All About Ceramics Crazing: Why Glazes Craze And How To Avoid Crazing

At the Hewitt Pottery we have been developing some glazes using local granites with a high felspatic content. The glazes are beautiful and sparkly but we have experienced some issues with crazing, so during our recent snowstorm, I spent some time reading through books and looking online to see what I could glean. I wanted to share the sum of what I have learned here in three sections: 1) Why is crazing a concern?, 2) Why causes crazing?, and, 3) What can we do to eliminate crazing?

1) Why is crazing a concern? 

Crazing can be an attractive feature of a pot and is often called "crackle" when intentionally used, such as on this tea bowl:

Intentional crazing, or "white crackle" glaze, on a tea bowl by Richard Brandt.

There are reasons why crazing is not ideal for functional pottery, however. Crazed pots may leak if the clay body is not totally vitrified, and potentially be unsanitary as bacteria can grow in the cracks. Structurally, crazing is also an issue as Michael Cardew points out in Pioneer Pottery, (p. 84), "It has been proved that glaze fit has a major effect on strength."

Cardew describes an experiment (recorded by Bettany and Webb in the British Ceramic Society's publication Transactions. Vol 40, p. 316), in which rods of porcelain are treated three different ways: some are left unglazed, some are dipped in a crazing glaze, and some in a sound glaze. The results after firing showed the comparative strengths of the rods in the proportions 40 : 100 : 160 (crazed : unglazed : sound). The results indicate that "vitreous ware with a non-crazing glaze may be three to four times stronger than ware which is crazed." Furthermore, a properly glazed pot will have a greater resistance to thermal shock. As a potter engaged in producing functional pottery, strength and resistance to thermal shock are important qualities, especially when one considers the competition of industrially produced wares.

The ideal, as Daniel Rhodes puts it in Clay and Glazes for the Potter (p. 255), for maximum durability and fit, "a glaze should be in slight compression over the body." See the kitten sweater below... it should be just a little bit snug for optimal cuteness.

This is Mango (not our cat, unfortunately).

2) What causes crazing?

John Colbeck (Pottery Materials, p. 61) says, "Crazing occurs when, on cooling, a body does not shrink more than the glaze." In other words, the glaze shrinks more than the body. It's as if Mango the kitten jumped into a tumble drier wearing her knitted sweater; the sweater would likely shrink more than her body and it would be quite a squeeze. This analogy doesn't totally work... but I wanted to include a picture of a cat in a sweater.

If you have a glaze that shrinks less than the clay body, then you can experience shivering, although this is more rare than crazing. {Side note from Michael Cardew on the difference between these defects: "It is always easy to tell the difference between shivering and crazing. Sometimes mild shivering may look like crazing, but the cracks are not on the surface of the glaze only; they can be seen right through the body" (p. 86).}

Crazing is not related to the shrinkage rate of the clay. As Rhodes points out (p. 255), all of the shrinkage happens when the wares are heating up. During the firing, whilst the pots are red hot and the glazes are still wet and molten, they fit the pots perfectly. It is upon the cooling of the kiln and the contraction of the wares that cracks form. The key point is that "some materials expand more when heated, and therefore contract more when cooling." This is called the coefficient of expansion. He goes on to say, "The cause of crazing, then, is always to be found in a high coefficient of expansion (and therefore contraction) in the glaze relative to the expansion of the body." I have copied out Rhodes' list (from English and Turner) of the expansion coefficients of some common materials used in ceramics:

SiO2          .05
Al2O3        .07
B2O3          .66
Na2O          4.32
K2O            3.90
PbO           1.06
ZnO           .07
CaO           1.63
MgO           .45
BaO           1.73

We can see from this list that oxides vary wildly in their coefficients of expansion: "Silica expands less than one eighteenth as much as sodium. Clay, being made up of alumina and silica, has a medium expansion; but some glazes, especially those high in soda, may have a high expansion" (p.255). So it's clear that your clay body and glazes will vary in their coefficients of expansion depending on the differing oxides present in them. Going into the science of this a bit deeper, we arrive at the formation of a substance called cristobalite. 

A Bit About Cristobalite:

John Colbeck explains (p. 62):

  • "Crazing needs to be considered in relation to clay bodies as well as glazes. The important factor to remember here is the role of cristobalite. Cristobalite, a crystalline form of silica, undergoes a contraction about 22°C (far below the temperature where glazes are molten). Cristobalite is formed quite slowly, at temperatures above 1,020°C, from the free silica which exists in bodies. It is not found in glazes because the free silica, whether high or low, react with fluxes to form the glaze solution. Thus bodies in which cristobalite has developed contract at 220°C as they cool, where glazes do not. It is this contraction of bodies which helps in the prevention of crazing by putting the cooling glaze under compression. Thus to diminish the tendency of a glaze to craze, any steps which assist the formation of cristobalite."

This leads nicely on to the important bit -- how to adapt glazes to reduce or eliminate crazing.

Two of my mugs from firing 96 at the Hewitt Pottery. If you look closely, you can see the crazing.

3) What can we do to eliminate crazing? 

