How it works · 05
Silver halide chemistry
How a few photons leave an invisible mark in a silver crystal, how a developer multiplies it millions of times, and why fast film is grainy.
From hours to a five-thousandth of a second
Show the data as a table
| Year | Entry | Exposure | Notes |
|---|---|---|---|
| c. 1826–27 | Heliography | 8 h | At least eight hours of sunlight; some estimates say days |
| 1839 | Daguerreotype | 10 min | Several minutes in bright sun, 1839 |
| 1839 | Giroux camera | 10 min | About 5–15 minutes in sun |
| 1840 | Petzval lens | 40 s | Under a minute for a daguerreotype portrait (1841) |
| 1835–41 | Calotype | 1.5 min | Around one to two minutes in sun (calotype, 1841) |
| 1842 | Cyanotype | 15 min | Several minutes to half an hour in sunlight |
| 1851 | Wet collodion | 5 s | Several seconds in daylight |
| 1854–56 | Ambrotype & tintype | 5 s | A few seconds in a studio |
| 1871–78 | Gelatin dry plate | 1/25 s | About 1/25 second in daylight |
| 1888 | Kodak camera | 1/25 s | About 1/25 second, the fixed shutter speed (1888) |
| 1891 | Lippmann plate | 2 min | One to several minutes in sunlight |
| 1907 | Autochrome | 2 s | Often a second or more outdoors |
| 1935 | Kodachrome | 1/40 s | About 1/40 second at f/8 in sun (ASA 10, 1935) |
| 1999–2000 | Camera phone | 1/5000 s | About 1/5000 second at f/1.8 in sun (recent phone) |
| 2010s– | Film revival | 1/400 s | About 1/400 second at f/16 in sun (ISO 400 film) |
Typical exposures in daylight. Each bar runs from the exposure out to the long end, so the shrinking bars show how much faster photography became. Exposure also depends on the lens and the light, so these are rough, period-typical figures.
Silver salts darken
Silver chloride, bromide and iodide, the silver halides, darken in light. Schulze showed in the 1710s that light, not heat, caused it; Scheele found that violet light worked fastest and that the dark stuff was metallic silver. Wedgwood and Davy made images this way around 1802 but could not stop them darkening. Printing-out papers such as the albumen print still worked by direct darkening, which needs minutes to hours of sun.
The latent image
A grain of silver bromide is a tiny crystal. When it absorbs a few photons, freed electrons gather at a defect and turn a few silver ions into a speck of metallic silver, a handful of atoms, far too small to see. This latent image was the real discovery behind the daguerreotype and the calotype, both of 1839–41: the picture is already there after a short exposure, only invisible.
Development: chemical amplification
A developer is a mild reducing agent. It cannot easily turn an untouched grain into silver, but a grain that carries a latent-image speck acts as a catalyst, and the whole grain, billions of atoms, becomes silver. One photon’s worth of change can be multiplied up to many millions of times. This is why exposures fell from hours to minutes once development was used, and to seconds with the wet collodion process. The gelatin dry plate of the 1870s, with grains ripened in warm gelatin, made exposures short enough to hold the camera in the hand.
Fixing
The grains that saw no light are still sensitive, so the image would fog as soon as it met daylight. A fixer dissolves them. Herschel found in 1819 that sodium thiosulfate (“hypo”) dissolves silver halides, and suggested it for photography in 1839; it is still used. Washing removes the fixer, which would otherwise stain and fade the image.
Grain and speed
A bigger crystal catches more light, so emulsions with large grains are faster but look grainier. Every film is a compromise between speed and fine detail. Silver halides are naturally sensitive only to blue and violet light; dyes added from the 1870s (colour sensitising) extended them to green (orthochromatic) and then red (panchromatic). Colour films stack three such layers, each turned into a coloured dye during development (agfacolor neu, kodachrome).
The chart below shows the result of all these improvements: from about eight hours for heliography in the 1820s to a few hundredths of a second for a hand camera by the 1880s, and on to the fraction of a millisecond a pixel needs today (see sensors).