Silver Salts and Light: Schulze to Ritter

A white powder goes black in the sun. That is the whole of the effect this page is about, and it was known for two centuries before anyone made a photograph with it. By the end you should be able to say what the powder is, what the black substance is, why blue light does it and red light does not, and why the effect on its own is not photography.

What a silver salt is

Dissolve silver metal in nitric acid and the silver goes into solution as a positive ion, Ag⁺. Evaporate, and you have silver nitrate, AgNO₃: white or colourless crystals, CAS 7761-88-8, relative molecular mass 169.87, extravagantly soluble - PubChem carries 122 g in 100 cm³ of water at 0 °C - and, in CAMEO’s physical description, “becoming black on exposure to light or organic material”. That clause is the whole history of this subject in eight words.

Add a solution containing chloride ions and the two leave solution together as a solid. This is precipitation, and Kodak’s 1928 primer gives the pattern for chloride and bromide alike.

AgNO₃ + NaCl → AgCl + NaNO₃

Precipitation of silver chloride

The sodium and the nitrate stay dissolved. What matters is the curdy white solid, silver chloride, AgCl, CAS 7783-90-6, relative molecular mass 143.32. Where the nitrate dissolves by the spoonful, the chloride essentially does not: 1.93 mg per litre at 25 °C against 1,220 g per litre for the nitrate at 0 °C, a ratio above a hundred thousand to one even allowing for the temperature mismatch. Silver bromide is less soluble still, 0.135 mg per litre, and is described not as white but as yellowish.

Hold on to that last detail. A white solid absorbs almost no visible light; a yellow one absorbs at the violet-blue end and reflects the rest, which is why it looks yellow. The colour of the powder already tells you something about the light it can use.

🔬Under the microscope

Silver chloride in a bottle is not a smooth substance. It is a crowd of crystals, each a lattice of alternating Ag⁺ and Cl⁻ ions. Speed, image colour and whether the image can be developed at all depend on the size of those crystals and the defects in them, not on the bulk formula. Part IV takes the lattice apart; for now, picture a stack of ions with occasional flaws, and light arriving one photon at a time.

Two centuries of nearly noticing

The blackening was seen repeatedly by capable people and led nowhere. Angelo Sala wrote in 1614 that powdered silver nitrate exposed to the sun turns black as ink, and that silver nitrate wrapped in paper for a year had blackened the paper. Robert Boyle recorded the white precipitate and its darkening in 1667 but blamed the air. Wilhelm Homberg showed the Paris Academy in 1694 a box turned from beef bone, soaked in silver-bearing aqua fortis, blackened in the sun and cut back on a lathe to expose the pale bone as a marbled pattern; he never separated light from heat, and laid no stencil on it. All three had the material and none had the question.

🕰A correction worth carrying

Georg Fabricius is often credited with discovering the light sensitivity of silver chloride in 1565. His De metallicis rebus does describe horn silver - native silver chloride - as a translucent mineral the colour of leather, but Eder is emphatic that it says nothing about any change in light, and traces the claim to a sentence in Arago’s 1839 report on the daguerreotype that later writers copied unchecked. When a date here looks suspiciously early, a nineteenth-century summary quoted at second hand is the usual explanation.

Schulze, 1727: the control that mattered

Johann Heinrich Schulze - born at Colbitz on 12 May 1687, professor of medicine at Altdorf from 1720, later at Halle, where he died in 1744 - was trying to reproduce Balduin’s luminous stone. He needed chalk moistened with nitric acid, and the acid to hand was the kind kept for separating silver from gold, which carries a little dissolved silver.

Working at an open window in bright sun, he was astonished to see the surface turn a colour he recorded as atro-rubentem et in coeruleum vergentem, dark red inclining to blue, while the side of the dish turned from the sun took no colour at all. He had gone looking for a carrier of light and found a carrier of darkness, so he called it a scotophorus and published under that title in the first volume of the Acta physico-medica at Nuremberg.

