A final look at Ptah-Sokar-Osiris

Julia Commander is a third-year graduate student in the Winterthur/University of Delaware Program in Art Conservation. She is currently completing a curriculum internship at the Penn Museum.

When we last checked in with the Ptah-Sokar-Osiris figure, I was working on finding a satisfactory cleaning approach. The figure has a darkened layer over the front surface, which obscures the beautiful patterns, colors, and hieroglyphs. My goal for cleaning was to clarify designs and improve legibility, although the sensitivity of the paint layers has made this an interesting challenge.

After cross-section analysis, I looked into instrumental techniques to better understand the condition issues. One promising technique was gas-chromatography mass-spectrometry (GC-MS) since the darkened layer was potentially a coating material. I took a sample by swabbing the dark layer from the wood substrate. Since only a small amount of material can be gathered this way, I collected several swabs in a glass vial for analysis. I sent this down to Winterthur Museum’s Scientific Research and Analysis Laboratory (SRAL), which has previously collaborated on samples from the Artifact Lab. Dr. Christian Petersen, a GC-MS specialist, sent back my spectra with some interesting results. He described the mixture as “waxy dirt,” which helps to clarify what likely happened to the surface. Wax may have been applied to consolidate the badly flaking paint, and this layer could have trapped dirt over time as the figure rested face-up in storage.

Focusing on the wax component did not immediately produce better cleaning results, and I continued testing gels with variations on solutions, application method, and timing. I eventually tried an application of Pemulen TR-2 gel, a polymeric emulsifier, with a proportion of solvent added. This gel was more effective for lifting the waxy grime and did not require excessive action on the surface. Used along with a silicone solvent barrier layer, I was able to lightly clean without lifting pigments from the surface. While I had some initial ideas about cleaning, this method was something that I only found through the process of trial and error.

L-55-29 detail, cleaning test

Even though I cleaned slowly in very small sections, the actual treatment step took much less time than the research, testing, and planning phases. Take a look at the results below.

L-55-29, before cleaning (left) and after cleaning (right)

Aside from cleaning, a few other steps were taken to stabilize the statue. The headdress, which is constructed from multiple pieces of wood, had a large gap that allowed the pieces to move individually. To add support and decrease movement, removable fills were made from Volara foam and Japanese tissue. These materials were turned into small “pillows” that were then pressure-fit into place.

L-55-29 headdress, shaping and fitting Volara foam fills

The figure, headdress, and base do not fit together in a stable arrangement. Instead of intervening further with the object itself, an exterior mount will be constructed to hold the components in place. This method has worked well with a similar Ptah-Sokar-Osiris figure, which you can see displayed in the Upper Egypt Gallery!

Another Ptah-Sokar-Osiris figure on display in Upper Egypt, showing the back of the figure and the mount holding the three pieces together.

Overall, this project provided quite a few challenges and an opportunity to explore cleaning techniques. Thanks for following along on this experience with Egyptian painted surfaces!

Analysis of the shabti box varnish

This is a long overdue post about the varnish on our beloved shabti box (in my last post I referred to the box as troubled…I’ve developed a tiny bit of a love-hate relationship with it, which I’m only now admitting).

A detail of the shabti box before treatment, showing the actively flaking and fractured orange-yellow varnish

A detail of the shabti box before treatment, showing the actively flaking and fractured orange-yellow varnish

Anyway, I’ve briefly mentioned that we believe that the varnish on our shabti box is a pistacia resin, but how did we come to this conclusion? I started out by doing some research into similar objects, and into painted wood from the New Kingdom in general. As I mentioned in a previous post, we know that some painted wooden objects were varnished with pistacia resin during this time period, and these varnishes often look like the coating we see on our shabti box. But there were some things about the coating, including the fact that it was actively flaking, and the fact that there are areas on the box where the paint is lost and where the coating extends over the loss onto the gesso below, which is strange.

In order to start characterizing the coating, I looked at the box under different light sources and did a microchemical spot test, all described here. All roads were leading toward the conclusion that the coating is pistacia resin, but since we had so many available samples (i.e. detached pieces of the varnish) I wanted to investigate further.

First, we turned to a resource that we have in-house: Fourier transform infrared spectroscopy, or FT-IR. FT-IR is a method of infrared spectroscopy, where IR radiation is passed through a sample, and some of the radiation is absorbed and some of it is passed through or transmitted. A spectrum is produced that represents the molecular absorption and transmission, which is unique to that material. I collected samples of detached varnish from the shabti box and from one of the shabti figures, and passed them along to Tessa de Alarcon, a conservator in our department, and consulting scholar Dr. Gretchen Hall. Here is what the spectra look like for each:

FT-IR spectra for samples of varnish from the shabti (top) and the shabti box (bottom).

FT-IR spectra for samples of varnish from the shabti (top) and the shabti box (bottom). The characteristic peaks are labeled on the top spectrum.

They look virtually identical, which confirms that the varnish on the box is the same as the varnish on the shabtis.

Dr. Hall then compared the spectrum for the shabti box sample to spectra for mastic (Pistacia lenticus) and terebinth (Pistacia terebinthus), both pistacia resins.

Spectra for (from top to bottom): the shabti box sample, a sample of terebinth, a sample of mastic from Chios purchased in Athens, and a sample of mastic from Kremer Pigments (the Kremer Pigment mastic sample spectrum was found in the IRUG database). IRUG = Infrared and Raman Users Group

Spectra for (from top to bottom): the shabti box sample, a sample of terebinth collected from the Uluburun shipwreck, a sample of mastic from Chios purchased in Athens, and a sample of mastic from Kremer Pigments Inc. (the comparative spectra were found in the IRUG database, IRUG = Infrared and Raman Users Group)

They all look very similar, with characteristic resinous acid peaks that occur between 1700 & 1720 cm-1 (carbonyl stretching) & the acid OH stretching that occurs ~1460 cm-1.

In order to see if we could classify the shabti box resin even further, Dr. Hall took a sample to Dr. Chris Petersen, Affiliated Associate Professor in the Winterthur/University of Delaware Program in Art Conservation (WUDPAC), where they analyzed it using Gas Chromatography-Mass Spectrometry (GC-MS). GC-MS is a technique that combines 2 methods of analysis, and in conservation we use it to analyze organic compounds.

Dr. Hall and Dr. Petersen ran the sample and here is what the GC-MS chromatogram looks like:

L-55-23A_GCMSlabeled2Dr. Petersen labeled the peaks and included their structures. The structures are consistent with pistacia resin, either mastic or terebinth. They did identify a peak for 28-norolean-17-en-3-one (#3 above), characteristic of heated pistacia resin, which could indicate that the resin was heated before application (which would have turned it from clear to a yellowed varnish). We cannot be certain what color the varnish was when it was first applied, but the analysis does confirm the fact that the shabti box and the shabtis all have aged pistacia resin coatings.

We are grateful to both Dr. Hall and Dr. Petersen for their work on this analysis!