Consolidation of a Glass Pitcher

There will be an exhibition of Islamic glass objects this fall. All the objects need to be examined before they are put on display. This includes writing or updating the object's condition report and photographs. I was given a small glass pitcher from Egypt to check and treat.

The pitcher is small and narrow, made from undecorated glass, tinted yellow-ish green. It was blown, as there is a little pontil mark visible on the underside of the base where it was attached. Bubbles are also evident in the glass, as are some manufacturing lines. The rim and spot are made from a wide disk pinched together. The handle is made from a separate, thick piece of glass, which is slightly darker and greener than the body. It twists slightly, connecting to the body further to the right than where it joins the rim. The handle is folded under and doubled up underneath the rim. There are ridges for decoration around the edge of the rim, in the middle of the top, and where the handle meets the body. The latter ring is mostly missing. The bottom is concave, with a thick central pontil mark.

Glass pitcher, before treatment.

Although originally green and transparent, much of the glass has turned brown, iridescent, and opaque. The worst of it is on the right side, where it covers most of the exterior. The left side is not as deteriorated, with some clear, transparent spots. The handle is the least weathered, with the interior and the left side of the handle clean, smooth, and transparent. Rainbow iridescence occurs when very thin deterioration layers are superimposed on one another, parallel to the original surface, that cause a distortion in the transmittance of light. Under ultraviolet light, some of the weathering crust faintly fluoresced yellow-white. The spots are seen mainly on the bottom and handle of the pitcher. It could be from microcracks in the glass, or the pitcher could have been consolidated in 1975, though no record of treatment remains.

The glass is heavily weathered from burial, with a beige enamel-like weathering crust over much of one side. In burial conditions, the alkali in the glass gets leached out and replaced with water. This process produces degradation layers with a different composition than the rest of the glass. In the best cases, this creates a weathering crust which protects the bulk of the glass. In the worst cases, the glass is completely dissolved. The discoloring weathering crust is not a layer of corrosion products on top of the glass, but rather the original glass after some of the chemicals have been replaced. Therefore, the weathering crust is never taken off. In removing the weathering, the original surface is lost and the glass is more vulnerable to chemical and liquid attacks. In this case, the crust is unstable and flaking off, particularly on the lower body. The glass is rough and iridescent where the patina layer has flaked off. In the object's old conservation file, there is a single slide from 1975 that shows that although this deterioration and delamination was happening then, it has since gotten worse. Some of the flakes have been retained in a bag, and could be analyzed later.

The worst flaking edges of the weathering crust were consolidated with a 10% solution of Paraloid B-72 in acetone. Glass is usually not consolidated or coated because it often seals water in the glass, which leads to deterioration. Paraloid B-72 is more permeable by air and water than other consolidants, and looks much like the glass. The Paraloid B-72 was applied with a small, soft brush along the laminated edge of the weathering layers on the right side and bottom of the pitcher. The adhesive was quite thin and runny. This allowed it to be brought into the weathering layers by capillary action. When the acetone dried, the consolidant tightened the layers and secured them to the rest of the glass. The excess Paraloid B-72 was removed with acetone so the consolidant is not visible on the surface. This should protect the crust and prevent further deterioration or delamination.

Top left: before treatment. Top right: applying the consolidant. Bottom: after treatment.

Parthian Vessel Fill, Part II : Microballoons

Filling a small gap in the Parthian vessel with plaster was a good experience, so I decided to try to gap fill with glass microballoons, or glass bubbles, to see the difference. I picked another small gap on the side of the large Parthian vessel and created a barrier by pressing dental wax on the inside of the vessel wall. The wax is used to hold up the fill and create a smooth interior surface, but it is not always necessary with glass bubbles, unlike plaster. When thick and set enough, the glass microballoons can be molded and pressed on without worrying about the fill material dripping or being unstable. The fill material looks much like plaster. It is a white powder which, when mixed with liquid, becomes thick and workable before drying hard. The plaster was mixed with water, but glass bubbles are mixed with adhesive. I slowly added the powder to a solution of 40% Paraloid B-72 in acetone, stirring it in until a paste formed. I had to add a lot more powder than I thought, but unlike plaster, you add and mix until it is the right consistency, rather than adding it all and then stirring, hoping it is the right ratio.

 Left: Microballoon powder, before mixing. Right: Wax barrier, from interior.

The bulked glass bubble mixture was thick, but it remained wet and workable longer than the plaster. I pushed the mixture into the gap with a spatula, taking care to get the sides filled entirely. I tried to get it as smooth and level to the surrounding ceramic as possible. Unlike plaster, microballoon fills are not carved down after drying. This fill shrinks less. As I was working, a film formed on top the fill. The top was dry and set, but still flexible enough due to the wet fill beneath it. It felt squishy and I could use my finger to press the fill into the proper shape.

