Solvent Action, Physical Development and Grain

Kodak’s data sheet for T-MAX 100 gives a diffuse rms granularity of 8. The sheet for T-MAX 400 gives 10. The sheet for TRI-X 400 gives 17. Three numbers, one scale, and it is tempting to read them as a ranking of three films.

They are not comparable. The T-MAX figures were measured on film developed in D-76; the TRI-X figure was measured on film developed in HC-110, Dilution B, and Kodak says so in the same paragraph as the number. Change the developer and the number moves. That is the whole subject of this page: grain is not a property of a film. It is a property of a film and a developer together, and the mechanism that links them is the developer’s ability to dissolve silver halide while it is reducing it.

The measured number, the perceived one, and the caption underneath

Part IV separated the two words and this page uses the separation rather than repeating it: granularity is measured, graininess is perceived, and the bridge between them is magnification. What Part IV did not do is read the small print, and the small print is where the developer lives.

Kodak’s footnote gives the reading conditions in full: the figure is read at a net diffuse density of 1.00, through a 48 micrometre aperture, at 12× magnification. Every one of those is a choice. Move the density and you sample a different part of the curve; widen the aperture and you average away the mottle you were trying to measure. And above the footnote sits the sentence that matters most here — the data in this section are based on development in a named developer at a named temperature.

✕Three numbers that cannot be put in a league table

T-MAX 100 at 8 and T-MAX 400 at 10 were both read in D-76 at 20 °C, so the difference between those two is a difference between the films. TRI-X 400 at 17 was read in HC-110 Dilution B. Between 10 and 17 sit a change of emulsion and a change of developer, and nothing on either sheet says how much of the gap belongs to which.

The reading skill is to check the conditions line before comparing any two published granularity figures, exactly as you would check the aperture and the density. A manufacturer who publishes both is being scrupulous; a table on a website that lists the numbers without them has thrown away the part that makes them mean anything.

What actually makes a negative look grainy: clumping

Kodak’s 1928 primer sets out the mechanism plainly, and it is not what most people assume. The graininess of a print is not the size of one developed crystal. It is clumping:

One of the causes of “graininess” or the coarsening of these tiny particles in the picture image is the fusion or clumping of the grains which occurs during development.

And then the observation that makes clumping a chemical problem rather than a geometric one:

Experiments have shown that several silver halide crystals in close proximity to each other, even though unexposed, may become developed and form a clump as a result of actual contact with an exposed crystal.

Read that carefully, because it contains something surprising. An unexposed crystal, carrying no latent image at all, can be dragged into development by touching a neighbour that is developing. The silver filament growing out of the exposed grain reaches the unexposed one, and the filament is metal: it is exactly the kind of conducting speck that Part IV’s electrode picture says a developing agent will donate electrons to. The clump is therefore a chain reaction across contacts, and anything that breaks the contacts breaks the chain.

Breaking contacts is what a solvent developer does.

Three ligands, three strengths, and one the course cannot put a number on

Dissolving a little silver halide during development is complex formation, the same chemistry as fixing, run deliberately weak. Part III’s table of ligands is the tool; this page uses it to rank the three solvents that appear in published developers.

AgBr + 2 Cl⁻ → [AgCl₂]⁻ + Br⁻

Chloride: the weakest of the three, and the one used in the fine-grain developers of the 1950s

AgBr + 4 SCN⁻ → [Ag(SCN)₄]³⁻ + Br⁻

Thiocyanate: five orders of magnitude stronger, and Kodak put it in a published formula

The formation constants come from the same appendix Part III used. Silver with two chlorides is 1.8 × 10⁵; silver with four thiocyanates is 1.2 × 10¹⁰. Thiosulfate, the fixing ligand, is stronger again. The ranking is the design space: chloride nibbles, thiocyanate bites, thiosulfate dissolves the image.

⚗The gap in the middle of this page, stated at the top rather than the bottom

Sulfite is the solvent nearly every fine-grain developer actually uses, and it is the one the course cannot quantify. Part III looked for a formation constant for a silver–sulfite complex and found none in a source it had read; the sulfite lesson looked for a measured solubility of silver bromide in sulfite solutions and found none either. Nothing on this page supplies either figure.

