Manganese stain reduction on an Ancient Greek terracotta vase

Susan Costello, Katherine Eremin, and Georgina Rayner


A 5th century BCE Greek red-figure terracotta pelike (jar) at the Harvard Art Museums exhibited areas of black manganese dioxide staining from burial. In addition to ceramics, these black stains are found on bone, glass and stone. They are not considered harmful to the object and are often left as part of its archaeological history. The disfiguring staining on this particular ceramic made interpretation of the painted design difficult necessitating treatment. Studies have been published on reducing manganese staining from glass, but very little was found for ceramics. Thus, a research project was undertaken to develop a safe method to reduce the manganese staining.

A variety of treatment techniques were investigated including Nd:YAG and Er:YAG laser cleaning, and application of a range of chemicals by swabs and poultices. The latter was deemed the most promising option and a variety of poulticing materials, chelators and reducing agents were investigated. To avoid testing on the pelike itself, treatment options were evaluated first on terracotta mock-ups with artificial manganese staining and then on an ancient terracotta plate fragment with archaeological manganese staining. Based on the results, treatment was carried out on the pelike using a poultice of bentonite clay with 80:20 deionized water:ethanol. Bentonite is mostly sodium or calcium montmorillonite but also contains minor amounts of other minerals. It was chosen because it has a high ion exchange capacity (80-150 meq/100g) and thus was able to break the stain’s bond to the ceramic. After the poultice was applied, allowed to slowly dry and removed, a cotton swab dampened in water reduced the manganese staining. Because ethanol is a less effective solvent than water for soluble salts, it replaced a portion of water to minimize the amount of salts brought to the surface during treatment.

The 80:20 ratio proved to be the most efficient at preventing the majority of salts while maintaining bentonite’s ability to reduce the staining. The thickness and the water content (Water content (Wc) = weight water/weight dry poultice) of the poultices were critical factors. Poultices used for effective treatment were about 3 mm thick with a Wc of approximately 5. If the poultice was too thin or the liquid content too low, the poultice dried quickly and was ineffective. The manganese staining was characterized by x-ray fluorescence spectroscopy (XRF), scanning electron microscopy with energy dispersive microanalysis (SEM-EDS) and x-ray photoelectric spectroscopy (XPS). SEM-EDS, XPS and Fourier-transform infrared spectroscopy (FTIR) were used to analyze the bentonite poultice and the manganese stained terracotta before and after treatment. Results showed that the terracotta surface was unchanged and no bentonite was left behind. XPS analysis enabled identification of the manganese species present on the terracotta before treatment. The treatment of the pelike significantly reduced the manganese staining and achieved the desired outcome of a clearer interpretation of the painted design. The results of this research project can inform future treatments of manganese stained ceramics.


2018 | Houston | Volume 25