Glass bubble gapfill, before setting.

After drying, the glass microballoon fill was hard and rough. The surface felt like soft sandpaper. Microballoon fills are often used because they match the feel and look of ceramic so well. I could see a tinted fill blending in perfectly with the texture of many ceramics. In my case, it was decided that the fill should be polished down. Paraloid B-72 is soluble in acetone, and, as the base of the fill mixture, so is the glass bubble fill. To smooth down the rough parts, I dipped a spatula in acetone. The acetone softened the fill, and I was able to slowly push the fill around slightly. Using this method of steady strokes over the surface with a spatula dipped in acetone, I was able to smooth the fill, flatten the imperfections, and create lines along the edges to simulate material loss like the other cracks. At the end, the fill felt much like the plaster one.

After polishing. The dark smudges are trapped dirt particles from the ceramic.

After polishing, I painted the fill. Using the same water-based acrylic paints as the plaster fill, I tried to match the buff ceramic color. Once again, I was not exact. It is much harder than it looks to match, especially since the paint looks different once it dries. This time, I tried to add a bit of a subtle spattered effect, like the ceramic, instead of a single color field. In theory, this break up of the color makes it look more natural and blend better.

Left: Close up of fill in process, with yellow and brown speckles. Right: The finished product.

Overall, I quite liked the glass microballoon fill. I thought that it was fairly easy to work with and looks good with the ceramic. For rougher pieces, leaving it unpolished would make the fill look quite natural. The glass bubbles fill was good over a small area, but could be more difficult to mold into the proper shape if the gap is too large, complex, or curved. I found mixing and working with plaster harder, but it would work better for a large gap.

Islamic bowl reconstruction

The Islamic bowl was difficult to reconstruct. It had many issues with it: large gaps, worn or broken edges, and laminated pieces. 

Top: Front of bowl, with gray areas indicating plaster; Bottom left: Back of bowl, with grey areas indicating plaster; Bottom right: pieces after disassembly.

I started with the easy part, the rim, and worked my way inward. The rim was mainly intact, but even so, there were weak joins and unstable sections. I glued each section together, placing the sections in a box of glass beads, which acted like a sandbox, to prop the pieces up. While the Paraloid B-72 dries, you want the join to be horizontal so that gravity helps keep the pieces in place. I then mounded up tissue for a support when I put the rim sections together.

Left: sections of the rim standing in glass beads; Right: the completed rim.

I put on as many pieces as possible, but then it became hard to place the pieces exactly. I then switched to gluing the small slivers together to create larger pieces. The center is mainly intact, but there are only a few places where a sherd touched both rim and center, and none of those had good join edges. 

Small sherds grouped into larger pieces, drying against a weight.

Claudia called Stephen Koob, a ceramics conservator at the Freer who has written books and articles on ceramic reconstruction methods. He said that he has seen a Seljuk bowl with laminated pieces, though not as bad as mine. It could be as issue with a particular clay or firing. He suggested using a synthetic material called Flügger to gap fill. Flügger is like plaster, but it is more flexible and stays workable for longer. It is used more in Europe than in the States, though it started to be used fairly recently.

To cast the bowl properly, all the joins had to be perfect and the angle of the sherds correct. I decided the best way to do this was to break up the rim into six sections and fix each before putting them back together. I used acetone to dissolve the worst rim joins, then I used a hairdryer to heat up the thermoplastic adhesive on each join to manipulate the pieces back into shape. I also heated and fixed the joins on the smaller interior pieces.

I made a new support for the bowl out of carved foam, cotton, and tissue. I used a razor and scalpels to carve out the general shape of the bowl from the pieces of thick foam. I cut a large depression in the middle so that the top of the base was level with the sides of the bowl. This was padded with cotton, then a sheet of tissue was placed over top.

Carving the foam.

I nestled the pieces into the support and pinned them in place. The central section was quite stable, but it took some work to get the rest of the ceramic to fit. I left off the top pieces of laminated sherds. When they lay over a back section, I cannot put the pins in place. The first time I tried to put all the rim and back pieces in, I was left with a centimeter-wide gap which I physically could not close. I took the pieces all off, padded the support again to change the angle of the sides slightly, then tried again. This time, I was able to close the gap and get all the pieces to fit. Although it is still not exact, the pinned pieces looked good.

Top: First attempt, with a large gap in the rim on the right; Bottom: Pieces pinned into the support.