What is sourced is the effect and the manufacturer’s own explanation of it, which is enough to formulate with and not enough to calculate with. Where this page says “more” or “less” about sulfite, it means exactly that, and any reader who finds a measured solubility curve for silver bromide in sodium sulfite has found something this course wants.

The published solvent developers, and what each one uses

Formula	Solvent	Quantity per litre	Source
Kodak D-76	sodium sulfite	100 g anhydrous	Kodak 1928 primer; Kodak Ltd 1949 handbook
Kodak D-23	sodium sulfite	100 g anhydrous, with no alkali at all	Kodak Ltd 1949 handbook
Kodak DK-20	sodium sulfite and potassium thiocyanate	100 g sulfite, 1.0 g thiocyanate	Kodak Ltd 1949 handbook
Kodak Microdol-X	not disclosed	not disclosed	described by Kodak, composition never published
ILFORD PERCEPTOL	not disclosed	not disclosed	described by ILFORD, composition never published

Kodak’s DK-20 is the one worth dwelling on, because it is a Tier-1 published formula containing a strong silver solvent, and the course can print it: Elon 5.0 g, sodium sulfite anhydrous 100 g, Kodalk (sodium metaborate) 2.0 g, potassium thiocyanate 1.0 g, potassium bromide 0.5 g, water to 1 litre, dissolved in the order given, used without dilution, about 15 minutes at 18 °C. Kodak calls it an “extra fine grain developer”.

Notice the arrangement. One gram per litre of a ligand five orders of magnitude stronger than chloride sits alongside a hundred grams of a weak one, and half a gram of bromide is there to hold back the fog that a strongly solvent bath invites. That is a formulator’s balance, and it is legible because Kodak published the numbers.

!Microdol-X and Perceptol: what the course will and will not say

Both are named in the curriculum as the fine-grain developers a reader will meet, and neither has a published composition. Kodak’s own literature describes Microdol-X as “an excellent fine-grain developer designed to produce low graininess and high sharpness of image detail”, supplied in powder and liquid forms, and says nothing about what is in it. ILFORD describes PERCEPTOL as “an extra fine grain film developer… designed for use when very fine grain negatives are required and a decrease in film speed is not important”, publishes its stock pH as 7.68 to 7.82 and its specific gravity as 1.110, and does not publish the formula.

Under the course’s provenance rule, a formula is printed only where its source can be traced. The composition of these two products is not something this course has read in a document, so it prints neither, and it does not repeat a composition attributed to them elsewhere. What it teaches instead is what the makers do state — behaviour, pH, speed cost — and the class of chemistry that behaviour belongs to. That is a smaller claim, and it is one you can rely on.

Dissolution and redeposition: physical development inside chemical development

Part IV drew the distinction the whole of this section rests on. Chemical development builds the image out of the silver already in the crystal. Physical development builds it out of silver dissolved in the bath. A solvent developer runs both at once, and the second one is the reason it works.

Dissolve, transport, redeposit: what a solvent developer does to one grain and its neighbour

1. 1Two crystals in contact — one exposed, one not; contact is the condition for a clump
2. 2Development starts, dissolution starts — the solvent attacks every face of both crystals, exposed or not
3. 3Contact broken — both crystals are smaller, the chain is cut, the clump does not form
4. 4The dissolved silver comes back — onto the filament as physical development, or into the gelatin as fog

A mechanism drawing, not a micrograph. Kodak's 1928 primer supplies the dissolution and the clumping; the redeposition is the physical development Part IV defined; the course has no photograph of any of it and describes none.

The redeposited silver goes somewhere, and where it goes decides whether the developer is a good one. Landing on a growing filament, it adds density without adding a new grain, which is why a solvent developer can dissolve some of the image and still reach a usable maximum density. Landing in the clear gelatin between the grains, it is fog.

Dichroic fog: the same reaction, one step too far

Kodak’s primer names the overshoot precisely, and the description is a diagnosis you can use:

Negatives developed in a developer containing an excess of sulphite or one containing hypo or ammonia may show dichroic or green fog. This appears yellowish-green by reflected light, and a pink color by transmitted light.

