Home Glass Chemistry Coefficient of Expansion (COE)



(with melting points for common metals)

All figures times 10 (-7)

Type of Glass

Coefficient of expansion

Bullseye tested compatible

(Also Uroboros 90)


Effetre (Moretti) sheets and rods

(some variation; should test)


Spectrum System 96

(also Uroboros 96)


Borosilicate (Pyrex)


Window (float) glass

(Also includes most bottles)

83 to 87 (depends on manufacturer)

May be even higher or lower

Source:  Manufacturer's data

Type of metal

Coefficient of Expansion

Melting point (°F)

Melting point (°C)





Brass, navy












Iron, cast












Steel, high carbon




Steel, stainless








Note:  These are for pure metals.  Alloys can vary widely.  I have seen other sources with slightly different COEs, but most are close to these figures.  (And besides, they're close enough for government work.)

Source:  U.S. Military Training Circular No. 9-237, "Welding Theory and Application."

Glass Compatibility

There are many different production methods and recipes for making glass. As a result, there are almost as many different types of glass as there are glass artists who want to use them. Types of glass range from basic window glass (called "float glass") to brightly colored stained glass (also called "art glass"), and many of the types of glass come in numerous sub-types and categories.

In addition, there are many types of glass coatings, such as iridescent and dichroic, which have unique properties when applied to glass. New types of glass and glass coatings are constantly being developed.

All of these different glass types are candidates for fusing, slumping, and other kiln-forming processes. Some can be used off the shelf, but others require testing to make sure they will work in the kiln. That’s because it’s likely that you will want to combine more than one different sheet of glass in your projects. If so, then you’ll need to make sure the glass you select is "compatible." Using incompatible glass may cause cracking or even shattering of the piece when it cools.


To better understand compatibility, let’s consider what happens when glass gets heated in a kiln. Like many other substances, glass expands when it gets hot and contracts when it cools. This change in density, which occurs at the molecular level, can be measured in a laboratory. A typical one inch piece of Bullseye brand glass, for example, will expand 0.0000090 inches for each 1 degree Centigrade (about 1.8 degrees Fahrenheit) increase in temperature. That’s nine-millionths of an inch!

This rate, which is commonly known as the Coefficient of Expansion (COE), is usually expressed as a whole number, rather than as a long decimal figure. Most Bullseye glass, for example, is said to have a Coefficient of Expansion of 90, and you will often hear glass artists refer to it as COE90 glass. Spectrum, another common glass, has a COE of around 96, while Corning’s Pyrex glassware has a 32 COE. Standard window glass, referred to as "float" glass by the glassmaking community, has a COE that is usually around 84-87, while Effetre (Moretti) glass, commonly used for lampworking, has a 104 COE.

These differences in expansion and contraction may not sound like much, but they are very significant on the molecular level. A 10 inch length of Bullseye glass, for example, will shrink about 0.046 inches (about 1 mm) in cooling from around 950 degrees Fahrenheit to room temperature. By contrast, a 10-inch piece of Spectrum glass will shrink about 0.049 inches over the same temperature range. That difference - .003, or three thousandths of an inch - sounds trivial, but it’s enough to ensure that you can’t fuse Bullseye and Spectrum together.

Two glasses with considerably different COEs are said to be incompatible. They cannot be fused together and should be kept in separate areas of the glass studio to prevent their accidentally becoming intermingled.

This is especially critical because you can’t always tell incompatible glasses just by sight. In the example below, Bullseye (90COE) and Spectrum (96 COE) glass has been fused together. All looks fine to the naked eye, but viewing the glass with a polarized film shows the underlying stress.

You can sometimes get away with using two different glasses where the COE is only one or two apart (say, a 90 with a 91), but not always. Sometimes even two glasses with the same Coefficient of Expansion can not be fused together. That’s because the laboratory test that determines COE takes place at a different temperature than the one the warm glass artist often uses.

There are really only two ways to know if your glass is compatible:

• Use glass that has already been "Tested Compatible" by the manufacturer.

• Conduct compatibility testing on your own.  This is a matter of fusing small squares of the glass to be tested to a base glass of known COE, then examining the fused strip by sandwiching it between two strips of polarized film.