Parthian Vessel Joining and Filling

After desalination and gentle cleaning of the Parthian vessel, the potsherd edges were sealed with a thin layer of a 10% solution of Paraloid B-72 in acetone. This is to prevent the stronger adhesive used to hold the pieces together from getting in the porous ceramic, and it helps remove that adhesive. If a piece was out of place, or in the future they needed to disassemble the vessel, it would be quite easy to do with a bit of acetone. The pieces were then joined with a 40% solution of Paraloid B-72.

For the most part, the pieces fit well, but the ceramic was thick and heavy. The bottom pieces slipped a little when I added the higher pieces, so I had to bandage the vessel and continually check to make sure the joins were correct. I also had to use more Paraloid B-72 to cover the entire edge. This meant that excess adhesive was squeezed out when the joins were very close. After it dried, I removed the excess adhesive with acetone-soaked cotton swabs.

Several gaps remain, particularly around the rim. The vessel could have had a spout at one point, which would have easily broken off. Fortunately, these holes are small and do not affect the structural stability of the vessel. Two of the pieces do not belong to this pot, both a slightly different color than this pot. One is made from a much lighter and more porous ceramic. The other is smoother and thinner. During the Seleucia excavation, they accidentally got placed with the wrong vessel.

It was decided that this would be a good pot to practice fills. The lowest large hole, irregularly shaped (visible in the top left and bottom right images), was chosen. I placed a piece of dental wax over the loss area inside the bowl to support the plaster during filling. To do this, I cut a square of wax, then heated it to make it workable. My hands were not warm enough, so I placed it in hot water. I could have also used a hair dryer if I wanted it even warmer and softer. I then placed it inside the vessel and pressed it into the ceramic. I pushed the edges of wax hard to stick to the rough ceramic to get a good seal around all the edges.

The vessel was then placed on its side so that the hole was horizontal. I then filled it with plaster made with deionized water. The dry plaster was added to some water until it started to mound up, then mixed until it reached the correct consistency. It needed to be wet enough to work with, but thick enough to build up, like frosting. I dabbed wet plaster into the loss along the edges and undercuts, where it might have been hard to fill if I had just poured the plaster. After the initial layer, I mounded up the plaster with a spatula. I needed to work fast as the plaster dried surprisingly quickly, and it was set before I finished placing all the plaster into the fill. Even so, my fill was fairly good, though a higher fill would have been preferable. It was just proud of the ceramic, but I could have mounded up more plaster. Its easier to take away excess plaster than it is to fill in depressions and pockets.

After drying overnight, I carved down the fill to the level of the vessel. I first used a flat spatula and scalpel. The plaster scraped away fairly easily, though it was a bit difficult getting to the same level across the entire fill. After I got down, I used sandpaper to smooth the plaster. I removed a little plaster around the edges of the fills to simulate cracks around a sherd. This undercutting also made the piece look better in the spot where the fill was too low. The plaster was then sealed with 20% Paraloid B-72.

I painted the plaster with watered down acrylic paint fill to blend in better. I made a small piece of plaster to test the colors on. I used titanium white, yellow ochre, raw sienna, burnt umber, and raw umber, trying to match the general beige color of the ceramic. I wanted it to match the background color so that it would not be jarring or distracting, but I wanted it to be very recognizable as a fake fill.

I used several layers, moving from lighter to darker shades. The first layers were very watery, to get the basic tone of the ceramic. Upper layers were more pigmented, to the correct shading. The first few layers did not stick well to the sealed plaster, so there were water marks and visible brush strokes, though these were covered in later layers applied with a dabbing motion. Because I did not mix up enough at first, I had to keep mixing the paints for each layer. This meant that some layers were slightly more yellow, others more grey. They created an effect that works well for this ceramic. I probably should have made the color a little lighter; it is easier to work up from a light base than cover a dark one. Suzanne also told me after the fact that acrylic paints often dry darker and cooler than when wet. Because the plaster was sealed in Paraloid B-72, I could wash off the paint with water and start over if it was needed. In the end, the fill had the same basic tone as the rest of the original ceramic, but it is clearly a fill.

Greco-Roman Sandal

I cleaned a Greco-Roman sandal which was excavated from Karanis, Egypt. The straps were missing, and only the left sole remains. However, it is in fairly good condition.

Left: top of sandal. Right: bottom of sandal, before conservation.

It was made from woven raffia, a type of fibrous palm. The long pieces were split into six sections at the edges and woven into an intricate border. Most sandals darken over time due to oxidation and deterioration, but this one is still light gold, close to how it would have appeared in antiquity. A knot in the middle and two weaker areas on the edges show where the toe thong would have gone. The bottom of the sole is more damaged than the top, with dry and cracking palms. Some of the bottom-most layer of raffia is damaged or missing.