The cause is given in the next sentence: dissolved silver salts “are reduced to metallic silver in a very fine state of subdivision”. Part IV already established why very finely divided silver is coloured at all while filaments are neutral black, so the two colours are not a coincidence — they are a particle-size measurement you can make by eye.

Two details in the primer’s account repay attention. It occurs “particularly in the shadow portions of the negative where no bromide is liberated during development” — because development releases bromide, and bromide is a restrainer, so the parts of the frame that are developing hardest protect themselves while the parts that are barely developing do not. And “fine grained emulsions are most susceptible”, which is exactly the wrong combination, since a fine-grained film is what a solvent developer is usually pointed at.

The same primer records a second route to the same stain, from the other end of the process: a fixing bath “old and exhausted” and carrying an excess of dissolved silver, or one that is not acid, will do it too. One appearance, two sources, one mechanism. Whenever silver is in solution near an emulsion and something reduces it, this is what you get; the course files the fixer route with fixing faults and the developer route here, and both are physical development happening where nobody wanted it. Bostick and Sullivan’s current instruction sheet for Pyrocat-HD gives the modern version of the warning for semi-stand work: dichroic fog “may result from extended development of high speed films”, and the remedy offered is a more dilute working solution and a shorter time.

The speed loss, and why it is not a fault that could be engineered out

Fine-grain developers cost film speed. ILFORD does not argue the point; it tabulates it.

Film and developer	Meter settings ILFORD publishes a time for
HP5 Plus (ISO 400) in ID-11 stock	400, 800, 1600
HP5 Plus in MICROPHEN stock	400, 800, 1600, 3200
HP5 Plus in PERCEPTOL stock	250 only
HP5 Plus in PERCEPTOL 1+1 and 1+3	320 only
FP4 Plus (ISO 125) in ID-11 stock	50, 125, 200
FP4 Plus in PERCEPTOL stock	50, 125

For a 400-speed film, ILFORD’s own chart offers PERCEPTOL at EI 250 and nothing faster: about two-thirds of a stop below the box speed, published as a meter setting rather than as a caveat. The sheet’s own description says the developer is “designed for use when very fine grain negatives are required and a decrease in film speed is not important”. That is a manufacturer stating the trade in the product description.

⚗Two mechanisms for the loss, and the course cannot weigh them

Speed, as Part IV defined it, is decided at the bottom of the curve, by the least-exposed grains that manage to develop at all. Two consequences of solvent action both land there.

The marginal speck is dissolved before it can work. A grain that received the smallest usable exposure carries a latent-image cluster of a few atoms on its surface. The solvent attacks surfaces. Given long enough, the speck goes into solution and the grain never develops, so an exposure that would have registered does not.

Less silver ends up in the image. Dissolving part of every grain removes silver from the negative altogether, so the same exposure yields less density — a loss of covering power, in the sense Part IV gave the term. Since film speed is defined by reaching a stated density above base, less density per unit exposure is less speed by definition.

Both are real and this course cannot tell you their proportions. No source it has read separates them. What follows practically does not depend on the split: the speed loss and the grain gain are consequences of the same reaction, so they cannot be separated by formulating more cleverly. A developer that dissolves less halide gives coarser grain and more speed; one that dissolves more gives finer grain and less speed. There is no third option, which is why every manufacturer’s range contains both kinds.

Non-solvent development, and the grain you see instead

Take the solvent away and the picture inverts. Nothing dissolves, so no marginal speck is lost and no covering power is thrown away: full speed, or a little above it. Nothing breaks the contacts either, so clumps form freely and the developed grains are as large as the emulsion allows. And with no dissolved silver in the bath there is no physical development to smooth the filaments, so each developed grain keeps its own hard outline.

The visual result confuses people, because a non-solvent developer produces a negative that is simultaneously grainier and sharper. The grain is coarser as a texture, and the boundary between one tone and the next is more abrupt. At a small enlargement the sharpness is what you notice; at a large one the grain is; and that crossover, not the chemistry, is what people are usually arguing about when they argue about developers.

