The origin of the colours in the photochromatic images finally understood

In 1848, in the Muséum d’Histoire Naturelle in Paris, Edmond Becquerel managed to produce a colour photograph of the solar spectrum. These photographs, which he called “photochromatic images”, are considered to be the world’s first colour photographs. Few of these have survived¹ because they are light-sensitive and because very few were produced in the first place. It took the introduction of other processes² for colour photography to become popular in society.

Solar spectra, Edmond Becquerel, 1848
Photochromatic images
Musée Nicéphore Niépce, Chalon-sur-Saône.

For more than 170 years, the nature of these colours has been debated in the scientific community, without resolution. Now we know the answer, thanks to a team at the Centre de recherche sur la conservation (CNRS/Muséum National d’Histoire Naturelle/Ministère de la Culture) in collaboration with the SOLEIL synchrotron and the Laboratoire de Physique des Solides (CNRS/Université Paris-Saclay). After having reproduced Edmond Becquerel's process to make samples of different colours, the team started by re-examining 19th century hypotheses in light of 21st century tools.

If the colours were due to pigments formed by the interaction with light, a variation of the chemical composition from one colour to another would have been observed. The spectroscopic measurements realized at SOLEIL by X-ray absorption (XAS / EXAFS) and high energy photoemission (HAXPES) on the ROCK, SAMBA and GALAXIES beamlines respectively, definitively led to the rejection of this first hypothesis. More precisely, the valence band observed by HAXPES (Figure 1(a)) on various colored samples differ only by a slight shift, which is attributed to the different sensitized layers thicknesses; the XAS spectra recorded at the Ag K-edge (Figure 1(b)) show that the sensitized layer does not undergo chemical change during the coloration.

Figure 1 : (a) HAXPES valence band spectra

Figure 1 : (b) XAS spectra at the Ag K-edge

If they were the result of interference, like the shades of some butterflies, the coloured surface should have shown regular microstructures about the size of the wavelength of the colour in question. Yet no periodic structure was observed using electron microscopy.

However, when the coloured plates were examined, metallic silver nanoparticles were revealed in the matrix made of silver chloride grains — and the distributions of sizes and locations of these nanoparticles vary according to colour. The scientists assume that according to the light’s colour (and therefore its energy), the nanoparticles present in the sensitised plate reorganise: some fragment and others coalesce. The new configuration gives the material the ability to absorb all colours of light, with the exception of the colour that caused it: and therefore that is the colour that we see. Nanoparticles having properties related to colour is a phenomenon known to physicists as surface plasmons³, electron vibrations (here, those of the metallic silver nanoparticles) that propagate in the material. A spectrometer in an electron microscope measured the energies of these vibrations to confirm this hypothesis.

This work was supported by the SACRe programme at the Université PSL, the Observatoire des Patrimoines de Sorbonne Université and the CEA and CNRS’s national network for transmission electron microscopy and atom probe microscopy.

-----------------------------------------

¹ Only a few dozen images of this type produced by Edmond Becquerel and then Niépce de Saint Victor are conserved in museum archives.

² For example, see: https://news.cnrs.fr/articles/the-birth-of-color-photography

³ This phenomenon, which explains the colours of objects as old as the Roman Lycurgus Cup, is studied today by physicists who hope for applications like ultrafast microprocessors and the improvement of various detector types.