What makes it an experiment is what he did next. His friends suggested heat, so he held the tube to a fire, turning the unexposed side towards the flame, until the glass was almost too hot to hold. Nothing changed colour. He shook the mixture uniform, hung a thread down the middle of the glass and left it in the sun for hours, finding a clean pale line where the thread had lain; he repeated it with horsehair, human hair and fine silver wire. Then he cut words out of paper, waxed the stencil to the bottle, and let the sun write them into the sediment.

🕰From the historical darkroom

Two details matter more than the picture. A single shake erased the writing and reset the material, so the images were demonstrations rather than objects: the permanence problem is there in the first photographic image ever made. And calcium nitrate alone proved insensitive, so the effect belonged to the silver. Eder reads the recipe as silver chloride and carbonate in silver nitrate against a white background - not far, he notes, from a printing-out paper.

✕The date you will see quoted

Many accounts date the experiment to 1717 and its publication to 1719, and the course could not find a source for either. The Latin memoir sits in the Acta physico-medica volume published at Nuremberg in 1727, and Eder - who reprinted that Latin text and wrote a monograph on Schulze - dates the work to 1725-1727. Until someone produces the 1719 printing, this course states 1727.

Thirty years later Giacomo Battista Beccaria, in Turin and apparently unaware of Schulze, ran the same test on silver chloride alone and concluded that the agent was light, not air as he had believed.

Scheele, 1777: what the black substance is

Carl Wilhelm Scheele was arguing about the nature of light, not about pictures, and used silver as evidence. He calls it well known that silver dissolved in nitric acid and poured on chalk blackens in the sun, that light reflected from a white wall does the same more slowly, and that “heat without light has no effect on this mixture” - then asks the question nobody had answered: “Should the black colour not be real silver?”

First, he proved that it is. He precipitated silver chloride with sal ammoniac, washed and dried it, and left it on paper in the sun for a fortnight, stirring to expose fresh surfaces. He poured aqueous ammonia on the blackened powder; some dissolved and a black residue did not. He washed that residue, dissolved most of it in nitric acid, and precipitated horn silver from the result. Something that dissolves in nitric acid and returns as silver chloride is silver metal.

Second, without recognising it, he performed the first fixing. The ammonia removed the silver chloride light had not acted on and left the image silver behind - the operation photography would spend sixty years failing to find. He had no picture to preserve, so the significance passed him by, and the answer to the permanence problem sat in print decades before Wedgwood and Davy gave up.

Third, he showed where the chlorine goes. He covered washed horn silver with distilled water, half in the sun and half in the dark. After two weeks the water over the blackened half, dropped into a silver solution, immediately threw down horn silver: it carried chloride. The water from the dark sample did nothing.

Then the experiment everyone remembers: a prism fixed at the window, the refracted sunbeam on the floor, a paper strewn with horn silver laid in the coloured light, “and you will observe that this horn-silver grows sooner black in the violet ray than in any of the other rays.”

=Expected result

By 1777 Scheele had established four things a modern textbook still states: light and not heat causes the darkening; the dark product is metallic silver; the halogen is released; and the response is strongest at the violet end. What he did not have was a reason.

Senebier, 1782, and the ends of the spectrum

Jean Senebier, librarian at Geneva, put times to it in a darkened room, with a spectrum thrown onto horn silver.

Light	Time to visible change
Violet	15 seconds
Purple	25 seconds
Blue	29 seconds
Green	37 seconds
Yellow	about 5 minutes
Orange	12 minutes
Red	20 minutes

The three slowest, he noted, never reached the depth of colour violet produced. These figures reach us through Eder and were judged by eye: treat them as an ordering, which is robust, not as ratios, which are not. He also recorded that horn silver sealed in glass turned violet within seconds, deepened for a minute, went umber after an hour and then stopped. A printing-out image runs out of steam, and we meet that ceiling below.

In 1800 William Herschel, measuring the heating power of the colours with thermometers, found the effect continuing past the red where there was no colour at all, and so discovered infrared. Within a year Johann Wilhelm Ritter looked for the other end: silver chloride on damp paper, the spectrum laid across it in a dark room, and the darkening beginning beyond the violet, where nothing could be seen. He published on 22 February 1801, noticing too that already-darkened paper went darker still at the violet end and lighter at the red - red light undoing what violet had done, an antagonism Part IV returns to.

