Flash Point Article

Stanford Memorial Church This article appeared in Flash Point, Volume 6, No. 2

IN DEFIANCE OF GRAVITY

THE RESTORATION OF STANFORD'S ANGELS

by William Kreysler

Four pendent mosaic angels adorn the high rounded walls just beneath the dome of the Stanford Memorial Church at the center of the Stanford University campus in Palo Alto, California. Originally installed near the turn of the century, the setting bed for the hundreds of thousands of glass tesserae was severely damaged during the Loma Prieta earthquake in 1989, causing some sizable chunks of glass and mortar to fall to the floor eighty feet below. Other sections were left hanging by the sheer geometry of their arched shape.

Archangel Our firm, William Kreysler & Associates, specialists in molded architectural products, was invited by Dinwiddie Construction Company to assist in the restoration of the angels. We worked closely with Terry Barnum, Dinwiddie's restoration specialist, and Lesley Bone of the de Young Museum, who was the project conservator. After inspecting the situation, the answer lay in one of two approaches. One was to hold all the tiles together, remove them from their setting bed, rebuild the mortar backing, and then reset the mosaics. The other was to hold the tiles where they were, go around behind and remove the structure, wire and mortar, and then rebuild the support structure, all from the backside.

Although the first approach was more conventional, it was time consuming and therefore expensive. The latter, without precedent and in defiance of gravity, held out the possibility of being faster, perhaps less risky and philosophically more appropriate, since it would preserve more of the original work. We, therefore, proposed a plan based on the repair-in-place approach. We would build a custom, precision-fitting, fiberglass support form and anchor it back to the structure around the perimeter of each mosaic angel. This form had to fit perfectly since any void between it and the face of the tesserae could prove catastrophic during the removal of the damaged mortar backing. The form also had to be completely insulated form the tesserae so as not to damage the surface in any way.

After such a precise support was finished and in place over the face of the mosaic, Dinwiddie's crew would carefully lift off the wire and damaged mortar from behind, leaving a 1-inch layer of mortar on the back of the tesserae themselves. This surface would then be meticulously cleaned and a new backing would be bonded to the back of the mosaics, tying them back to the existing steel framework.

In principle all of this sounded great. Making it work was a different matter. There were several problems. How would the support mold be made to fit the irregular shape? How would it be kept from damaging the glass surface? If the support mold were made of a structural material that would be molded to the surface (which turned out to be the case), how would we keep it from sticking? Then, assuming these problems were solved, how would the damaged loose mortar, which varied in thickness from 3/4 to 2 1/2 inches, be removed without pulling up pieces of the mosaic? And finally, how and what would be used to bond the whole back together and how would that be tied to the structure?

After much analysis and debate, a consensus was finally achieved. The plan was basically as follows:

First, we would survey the face of each angel to develop a topographical map that would be approximately accurate, plus or minus 1/4 inch. Next, a protective coating, specially developed by conservator Tracy Powers, would be applied to the face of the tesserae. Meanwhile, we would build a series of one foot wide, fiberglass-reinforced plastic (FRP) ribs that, when set side by side, would cover the entire mosaic surface (Figure 1). FRP is easily formed. When cured it is extremely strong, lightweight and durable. It is also easily drilled, cut, and bonded to, thus allowing for unexpected changes on the job. The ribs would be held in place by bolting them top and bottom through the mosaic itself in specified locations where the glass tesserae had been carefully removed (Figure 2). (A total of 1000 glass tiles were recorded, removed, and stored to be replaced once the ribs were taken down.)

Next came the difficult task of converting the approximate fit of the fiberglass ribs to a perfect fit. The plan was to saturate a special flexible foam blanket with a thermosetting polymer and lay it on the face of each rib. The wet foam would be compressed up against the mosaic face, conforming to the shape and filling whatever voids existed (Figure 3). When the polymer in this wet blanket cured, it would harden into a surface capable of withstanding the compressive forces anticipated. The ribs were thus successfully installed. The space between each rib was bridged with fiberglass and and resin applied in the field. When finished, we had a perfectly fitting lightweight, yet extremely strong, one-piece support form (Figure 4). The work could then begin behind the mosaic.

Dinwiddie's team began carefully cutting the backing loose with electric grinders and chisels, using gauges to control the depth. Meanwhile, Juri Komendant, our structural engineer, and Serge Labesque, my partner, chief designer and resident wizard, came to the realization that the flimsy little tees that were holding the wire and plaster and mosaic were also the main lateral support for the entire dome above the mosaics. The mosaics were actually involved in the structure of the building! This changed the game plan a bit since cutting away the mosaic's mortar meant we were reducing the stability of the dome. We all agreed on a modified procedure to minimize the risk to within acceptable tolerances.

As the mortar backing was being removed and the bracing upgraded, we were busy testing epoxy resins. We had decided the best way to consolidate the mosaic fragments as they rested on our support forms was to trowel a layer of thickened epoxy onto the back of the mortar and then build up a laminate of glass fiber and epoxy resin over the entire back (Figure 5). This left three problems unsolved. First, we needed to be sure the epoxy laminate allowed vapor to pass through the tile and the remaining mortar. Condensation and the permeability of the original material gave rise to concerns about trapping moisture at the mortar/epoxy interface. Experience has shown that this can cause delamination problems. Second, the thin fiberglass and epoxy laminate had to be stiff enough to provide adequate rigidity. And finally, the whole had to be tied back to the existing framework.

The vapor problem was solved by perforating the laminate. The stiffness problem was solved by adding ribs to the back of the laminate in a waffle iron pattern. Glass fiber and epoxy is stronger than steel in some ways but its weakness (sometimes its advantage) is that it is flexible too. It bends quite a bit without breaking. The fiberglass reinforced plastic (FRP) ribs solved this problem by keeping the whole from bending.

The third problem, tying everything back to the steel t-bars, was solved by allowing the stiffening rib pattern to include ribs that filled the space between the laminate and the underside of the steel framework (Figure 5). The framework was then wrapped in glass fiber that also went down over the ribs and onto the backing (Figure 6). The result is an extremely strong, well-bonded, composite panel that is stuck to the back of everyone of several hundred thousand tesserae. A crack perpendicular to the plane of the tesserae could occur and not a single piece would fall. The whole backing weighs about 1/10^th of what it replaced and is many times stronger.

After the laminate is cured, the temporary FRP ribs were removed from the front. The protective layers were peeled off and the mosaic was exposed undamaged, unchanged, and ready to face the ravages of time with the help of some of the world's most sophisticated materials. Composites or advanced composites, as some people refer to them, are among the most sophisticated materials available. And yet they are really not new at all. All that's new is the synthetic fibers recently developed and the extremely tough and durable polymers they reinforce. The idea of fiber reinforcement is as old as civilization. Ever since the first human discovered that letting a bit of grass or straw get mixed in with clay made for a much stronger dried mud or adobe, composites have been an integral part of our built environment.

Questions? Email us or
call 707-552-3500

Last updated: 09/21/06