ILFORD’s own recommendation table on the HP5 Plus sheet says it in two rows. Under Finest grain: PERCEPTOL, stock. Under Maximum sharpness: ID-11 at 1+3. The same manufacturer, the same film, two different products and two different dilutions, pointing in opposite directions.

Dilution is the lever, and it moves two things at once

D-76 stock carries 100 g of sulfite per litre. D-76 at 1+1 carries 50. That is the simplest intervention available on this whole page — you change the solvent concentration by a factor of two with a measuring cylinder and no new chemicals.

c₂ = c₁ × V₁ / (V₁ + V₂)

Sulfite in a diluted stock

where c₁ is the concentration in the stock, V₁ the volume of stock taken and V₂ the volume of water added. For D-76 at 1+1, 100 g/L becomes 50 g/L; at 1+3, 25 g/L. ILFORD publishes the same progression as specific gravity rather than composition — ID-11 stock at 1.090, 1+1 at 1.047, 1+3 at 1.022 — which is the dilution visible in a hydrometer reading.

Kodak states the consequence in one sentence on two separate publications: “For greater sharpness, but with a slight increase in graininess, you can use a 1:1 dilution of this developer.” The chemical data sheet adds that the dilution “requires longer development times, which may result in a slight increase in graininess”. Less solvent, less clump-breaking, coarser grain; less solvent, sharper edges.

ILFORD's published times for HP5 Plus at EI 400, against the dilution of the same developer

• ID-11 with HP5 Plus at EI 400, spiral tank

Show the numbers behind this plot

Three points from ILFORD's published development-time table for HP5 Plus at EI 400 in ID-11, plotted against dilution. At stock, meaning zero parts of water, the time is 7.5 minutes. At 1+1, one part of water, it is 13 minutes. At 1+3, three parts of water, it is 20 minutes. The three points do not lie on a straight line through the origin, and the caption makes the point that the time does not scale with the dilution factor: 1+1 halves the concentration but lengthens the time by only about seventy per cent, and 1+3 quarters it but lengthens the time by a factor of 2.7 rather than four. The same three dilutions correspond to sulfite concentrations of one hundred, fifty and twenty-five grams per litre in the analogous Kodak formula, and to ILFORD's own recommendation of the most dilute one for maximum sharpness.

Series	Parts of water per part of stock	Development time at 20 °C (minutes)
ID-11 with HP5 Plus at EI 400, spiral tank	0.00	7.50
ID-11 with HP5 Plus at EI 400, spiral tank	1.00	13.00
ID-11 with HP5 Plus at EI 400, spiral tank	3.00	20.00

Points are ILFORD's published starting-point times for a spiral tank, not the course's measurements. The line between them is drawn only to join the three published points and is not a fitted model.

🧮The confounded experiment, and how Part IX unpicks it

A reader who develops one roll in D-76 stock and one in D-76 at 1+1 and compares the grain has changed four things at once, and the comparison cannot tell them apart.

1. Solvent concentration halves, which is the variable the experiment meant to change.
2. Development time lengthens — from 7½ to 13 minutes on ILFORD’s figures for HP5 Plus in ID-11 — so the emulsion sits in the bath nearly twice as long, and the solvent that is present has longer to work.
3. Developing agent concentration halves too, so the rate at every point in the emulsion falls and local exhaustion becomes easier to reach, which is the subject of the next lesson.
4. Capacity per roll falls, because half the agent is present in the same volume. Kodak’s instruction for 1:1 is to use it once and discard it, and to start with enough volume.

Two of those push grain finer and two push it coarser, which is why the published outcome — slightly sharper, slightly grainier — is a small net effect rather than a large one. The way to separate them is a compensated-time arm, in which the dilute developer is given whatever time brings it to the same contrast, so that contrast is held and only the chemistry differs. Part IX builds that arm, and it is worth knowing now that the obvious version of this experiment is the one that answers nothing.