🕰Where the word actinic comes from

Wall’s Dictionary of Photography derives it from the Greek aktis, a ray, especially of the sun, and defines it as “that portion of light which effects chemical change, in distinction to those portions which furnish light and heat”. Then comes the sentence that saves the definition: “the division is entirely arbitrary, as all depends upon the substance exposed.” Actinic is not a property of light but a relation between one light and one material. The course could not establish who first used the word in a photographic context.

What photolysis actually is

Everything above is phenomenology. Here is the mechanism, in the form latent-image theory has used since Gurney and Mott, as set out by the chemist Mike Ware.

⚗The chemistry

1. A photon absorbed by the crystal lifts an electron into the conduction band. The electron migrates until a defect or impurity traps it, where it reduces a silver ion to a silver atom: Ag⁺ plus an electron gives Ag.
2. Further photoelectrons and mobile silver ions from within the lattice arrive alternately at the same trap, so the speck grows atom by atom.
3. The positive hole left behind migrates to the surface, takes an electron back from a halide ion and leaves a neutral halogen atom. Two of those make a halogen molecule.
4. About four silver atoms make a development centre, where a developer can attack and convert the whole crystal. Making a speck you can see takes something like a million times more light.

Written as a net equation, in favour of the bookkeeping and against everything interesting:

2 AgCl + light → 2 Ag + Cl₂

Net photolysis of silver chloride

Nothing in it tells you that the silver appears in specks at defects, that the halogen has to go somewhere, or that the reaction reverses if it does not.

Where the chlorine goes, and why it decides everything

With no halogen acceptor present, the liberated halogen attacks the fresh silver at the surface and re-forms the halide. Irradiating a pure crystal therefore gets you almost nowhere: the particles stop growing near 10 nm and the yield saturates at an optical density of order 0.02, a barely perceptible greying. That is the ceiling Senebier met after an hour. Where something will take the halogen the reaction continues: cellulose, starch and gelatin react with it directly, and water disproportionates it, with excess silver ions pulling the equilibrium over.

Cl₂ + H₂O + Ag⁺ → AgCl + HOCl + H⁺

Water and excess silver as the halogen acceptor

Notice the product: more silver chloride. The halogen is partly recycled, so photolysis eats into the free silver in the coating instead of stopping at the crystal - and it releases hydrogen ions, so the coating slowly acidifies as the exposure runs.

Why violet acts and red does not

A photon’s energy rises as its wavelength falls, and a material with a threshold ignores light below that energy however much you supply. Semiconductor practice states the relation compactly: Hamamatsu gives a detector’s cut-off wavelength as

λ_c = 1240 / E

Threshold wavelength from threshold energy

where λ_c is the longest wavelength that still works, in nanometres, E is the threshold energy in electronvolts, and 1240 is Planck’s constant times the speed of light in those units. The example is silicon, not silver chloride; the shape of the argument is what transfers. For pure silver chloride Ware takes the sensitivity as ultraviolet and violet, negligible above about 420 nm - essentially the boundary Senebier’s table shows, found by counting seconds.

🔬Deeper: the real material sees further than the pure crystal

Ware tested a photogenic-drawing paper behind a yellow filter cutting everything below 470 nm and found the fogging rate only halved: real response well beyond what the pure salt absorbs. Print-out silver absorbs visible light itself and passes the energy on, so the image sensitises the material that made it. “The paper cannot see red” is therefore a working rule and never a guarantee, which is why safelight recommendations carry distances and time limits rather than a promise.

Printing-out is not developing-out

The image Schulze, Scheele and Senebier made is a printing-out image: light does all the work, the picture appears while you watch, and you stop when it looks right. Conservators describe it in exactly those terms, against a paper whose image appears in a bath.

The physical difference is particle size, and it shows as colour. The Getty Conservation Institute’s atlas states that silver formed photogenically is in much smaller particles than silver formed by development; Ware puts numbers on it, 10 to 100 nm against micron-sized filaments, and explains the colour by surface plasmon resonance in particles smaller than the wavelength of light. So a printing-out chloride image runs from light yellow-brown through red to darker brown, a printing-out bromide image is cooler and greyer, and a developed image is neutral black.