First, I brushed off the loose soil. I used a small vacuum, with the suction turned down low and gauze placed over the opening. I brushed from the center of the sandal to the edges, pushing the soil toward the vacuum. Not too much that came off with this method, as it had previously been brushed off.

To clean off the excess dirt, embedded from when it was buried in Karanis, Egypt, I used a 1:1 solution of ethanol and deionized water. I made small cotton swabs and shaped them either fat, for the flat sections, or pointed, for the edges. I moistened the swabs in the solution, then blotted it on paper towel. I wanted them to be just slightly damp, as I did not want to get the sandal wet. That would have caused more damage than good.

I gently rolled the cotton along the fibers. The idea is to pick up the dirt by rolling over it and lifting it off, rather than wipe and potentially work it into the material. I had to go over several places a couple times before the soil would come away.

I had to let the area dry and rest before returning, to make sure I was not keeping the organic material damp. There were some areas on the back that I could not clean very well, as they were deteriorated or detached. Too much motion would have damaged these areas more. The hardest part was the border. The dirt trapped in between the weaves was not being picked up by the cotton swabs. I had to use a small needle and brush to remove the soil there.

In the end, most of the dirt was removed, but not all of it. They did not want it to be perfect. It still looks worn and old, but it is brighter and cleaner now.

Left: front, after conservation. Right: bottom.

BTA Treatment of Bronze Figurine

I finished treating the Greco-Roman bronze figurine of a bear or wolf on top of a man. I scraped off all the active corrosion, getting as close as possible to the copper but not scratching it. The worst places were underneath the animal and in the incisions on the man. Unfortunately, it was quite hard to get the scalpel into these locations. Although I was able to remove most of the chlorides, there are probably still some spots of corrosion. It is hard to remove them all by hand, even using a small scalpel and microscope. The corrosion layer that remains is not active, and serves to protect the metal underneath. This layer is green and orange on my figurine, whereas the active corrosion was whitish and light green spots.

Satisfied with the results, I inhibited the bronze figurine to hopefully prevent future corrosion. To do this, I immersed the figurine in a 3% solution of benzotriazole (BTA) in ethanol.

BTA is used to coat industrial copper, just as pipes, to prevent corrosion. There is debate over its usefulness in conservation, as the old copper alloys treated this way generally already have corroded. BTA chemically binds to the surface copper, creating a barrier between the metal and the moisture of the environment, but corrosion underneath can erupt through the coating. However, it does seem to slow down the deterioration.

BTA comes as white, long pellets. In general, BTA is dissolved in water or ethanol at 1-3%. I dissolved 30 grams of BTA in 1000 ml of ethanol. This was done by gradually adding the pellets to the cold ethanol and stirring, but some people heat it and apply it hot to make the entire process faster. The solution can be sprayed or painted on, though the most effective method is to submerge the piece for 24 hours. The conservators I work with chose to use the higher percentage and reduce the time in the bath to just a few hours. In this case, I left the figurine in the BTA for four hours.

I took out the bronze figurine and let it air dry. After this treatment, gloves should be worn when handling the figurine. Occasionally, excess BTA dries on the surface as a white crystal, but that can be brushed off or wiped off with ethanol. The white patch that appeared on mine probably occurred from bleaching the adhesive used to put the accession number on the figurine. Sometimes the copper alloys are coated, such as with wax or lacquer, but the Kelsey's collection is in a stable environment so this step is unnecessary for my piece.

Consolidation of Ceramics

Today, I consolidated both the Islamic bowl I have been working on and a second Parthian jug.

The jug is larger and thicker than the other jug I did. I had previously cleaned and desalinated the ceramic. It took a long time to desalinate, over two weeks. There are still some soluble salts in the fabric, but the conductivity plateaued. The jug is unglazed and fairly porous. Although a couple pieces are missing, for the most part it is whole, broken up into two dozen pieces.

I used the 10% solution of Paraloid B-72 in acetone that I had made up last week to consolidate the break edges. Using a small paint brush, I sealed all the edges of these two ceramics, which took several hours total. The Parthian jug was easy. The break edges were wide and relatively cleanly broken. The Islamic bowl was harder. The pieces are quite small, and I only wanted the consolidant on the ceramic, not the glaze. When some went onto the glaze, I cleaned it off with acetone. For the laminated pieces which are just one sided sherds of glaze, I consolidated the non-glazed side.

The Paraloid B-72 dried quickly. Some of the ceramic edges look slightly darker than they were, but otherwise they appear to be the same. They feel slightly smoother now though. The consolidation will prevent the ceramic from soaking in the stronger adhesive that I will use next week to join the objects and will protect the pieces from crumbling. It will also make it easier to take apart next time if someone needs to treat it further.