The make is the plain-silver orthochromatic emulsion: 0.0294 mol of silver, 44 mol per cent halide excess, 2.0 mol per cent iodide, dyed with 3 to 4 drops of a 2 per cent erythrosin stock into the salted gelatin before precipitation.
• Two published doses agree within a factor of 1.7 across seventy-two years, and neither is the optimum, because the optimum depends on the dye and the emulsion together.
• The restrainer replaces the stabiliser, at about 170 to 340 mg of bromide per mole from two sources.
• The real cost of the dye is a safelight, not a hazard: a deep red filter, a 15 W bulb, not less than 1.2 m, and a fresh fog test on your own material. That is the one place in this course where spectral sensitivity becomes something you can feel.
• Panchromatic sensitisation is Level D, study only: no hazard record exists for pinacyanol chloride or its relatives, and the work demands darkness.

Check your understanding

Question 1. What is the honest form of words about erythrosin's hazard classification, and why does the precise wording matter?

• It presents no hazard at all, since it is a permitted food colour
• It is unclassified under GHS on a thin evidence base — thirty-seven of thirty-nine notifiers report it does not meet the criteria, but only about five per cent of companies supplied information and two notifiers did lodge hazard codes — and there is no British workplace exposure limit, which under COSHH means exposure is reduced as far as reasonably practicable with no figure to work to
• It is a carcinogen, because the FDA revoked its food authorisation
• Its hazards are unknown and it should not be handled

Show the answer and why

Answer: It is unclassified under GHS on a thin evidence base — thirty-seven of thirty-nine notifiers report it does not meet the criteria, but only about five per cent of companies supplied information and two notifiers did lodge hazard codes — and there is no British workplace exposure limit, which under COSHH means exposure is reduced as far as reasonably practicable with no figure to work to

Two failure modes sit either side of the right answer and the course refuses both. "Not classified" is a statement about what companies notified, not a finding of safety, and PubChem records that the evidence base behind it is thin and incompletely disclosed. But the absence of a classification is also not a reason to treat the substance as unknowable. What follows practically is the same either way: the exposure route that exists is dust from a single weighing, so the powder is bought in the smallest quantity, made up in one operation, weighed over a tray in still air with gloves, eye protection and a mask, and the stock is labelled. The exposure limit's absence is what makes that discipline mandatory rather than optional, because there is no figure to demonstrate compliance against.

Question 2. The FDA revoked the authorisation for FD&C Red No. 3 in January 2025. What must a page that mentions this also say?

• Nothing else; the revocation speaks for itself
• That the revocation was compelled as a matter of law by the Delaney Clause after two studies found cancer in male rats, and — from the same notice — that the agency states the rat mechanism is hormonal and does not occur in humans, that relevant human exposures are much lower, and that claims of human risk are not supported by the available evidence
• That the ban applies only in the United States and can therefore be ignored
• That the dye is safe because it is still permitted as E127 in Britain

Show the answer and why

Answer: That the revocation was compelled as a matter of law by the Delaney Clause after two studies found cancer in male rats, and — from the same notice — that the agency states the rat mechanism is hormonal and does not occur in humans, that relevant human exposures are much lower, and that claims of human risk are not supported by the available evidence

Both halves, or neither. A reader who meets "banned by the FDA" somewhere else and finds no mention of it here stops trusting the course; a reader who meets the ban without the mechanism draws a conclusion the regulator explicitly disclaims in the same document. It also matters what the notice is about: eating the dye, at exposures relevant to food, and not dissolving a gram of it once and dispensing drops. Note that the Delaney Clause is a zero-tolerance legal instrument rather than a risk assessment, which is why a finding in rats compels an outcome even where the agency states the mechanism does not transfer. Options 3 and 4 are the two lazy answers the safety-language rule exists to prevent.

Question 3. Why is the double-jet taught here and not performed?

• Because it is dangerous
• Because no source at any tier this course could reach publishes a home double-jet make, and none quantifies the error a constant-rate single jet introduces, so building a schedule out of an industrial patent and a practitioner's note would be inventing a procedure
• Because it needs equipment that cannot be bought
• Because a single jet gives the same result

Show the answer and why

Answer: Because no source at any tier this course could reach publishes a home double-jet make, and none quantifies the error a constant-rate single jet introduces, so building a schedule out of an industrial patent and a practitioner's note would be inventing a procedure

The reason is evidential rather than practical or hazardous — a jacketed vessel, two syringe pumps and a silver electrode can all be bought, and none of them is dangerous. What does not exist is a published, tested domestic procedure, and Rule 6 does not permit assembling one. Option 4 is wrong in a specific and useful way: a single jet at constant volume per minute lets the halide excess fall continuously through the run, so pAg drifts and the crystals grown at the beginning were not grown under the conditions of those grown at the end. That is exactly what a controlled double jet fixes, and the patent literature states that a change in pAg alone is enough to move a double-run precipitation between regular cubes and octahedra.

Question 4. Your dyed emulsion measures no faster than the undyed one on an unfiltered step wedge. What has happened?

• The dye failed
• Probably nothing: a spectral sensitiser adds sensitivity in a region the crystal could not see, so on an unfiltered exposure dominated by blue you should not expect a large change — read the green-filtered strip, which is what the dye was for
• The emulsion was over-ripened
• The safelight fogged it

Show the answer and why

Answer: Probably nothing: a spectral sensitiser adds sensitivity in a region the crystal could not see, so on an unfiltered exposure dominated by blue you should not expect a large change — read the green-filtered strip, which is what the dye was for

The unfiltered strip is mostly reporting blue response, and blue is the region silver bromide already handled by itself. The dye's contribution shows up where the crystal was previously blind, which is why the test that answers this question is the four-filter set and not the plain wedge. Two other things are worth holding in reserve, though. Over-dyeing genuinely does cost useful speed, and rapidly, so an unfiltered result that is clearly worse rather than merely unchanged points at the dose. And a dyed emulsion under the old safelight will fog, which raises base density and eats threshold — which is why the fog test is re-run on this material rather than assumed from the earlier projects.

Question 5. The project asks for a stabiliser at a stated dose and the page supplies a bromide restrainer instead. What is the exact reason, and why does the distinction matter?

• Tetraazaindenes and benzotriazole are unsuitable for hand-made emulsions
• No source at any tier in this corpus gives a dose for either compound in an emulsion, so the course cannot publish one; the substitution is an absence of evidence, not a judgement about the compounds, and the two are different claims a reader is entitled to tell apart
• They are too hazardous at Level B
• The bromide restrainer works better

Show the answer and why

Answer: No source at any tier in this corpus gives a dose for either compound in an emulsion, so the course cannot publish one; the substitution is an absence of evidence, not a judgement about the compounds, and the two are different claims a reader is entitled to tell apart

Duffin's "Restrainers" section is about the adenine naturally present in gelatin — 30 to 50 parts per million in hide gelatins against 4 to 6 in ossein ones — which is good mechanism and not a dose for anything you would add. Nothing else in the corpus doses either compound in an emulsion. The bromide restrainer takes their place because it has two independent published doses, about 170 milligrams of bromide per mole of silver from a 1941 manufacturing formula and about 340 from a 2013 domestic one. Keeping "we could not source this" apart from "we chose not to do this" is what lets a later reader with a better library finish the job rather than assume it was already decided.

Check my answersClear