Under the microscope: what you can see, and what you cannot

The instrument is the one Part IV costed out, and the arithmetic there stands: the structure that matters is the mottle Kodak measure through a 48 µm aperture, and you need enough magnification to resolve variation on that scale rather than to resolve a single crystal.

What is worth looking at. Take one negative developed in a solvent developer and one in a non-solvent developer, at the same density — a mid-tone, not a highlight — and look at the texture of the mottle rather than at any individual speck. The solvent-developed area should read as finer and more even; the non-solvent one as coarser and more clotted, with the clumps larger and the clear spaces between them larger too. That difference is the clumping mechanism, made visible.

What you will not see, at any magnification you own. You will not see redeposited silver as a distinct thing, because it lands on filaments that are already there. You will not see the marginal grains that were dissolved before they developed, because they are not in the negative. And you will not see the boundary between chemical and physical development, because both are silver.

✕A note on what the course has not photographed

This page describes no micrograph, because the course holds none. The research plan proposed comparing two provided micrographs and this page cannot honour that, because no licensed image of a solvent-developed and a non-solvent-developed grain pattern exists in the course’s collection. Drawing one would be worse than saying so: a drawn micrograph is a claim about appearance dressed as evidence.

What you can do instead is make the pair yourself, from the two developers the lab has you mix, and it is a better exercise than looking at somebody else’s.

Recap

• Granularity is measured with its conditions attached: a density, an aperture, a magnification — and a developer. Kodak’s published figures name all four, and two figures measured in different developers cannot be compared.
• Graininess is clumping, not crystal size. Kodak’s 1928 primer records that an unexposed crystal touching a developing one can be dragged into development with it, so the visible grain is a chain of contacts.
• A solvent developer breaks the contacts by dissolving a little of every crystal while development proceeds. Chloride, thiocyanate and sulfite all do it; their formation constants rank them, and sulfite’s is the one the course has no figure for.
• Kodak D-76, D-23 and DK-20 are published and are printed here. Microdol-X and PERCEPTOL are not published, and are taught by their behaviour and their makers’ own claims instead.
• The dissolved silver is redeposited, on the filaments as physical development or in the gelatin as dichroic fog. Same reaction, one step too far.
• The speed loss and the grain gain are the same reaction, so no formulation separates them. ILFORD publishes PERCEPTOL at EI 250 for a 400-speed film and says in the product description that the loss is the point.
• Dilution is the practical lever, and it moves solvent concentration, time, agent concentration and capacity together. Kodak’s published outcome for D-76 at 1:1 is sharper and slightly grainier.

Check your understanding

Question 1. Kodak publishes a diffuse rms granularity of 10 for T-MAX 400 and 17 for TRI-X 400. What is the first thing to check before reading that as a comparison between the two films?

• The film speed of each
• The developer each figure was measured in
• Whether the films were 35 mm or 120
• The year each sheet was published

Show the answer and why

Answer: The developer each figure was measured in

Both sheets state the conditions in the Image Structure section, and they are not the same: the T-MAX figure is for development in D-76 and the TRI-X figure for HC-110 Dilution B. A granularity number carries a density, an aperture, a magnification and a developer, and comparing two figures across a change of developer measures both the film and the chemistry without saying how much of the difference belongs to which.

Question 2. Kodak’s 1928 primer says that an unexposed crystal in contact with a developing one may be developed too. Why does that make grain a chemical problem rather than a purely geometric one?

• Because it means the emulsion was coated too thickly
• Because the growing silver filament is a metallic conductor and behaves as a development centre for its neighbour, so anything that breaks the contact breaks the chain
• Because unexposed crystals are larger than exposed ones
• Because it only happens above 24 °C

Show the answer and why

Answer: Because the growing silver filament is a metallic conductor and behaves as a development centre for its neighbour, so anything that breaks the contact breaks the chain

Part IV’s electrode picture makes a developing grain a conducting speck that a developing agent will donate electrons to, and a filament reaching an unexposed neighbour supplies exactly that. Clumping is therefore propagation through contacts, which is why dissolving a little of every crystal — a chemical intervention — changes the visible grain without changing the emulsion at all.

Question 3. You dilute D-76 from stock to 1+1 and develop for the longer published time. Which of these changed?

