Ceramic Recipes logo

The Lure of Lithium

Lithium is the lightest chemical element used in ceramic work. In fact, hydrogen and helium are the only known elements that are lighter. Because lithium is such a powerful flux, adding even a small amount of it to a glaze recipe can produce a big change in the fired result.

Defining the Terms

Coefficient of Thermal Expansion (CTE)—The amount of change in dimension of a material in response to a given change in temperature.
Both linear and volumetric CTE are measured,
so it’s important to know which is being discussed.
Linear CTE, which is typically used in ceramics,
is a very small number and is expressed in
scientific notation: nx10-6/ºF, where n is different
for each material.

Shivering—A glaze fault, occasionally called peeling, in which the glaze is in such powerful compression that it cracks and can literally pop off the ware. The root cause is that the glaze has a much smaller CTE than the body and thus does not shrink as much during cooling.

Crazing—A glaze condition in which the glaze cracks because it is in tension with the clay body, typically causing craze lines. Crazing weakens ware by about 1/4 to 1/5 the strength of uncrazed ware. Crazing is only considered a fault where it is unintended. Craze lines are sometimes emphasized for a decorative effect by filling the cracks with a colored stain or oxide, or rubbing with tea or ink.

The Well Rounded Flux

Lithium is a powerful, useful, alkaline glaze flux. It brightens glaze colors much like sodium and potassium do. However, lithium also promotes clearer transparent glazes and lowers both the melting point and the viscosity of glazes even more than sodium or potassium. Lithium also contributes to a harder fired glaze surface than either of those fluxes. It is well known as a flux in flameware clay bodies (see “Techno File: Flameware” May 2011, CM).

Lithium dramatically lowers the linear coefficient of thermal expansion (CTE) of glazes and clay bodies. This is part of why lithium is so useful in reducing susceptibility of flameware to thermal shock. However, the amount of lithium used must be carefully controlled. Too much lithium in a glaze is certain to cause that glaze to shiver from most ordinary clay bodies. Too much lithium also causes the formation of crystals, effectively turning a clear glaze opaque.

Magnesium is the only flux that has a lower CTE than lithium. Potassium and sodium, have CTEs on the order of five times larger than that of lithium, and the alkaline earth fluxes (calcium, barium, and strontium) have CTEs twice as large as lithium. 

Chemically, lithium is termed an alkali metal. It occurs in the same group of flux elements as the familiar sodium and potassium. When lithium reacts with other elements to form chemical compounds, covalent bonds result. These very strong bonds are the reason lithium contributes higher hardness to glazes than either sodium or potassium.

Lithium in Moderation

As a rule of thumb, use no more than 0.2 moles of Li2O per mole of total fluxes on any functional glaze at any firing temperature. However, even half that amount can cause shivering, so glaze fit (the degree of mismatch between glaze and clay body CTE) will usually determine how much lithium is too much.

Successfully using lithium as a glaze ingredient requires careful control of the glaze chemistry. The chart to the right shows three glaze recipes that illustrate this. The first glaze, 4-3-2-1, was made popular by Bernard Leach. The Lithium A and Lithium B recipes are, chemically, almost identical to the Leach 4-3-2-1. They differ only in that the Lithium A recipe replaces about half of the sodium and potassium of the 4-3-2-1 with lithium, while the Lithium B recipe replaces more than three quarters. Simple substitutions, right? Look at how dramatically different the  Lithium A and  Lithium B recipes are from the original.

Only 3% lithium carbonate was added to Lithium A, but to remove a molar equivalent amount of sodium and potassium, collectively, it was necessary to remove two thirds of the Custer feldspar from the original 4-3-2-1 recipe. Silica and alumina, which had been supplied by the feldspar, were then added back in the form of silica (about 50% more) and kaolin (more than doubled).

Just over 20% spodumene was added to Lithium B. To remove a molar equivalent amount of sodium and potassium, collectively, it was necessary to remove 85% of the Custer feldspar. Because spodumene is similar to feldspar, the silica and kaolin amounts in the recipe changed little.

The 4-3-2-1 is a basic, transparent, no frills glaze. When fired to cone 8 on porcelain,  there was some crazing within an opaque matte finish. With the addition of 3% LiCO3 in Lithium A, or 21% spodumene in Lithium B, the result is very similar in surface and color, but without crazing. Rather than working as simple one-to-one substitutes for other fluxes, lithium is such a light element that it is more powerful per gram than any other flux and thus fewer grams of flux are necessary to produce a desired result than with any other flux.


The point of the glaze chart is not to present recipes for studio use, but rather to show how easy it can be to substitute a small amount of lithium carbonate and produce a dramatic change in the recipe and the fired results. The CTE of the recipes indicate the original 4-3-2-1 recipe is likely to fit most stoneware clay bodies. Lithium A, however, might shiver off those same stonewares but will fit many porcelains which are formulated to have lower CTEs similar to glazes. Lithium B may even shiver off some porcelains.

This article was excerpted from the September 2011 issue of Ceramics Monthly, which can be viewed here.


Click Image to Enlarge



This item will be permanently deleted and cannot be recovered. Are you sure?

Once you submit this recipe for moderation you will no longer be able to edit its content.


Leave a Reply

You must be logged in to post a comment.

Enter Your Log In Credentials
This setting should only be used on your home or work computer.

Larger version of the image

Send this to a friend