📷The photography

This is the first place in the course where a molecular fact becomes a visible one. A warm brown print is not a filter or a paper stock: it is small particles. Every warm-tone paper and sepia toner later in the course steers particle size, and you can now predict the direction of the shift before you read the box.

Why this is not yet photography

Kodak’s teaching primer put the limit better in 1928 than any modern summary: the earliest photographs were made by coating paper with silver chloride and forming images by its darkening under the action of light, but the sensitiveness of the silver chloride was too slight to use that way in a camera. The way round it is that light need not do the whole job - expose briefly, then continue the image by chemical action. That is the sensitivity problem stated by the people who solved it, and with the unexposed salt still in the sheet and no way to make copies, all three problems are on the table.

Deeper: why silver nitrate on plain paper is so feeble

Wedgwood used paper and leather soaked in silver nitrate, and it worked badly. Ware gives the reason in one line of redox bookkeeping: reducing Ag⁺ to Ag needs an electron, and silver nitrate solution has nothing to supply one, since nitrate is already fully oxidised and water needs a larger driving force than the silver couple provides. Silver nitrate alone is not photosensitive. It becomes so when something readily oxidisable is present - on paper, the cellulose and the sizing - but slowly, because the crystal, where electron transfer is easy, is missing. Form the chloride in the paper by soaking in salt and then in silver nitrate and you have crystals in a matrix that will accept the liberated chlorine. That is photogenic drawing paper.

Fact, interpretation and a sentence to be careful with

Take “Schulze discovered photography” apart. Schulze is fact. Discovered is defensible for the light sensitivity of silver salts under a heat control, which is what he demonstrated and published. Photography is interpretation, and Eder’s: he is the historian who advanced the claim and had a case to make. Schulze’s images died with a shake of the bottle, and whether that is photography depends on a definition that is ours, not his.

The same care applies at the century’s other end. Elizabeth Fulhame, in An Essay on Combustion with a view to a New Art of Dying and Painting (1794), described silk soaked in a solution of gold and hung in the sun, where a purple tinge and specks of reduced gold appeared over days - and ran a control, the same silk in a dried atmosphere, which produced nothing in three months. She applied the method to maps, drawing rivers in silver and cities in gold. Larry Schaaf has argued that this earns her a place among the forerunners of photography, and Herschel named her in his first written account of the subject in 1839. Whether light-drawn maps are photography is a question about definitions; that she did the experiment, and the control, is not.

Recap

• A silver salt is Ag⁺ with an anion. The nitrate is very soluble, the halides essentially are not, and mixing the two precipitates the halide.
• Schulze, published 1727, showed that light and not heat darkens a silver-bearing mixture, and made stencil images that a shake destroyed.
• Scheele, 1777, showed the black product is metallic silver, that ammonia dissolves the unchanged chloride and leaves it - fixing, unrecognised - that the chlorine is released, and that violet acts fastest. Senebier timed the colours; Ritter found the action continuing beyond the violet in 1801.
• Photolysis is electron transfer inside the crystal, and the fate of the released halogen decides whether the reaction continues or reverses.
• Print-out silver is small particles and therefore coloured; developed silver is large filaments and therefore black.
• It is not photography yet: printing out needs about a million times the exposure a developed image needs, and nothing here removes the salt light did not touch.

Check your understanding

Question 1. Schulze held the tube of silvered chalk close to a fire until it was almost too hot to hold, deliberately turning the unexposed side towards the flame. What was he testing?

• Whether the mixture would dry out faster when warm
• Whether heat, rather than light, was causing the colour change
• Whether the chalk would decompose at high temperature
• Whether the colour could be reversed by heating

Show the answer and why

Answer: Whether heat, rather than light, was causing the colour change

This is the control that turns an observation into an experiment. Sala, Boyle and Homberg had all seen silver preparations blacken, and Boyle blamed the air. By varying one factor - heat - while keeping light out of it, and getting no colour change, Schulze eliminated the rival explanation. Every experiment page in this course asks you to name your control for the same reason.