The advantage of testing for compatibility is that a broader range of colors and textures are available. Also, most manufacturers charge a premium for "Tested Compatible" glass, so it’s generally less expensive to use glass that has not been tested.


top pix

bottom pix

Glass Types and Forms

In addition to compatibility, glass artists also differentiate among different types of glass in many different ways. One of the major criteria for differentiation is the transparency of the glass. Opaque glasses that do not transmit light are generally referred to as "opaques", as "opalescent" glasses, or as "opals." See-through glasses of various colors are usually called "transparent" or "cathedral" glasses. Combining more than one different opalescent or cathedral glass or color in a single kiln-formed work is common.

Several different companies offer lines of tested compatible glass, with the largest and most popular being Bullseye and Spectrum.  Other companies offering tested compatible glass include Uroboros, Effetre (Moretti), Wasser, and Gaffer.

Bullseye, which has produced tested compatible glass since the 1970's, is generally acknowledged as the market leader, with a broader product offering than Spectrum or other brands.  Spectrum's tested compatible program, initially launched in Spring 2000, contained glasses made by both Spectrum and Uroboros, and is marketed under the "System 96" name.  Although the two product lines behave similarly in the kiln, they are not compatible, so most glass artists and hobbyists choose one or the other brand as their primary glass for fusing and slumping.

It should be noted that Bullseye, Spectrum, Uroboros, and many other firms also manufacture glass that is not guaranteed compatible. (A complete list would also include companies such as Armstrong, Desag, Freemont, GNA, Kokomo, Wissmach, and Youghiogheny.  Sometimes the glasses made by these companies tests compatible for fusing, but often it does not.  If you wish to use any of these glasses for kiln-forming projects involving more than a single sheet of glass, you will need to test for compatibility.

Virtually any stained glass, whether tested compatible or not, can be treated with an iridescent coating that causes the treated side of the glass to take on a metallic sheen. Some liken this effect to a shimmering rainbow. The shimmer goes away when the piece is lit from behind, allowing the normal color of the glass to shine through.

Another popular kind of glass coating, called "dichroic",  has the unusual property of reflecting one color while it transmits another. This means that the different colors can be viewed by examining the glass at different angles.   This unique glass is manufactured by spraying a thin chemical film on the glass. This must be done in a controlled environment in a vacuum chamber, making dichroic glass one of the most expensive glasses made for kiln-forming. Because of this expense, dichroic glass is more commonly used in jewelry and similar items, or as an accent in larger scale fusing projects.

One final type of glass that is often used for kiln-forming is "float" glass.  Made by "floating" molten glass on a bath of molten tin, float glass is better known as common window glass. It is inexpensive and widely available. It also works well in the kiln, but care should be taken to test for compatibility if different brands and types of float glass are mixed together. If at all possible, cut pieces to be fused together from the same glass sheet.

Although some colored varieties of float glass are available, it is most commonly found in a clear (often slightly greenish) formulation.  It tends to slump and fuse at slightly higher temperatures than most art glass (about 75 to 100 degrees F higher), and can be prone to devitrification. Its COE depends on the specific formulation used and can be as low as 83 or as high as 90, but it generally ranges from 85 to 87.


Glass is available in several different shapes and sizes, the most common of which are:

Sheet glass - relatively flat, up to 30" wide.  Sheet glass is typically 1/8" (3mm) thick, but thinner and thicker varieties are available.

Frit - small, irregularly shaped glass pieces, available in sizes that range from around 1/2" (6mm) in width to powders.

Stringer - long, thin threads of glass, about 1mm in diameter and up to around 18 inches (45cm) in length.  Some thicker stringer and related shapes (called "noodles") are also available from some manufacturers.

Rods - round cylinders of glass, about 4-5mm in diameter and up to around 18 inches in length

Shards and confetti - slices of glass that are slightly thicker than a sheet of paper

Cullet, billets, patties, and dalle - various sizes and shapes of glass "chunks" used primarily for casting

For the scientific world: COE, or coefficient of expansion, doesn't really have to do with atomic vibrations - it's the amount that a material expands with a change in temperature relative to a standard.  For example, if you have a stuck pickle jar lid, you can run it under hot water - this causes the metal to expand faster than the glass, in part because metal has a higher COE (although it also conducts heat faster than glass, which helps too).


I was thinking of temperature - which is a measure of the average kinetic energy (KE = 1/2 mv2) of the atoms or molecules in a substance.  As you heat something up, you excite the atoms, which make them vibrate more, causing them to separate more, thus expansion.

Because we mix materials (and COEs) when we mix glasses, certain stresses can form when we first cool the marble.  The annealing process is a really slow heat and cool that allows the stresses to redistribute and even out throughout the material - eventually making it less prone to sudden failure.