• The sulfite concentration, from 100 to 50 g/L
• The developing agent concentration
• The time the emulsion spends in the bath
• The molar mass of sodium sulfite

Show the answer and why

Answer: The sulfite concentration, from 100 to 50 g/L, The developing agent concentration, The time the emulsion spends in the bath

Dilution scales every dissolved component together, so agent and solvent fall in the same ratio, and the published time then lengthens to compensate for the slower rate. Three variables move at once, which is why the observed result — Kodak’s "greater sharpness, slight increase in graininess" — is a small net effect and why a controlled comparison needs a compensated-time arm. Molar mass is a property of the compound and is unaffected by anything you do in a graduate.

Question 4. ILFORD publishes development times for HP5 Plus in PERCEPTOL stock only at EI 250, while ID-11 stock is listed at 400, 800 and 1600. What is the chemical reason?

• PERCEPTOL contains a restrainer that fogs the film at higher speeds
• PERCEPTOL is a more dilute solution and simply develops more slowly
• PERCEPTOL is a strongly solvent developer, and the dissolution that gives it fine grain also removes marginal latent-image centres and silver from the image, both of which cost speed
• PERCEPTOL is intended only for sheet film

Show the answer and why

Answer: PERCEPTOL is a strongly solvent developer, and the dissolution that gives it fine grain also removes marginal latent-image centres and silver from the image, both of which cost speed

Speed is set at the foot of the curve by the least-exposed grains that develop at all, and solvent action attacks exactly what those grains depend on: a latent-image cluster of a few atoms sitting on a surface that is being dissolved, and the silver that would have made the density. Both mechanisms are real and the course cannot weigh them against each other; what matters practically is that they are the same reaction as the grain benefit, so the trade cannot be formulated away.

Question 5. A negative shows an overall veil that looks yellowish-green in reflected light and pink when you hold it up to a lamp. Which two very different causes should be on your list, and what do they share?

• Over-exposure and under-development; both give too much density
• An excess of silver solvent in the developer and an exhausted or non-acid fixer loaded with dissolved silver; both put silver in solution near the emulsion where something reduces it to very finely divided metal
• Safelight fog and camera flare; both are optical
• A stale film and a warm darkroom; both are storage faults

Show the answer and why

Answer: An excess of silver solvent in the developer and an exhausted or non-acid fixer loaded with dissolved silver; both put silver in solution near the emulsion where something reduces it to very finely divided metal

Kodak’s 1928 primer describes both routes with the same two colours, and gives the same mechanism for each: dissolved silver salts reduced to metal in a very fine state of subdivision. Part IV explains why finely divided silver is coloured while filaments are neutral black, so the colour is a particle-size reading. The diagnosis matters because the two causes need opposite fixes — less solvent or less time on one side, a fresh acid fixer on the other.

Question 6. Kodak’s DK-20 carries 1.0 g of potassium thiocyanate per litre alongside 100 g of sodium sulfite and 0.5 g of potassium bromide. Why so little thiocyanate, and why is the bromide there at all?

• Thiocyanate is expensive, and the bromide is a preservative
• Thiocyanate forms a far stronger silver complex than sulfite, so a small quantity does a large amount of dissolving; and the bromide restrains the fog that a strongly solvent bath invites
• Thiocyanate is a developing agent and the bromide is the alkali
• Both are present only to adjust the specific gravity

Show the answer and why

Answer: Thiocyanate forms a far stronger silver complex than sulfite, so a small quantity does a large amount of dissolving; and the bromide restrains the fog that a strongly solvent bath invites

The formation constant for silver with four thiocyanates is about 1.2 × 10¹⁰, against 1.8 × 10⁵ for silver with two chlorides, so on a molar basis thiocyanate is an incomparably more effective solvent and is dosed accordingly. More dissolved silver in the bath means more opportunity for physical development in the wrong place, which is dichroic fog, and a restrainer raises the threshold that has to be crossed before an unexposed grain develops. Reading a formula as a set of balancing decisions rather than a list is the skill the assignment page asks for.

Check my answersClear