Question 2. Scheele poured aqueous ammonia onto silver chloride that had been blackening in the sun for a fortnight. Part dissolved and a black residue did not. What had he just done?

• Developed the image, by converting exposed silver chloride to silver
• Toned the image, by replacing silver with another metal
• Fixed the image, by dissolving the silver chloride light had not acted on
• Bleached the image, by re-oxidising the silver back to silver chloride

Show the answer and why

Answer: Fixed the image, by dissolving the silver chloride light had not acted on

Fixing means removing the light-sensitive salt that was not exposed, leaving only the image substance, and that is exactly what the ammonia did. Scheele was testing whether the black material was silver, not preserving a picture, so he did not see what he had. Development is the opposite operation: it amplifies an invisible latent image chemically rather than removing anything.

Question 3. A pure silver halide crystal is irradiated with strong violet light with nothing present that can take up the liberated halogen. What happens?

• It darkens without limit until all the silver halide is reduced
• It darkens to a faint, self-limiting grey and then stops
• Nothing happens at all, because photolysis needs an acceptor to start
• It darkens, then returns to white spontaneously in the dark

Show the answer and why

Answer: It darkens to a faint, self-limiting grey and then stops

Photolysis starts perfectly well - a photon frees an electron and a silver atom forms - but with no halogen acceptor the released halogen attacks the fresh silver and re-forms the halide, and the two processes reach a standoff. Ware gives a limiting particle size near 10 nm and a limiting optical density of order 0.02, which is barely visible. The surroundings of the crystal, not just the crystal, decide whether you get an image.

Question 4. Silver chloride is a white solid; silver bromide is yellowish. On the basis of colour alone, what would you predict about their response to blue and green light?

• Identical, because colour and sensitivity are unrelated
• The chloride responds further into the visible, because white reflects more light
• The bromide responds further into the visible, because a yellow solid absorbs at the violet-blue end
• Neither responds to visible light; both need ultraviolet

Show the answer and why

Answer: The bromide responds further into the visible, because a yellow solid absorbs at the violet-blue end

A solid looks yellow because it absorbs violet and blue and reflects the rest, and a photon that is not absorbed can do no chemistry, so the yellower halide can use light that passes straight through the white one. The prediction is right, and it is part of why bromide emulsions are faster than chloride ones in blue-rich light. Do not over-read it: absorption is necessary but not sufficient, and a real coating also depends on crystal size, on halogen acceptors and, in modern papers, on added dyes.

Question 5. Two prints of the same negative are made on the same paper: one printed out in sunlight until fully visible, the other exposed for a fraction of a second and developed. What do you predict about their colour?

• Both neutral black, because both images are metallic silver
• The printed-out one warmer or redder; the developed one closer to neutral black
• The developed one warmer; the printed-out one neutral
• Colour depends only on the paper base and cannot be predicted

Show the answer and why

Answer: The printed-out one warmer or redder; the developed one closer to neutral black

Both images really are metallic silver, which is what makes the first answer tempting, but the colour comes from particle size rather than composition. Light alone builds many small particles, 10 to 100 nm, which absorb selectively and read as warm browns and reds; development grows a few large filamentary particles from each speck, and those read as neutral black. The Getty atlas records this contrast between printing-out and developing-out silver gelatin papers.

Question 6. Why is "silver chloride is sensitive to actinic light" a poor sentence?

• Because silver chloride is not in fact sensitive to light
• Because actinic means ultraviolet only, and silver chloride also responds to violet
• Because actinic is defined by its effect on a material, so the sentence says only that the light which affects it affects it
• Because actinic is obsolete and has no modern equivalent

Show the answer and why

Answer: Because actinic is defined by its effect on a material, so the sentence says only that the light which affects it affects it

Wall's 1912 dictionary makes the point itself: the division of the spectrum into actinic and non-actinic is entirely arbitrary, because everything depends on the substance exposed. Actinic is a relation, not a property of light, so using it to explain a material's sensitivity is circular. Say which wavelengths, on which material, and how fast instead - which is what Senebier's table does, and what a spectral sensitivity curve does.

Check my answersClear