Between 30 BC and the 3rd century AD, during which period Egypt was a province of the Roman Empire, a practice developed of attaching a portrait of a mummified individual to their mummy wrappings. Approximately 1,100 of these paintings have been collected over the centuries, the majority during the 19th and early 20th century – but, as many were bought and sold as works of art instead of archaeological artefacts, information about their context and provenance has disappeared. In 2013, to mitigate this loss and try to reconnect these mummy portraits once more, the Ancient Panel Paintings: Examination Analysis and Research (APPEAR) project and database was created, through the collaboration of more than 40 institutions. As part of this project, one team of researchers, from the University of Utah and Boise State University, recently analysed a microscopic sample of purple pigment taken from one painting – a 2nd-century AD example known as Portrait of a Bearded Man, which was originally from Faiyum in Egypt but now resides in the Walters Art Museum in Baltimore, Maryland. Could a sample of paint, no bigger than the width of a hair, provide any clues about this painting’s origins? In this month’s ‘Science Notes’, we examine the project’s findings.
The paint sample was removed from the clavus – a stripe on the shoulder of the bearded man’s tunic. In ancient Egypt, purple dyes were often reserved for royalty or those of high status, especially the highly valued Tyrian purple made from molluscs. But, by the time the portrait in question was painted, purple had been democratised and was used by all social strata. To examine the paint’s make-up, and to learn more about which type of purple this may be, the team broke up this already minuscule sample into even smaller parts to be analysed using multiple techniques.
First, based on the fact that the clavi glowed a pinkish-orange under ultraviolet radiation, it was thought that the purple paint was probably made from an organic lake pigment, which has nothing to do with a body of water but instead with how it is made: by mixing an organic dye with an inert binder, such as a metallic salt, in order to make it insoluble in water. This was confirmed using techniques such as XRF and SEM-EDS (see CA 366 and 367), which detected the individual elements that make up the sample: aluminium, titanium, silicon, potassium, iron, and sulphur – elements expected from a lake pigment – were all detected. The lack of bromine, however, confirmed that it was probably not made of Tyrian purple.
Interestingly, chromium and lead were also detected. The reason for their presence was not immediately clear – though the presence of the lead could be due to the crucibles in which the pigment was made. Descriptions of the dyeing vats discovered during Petrie’s 1908 excavations of the Ptolemaic Egyptian dyers’ workshops of Athribis suggest that they were lead-lined, so it is very possible that some of this lead would have been transferred to the pigment during production. The chromium remains unaccounted for, but another institution participating in the APPEAR programme also detected chromium in another mummy painting, Portrait of a Young Girl. While the two had not been connected before stylistically, re-examination does suggest similarities, including the panel shape and background tone.
Another way the paint was analysed was through atom probe tomography (APT), which uses a microscopic needle inserted into the sample to examine the distribution of elements within it. To the researchers’ knowledge, this technique had never before been used on a paint sample, and this study represents the first successful analysis of organic pigment in this way. This analysis, along with the assessment of a cross-section of the sample using transmission electron microscopy (TEM), showed three distinct ‘layers’ to the pigment. The outer layer was characterised by large crystals, in contrast to the inner layers which were made up of many small crystallites. This heterogeneity is unusual for a lake pigment, which before were thought to always be homogenous.
This layering suggests that the purple pigment may have actually been ‘upcycled’ using the dye sludge that remains at the bottom of dye vats. This would be in keeping with other apparent recycling efforts during the Roman period, such as the red dye used in the ‘red-shroud mummies’ from the Río Tinto site in Spain, which was shown to have been made using the lead oxides that remain after silver has been refined.
The full results of the study were recently published in the International Journal of Ceramic Engineering and Science and can be read for free at https://doi.org/10.1002/ces2.10075.