To Anneal - or not to Anneal

by Woodrow Carpenter
(First published in Glass on Metal, vol. 23 # 3, June 2004, reproduced with permission of the author)

At the Enamelist Society Conference in 2003, two enamellists had questions concerning annealing. One, with glass working experience prior to enamelling, was puzzled that annealing is not a requirement in enamelling. The second, who had no glass working experience prior to enamelling, that someone would think annealing could be of value in enamelling. They wanted to know why it is required in glass working and not in enamelling. The answer was: 'the metal keeps the glass in compression.' This caused many questions, which leads to this article.

First, why does the glass worker have to anneal? There are two primary reasons. The first is easy to understand with common everyday logic. The thermal conductivity of glass is relatively low. When the surface of a glass rod is heated, there is a sizable lag in time before the center of the rod reaches the same temperature. And by that time, the surface is even hotter. Since glass expands with temperature, it is obvious the surface is expanding faster than the center. This sets up a strain which is opposed by internal restoring forces called stresses. If the stress exceeds a certain limiting value, rupture will occur. When cooling, the surface cools and contracts faster than the center, setting up stresses and strains in a similar manner. 

The second primary reason is not so obvious. Figure 1 shows the thermal expansion curve of a typical soda lime glass from room temperature to its softening point. The rate of expansion, #1, is relatively constant from room temperature to 900°F - 950°F, where it begins to rapidly increase its rate of expansion, #2. This is where the glass begins to change from the behavior of an elastic solid to that of a viscous liquid. This is known as the glass transformation temperature. Here, the molecular structure of the glass changes without the glass visibly changing. Our interest is the internal stresses which form as the glass cools from its softening point to room temperature.
More than 5000 years ago, practice showed that glass articles cooled directly from their working temperature to room temperature would usually crack or shatter. Practice also showed if the cooling glass was allowed to dwell at a certain point, and then gradually cooled to room temperature, it would not crack or shatter. We know now that certain point as the transformation area. The area where glass must remain in order to relieve stresses which may have formed while cooling to this stage as well as any stresses which may have formed because of the glass structure rearrangement while changing from a vitreous liquid to an elastic solid. Now the stress-free glass must be cooled slowly, because of the first primary reason given above in order to be stress-free at room temperature. Otherwise, it would crack or shatter.

Second, why do enamellers seldom anneal? Figure 2 shows the overall expansion of enamel is lower than that of copper. Enamel is selected with this basic requirement in mind. The rate of expansion of copper is linear throughout the normal enamelling temperature range. The enamel has two rates of expansion and a transformation zone similar to the glass worker's material. Point #1 in fig. 2 shows a small cross section of an enamelled copper article starting to cool. The enamel is a viscous liquid conforming to the metal surface. Neither the enamel, nor the metal places much stress on the other, even though they are contracting at different rates. At some point near the lower end of the enamel's transformation zone, due to the different rates of contraction, the enamel will be in slight tension, point 2, figure 2. This can be easily observed.
Apply a coat of enamel to one side of a 1"x 6" copper strip and fire, laying flat, on a sturdy flat firing fixture. When it comes out of the fire, the strip will be flat. Within 20 - 30 seconds, the ends will begin to raise. Point 2, figure 2, and photo #1. 45 - 60 seconds later, the ends will begin to lower, in about 60 seconds the strip will be flat again. Point 3, figure 2, and photo #2. Soon, the center begins to raise, giving the strip a convex bow, point 4, figure 2, and photo #3. The contraction of the metal has caught up with and exceeded the contraction of the enamel, removing the slight tension (bad) stress and replaced it with compression (good stress).The need for annealing has been removed and a needed stress applied. Otherwise, upon reheating, the metal's faster rate of expansion would put additional tension on the enamel, causing large cracks and sometimes large pieces to jump off.*