Adapting your clay body:

Conventional wisdom suggests that adding silica to your clay body is the first port of call. Cardew says you can increase silica in either body or glaze, but that it tends to be more effective in the body.

Colbeck says you can assist the formation of cristobalite by adding silica to the body because this "will increase the free silica in it which is available to form cristobalite" (p. 62).

Rhodes agrees but warns, "bodies that contain more than about 25% of silica may be hard to fire without dunting or cracking." Conversely, bodies with "less than about 10% of silica... may be expected to be difficult to fit with glazes" (p. 256). There is clearly a sweet spot to be found with the amount of silica, and subsequently cristobalite, in your clay body. In A Potter's Book, Bernard Leach recommends 5-15% cristobalite in a body. He says this is enough to produce the cristobalite squeeze, "which exerts a centripetal compression on a glaze which tends to prevent crazing" (p.176). He notes that cristobalite can be cheaply acquired as powdered silica-brick waste.

Leach goes on, "the addition of powdered flint is usually the first alteration to a body to prevent crazing, but more important than an increase in quantity is a decrease in the particle size of silica." This is something that none of my other sources mentioned, but it is worth considering. He also mentions the option of increasing the cristobalite content using talc as "it acts as a catalyst and assists the transformation of silica" (p. 176). He also makes a distinction over vitrification; "in non-vitrified bodies the amount of flux should be increased and in non-vitrified bodies the reverse is true. Finally, in bodies which contain ball clay and china clay, the former should be increased, the latter decreased" (p. 177).

This advice is all well and good if you have the option of changing your clay body easily and testing it extensively, but many potters do not have this luxury. Changing the formulation of the glaze may be an easier option, or the only option.

Adapting your glaze:

The aim here is to reduce the coefficient of expansion of the glaze (to stop it contracting as much on cooling). This means adding oxides with low coefficients of expansion and decreasing some of the materials with higher coefficients of expansion. As Rhodes points out, this can be tricky "without altering the maturing temperature or appearance of the glaze" (p. 255). He recommends:

      1) increasing the silica
      2) decreasing the feldspar
      3) decreasing materials containing potash/soda
      4) increasing the boric oxide
      5) increasing the alumina

Leach also recommends increasing the silica content and possibly borax or raw boracalite (B2O3), and/or decreasing the alkaline content of the glaze.

Thinning down the glaze may also help reduce or eliminate crazing; as Colbeck says "thick layers of glaze are always more prone to crazing than thin" (p.62).

The website/database Digital Fire has some excellent articles on the subject. It recommends decreasing the potassium oxides and sodium oxides present in your glazes: these are typically found in potash feldspar, soda feldspars, nepheline syenite and frits. The issue with reducing your these is that the glaze may be less inclined to melt, so then you have to add some more flux and these adjustments may alter the look of the glaze. Digital Fire also recommends increasing your magnesium oxide. Talc and dolomite are excellent sources of MgO and purportedly effective at high temperatures, (for cone 6 you may want to use frits like Fusion F69 or Ferro 3249). They show one test where an addition of 10% talc helped eradicate crazing.

One of my teapots from Firing 96 at the Hewitt Pottery. The crazing is particularly obvious where the glaze is thick around the lid.

Final notes: 

You can tell how much your glaze does not fit your body by looking at the cracks: a network of lots of small cracks means you have greater stress than a few larger cracks. It is easier to fix the latter as you might expect.

Rhodes also posits that over firing can cause crazing, "if the firing has proceeded to the point where the free silica in the body has entered into glassy melts with the other materials, it does not go through any crystalline change upon cooling and so does not lose volume and put the glaze into compression" (p. 256).

Another issue can be removing pots from the kiln whilst they are still too hot: the kiln must be under 200°C. The heat shock of opening the kiln too soon or even putting wares atop of an oven/stove can induce crazing.

This post has been about primary crazing, but there is also such a thing as secondary crazing. This can happen with bodies which are not fully vitrified. Colbeck says porous bodies can "subsequently absorb water, causing the body to expand fractionally" (p. 62) which can craze a previously uncrazed glaze, months or even years after coming out of the kiln.

I don't want to end on a downer, but Daniel Rhodes does say that in glazes with a high content of "soda or potash in the form of feldspars, frits or raw alkalines," it may be "impossible to correct crazing without completely altering the character of the glaze" (p. 255). Our celadon glazes do contain a very high proportion of feldspar so the exercise could be tricky. It is worth a shot though -- especially if you have the ability to adjust your clay body as we do. I plan to test various methods and see what works best.

I hope some of this was helpful. Like a wise karate master from the movies, I leave you with these simple words...

Glaze on. Craze off. Glaze on. Craze off. Glaze on. Craze off. 

Works Cited

Cardew, Michael. Pioneer Pottery. London: A. & C. Black, 2002. 
Colbeck, John. Pottery Materials: Their Composition, Preparation and Use. London: Batsford, 1988.
Leach, Bernard. A Potter's Book. London: Faber and Faber, 1945. 
Rhodes, Daniel. Clay and Glazes for the Potter. Third ed. N.p.: Krause Publications, 2000.