Unified molecular glaze calculation is often looked on by potters as the ultimate tool for accurate understanding of ceramic glaze behavior. While it certainly is the fundamental basis for truly understanding glaze chemistry, there are a number of variables that the Seger molecular calculation approach doesn't take into account. Hermann Seger's system assumes that there is a completely homogeneous, evenly distributed mixture of the chemistry present. In the real world of glaze batch materials sourcing and melting glass layers onto clay, things aren't always so nice and neat.

First, let us take a fired sample of an oxidation glaze fired in the Orton cone nine range which contains no soluble raw materials, and look at it in cross section at a greatly enlarged level. Picture the glaze and clay body sliced in two, and you are looking at the cross-section edge-on in a greatly magnified view. For the purposes of reference, the outer surface layer of glaze is "UP" and the clay body layer is "DOWN".

On the outer surface of the glaze (UP), the composition of the glass is pretty much what you would expect from the results of a molecular calculation utilizing Seger's approach. As you move toward the clay body (DOWN) the composition remains pretty constant for a while. Then, as you draw nearer the clay body, the composition starts to change as the effects of the materials in the body start to appear. Crystals developing in the clay body continue to grow into the glaze layer and the glaze erodes the clay surface and begins to dissolve body materials. Bits of feldspar and particles of silica in the body become fluxed by the glaze fluxes and begin to melt. And so on.

So the composition of this boundary layer or interface layer is different than that which would be predicted by molecular calculation alone. This change is gradual, not like a distinct line or plane. It is more of a zone of interaction. In some areas of the cross-sectional slice, it can often be hard to tell where the glaze stops and the body begins.

As a very broad generalization, this effect is less and less pronounced as the maturing firing temperature of a glaze goes lower. It is also greatly dependent on the exact composition of the clay body itself. This is one reason why glazes can look so different on differing bodies, and why certain glaze defects can occur on some bodies and not on others. A body that has a lot of "glassing materials" will have more of a tendency to let various components bleed into the glaze. A body that is prone to crystalline development will have more of these elements grow into the glaze. A body high in coloring materials will have more of an impact on color development in the lower glaze layers.

If the body formulation contains a chemical that has some unique properties, then these properties can then affect the glaze too. For example, if a flameware body contains spodumene, supplying lithium oxide as a flux, that lithium will tend to find it's way into the lower part of the glaze layer and can tend to cause effects which are promoted by lithium in the glaze melt, such as a change in the Coefficient of Thermal Expansion or color rendition with certain colorants. This is true for many materials that could be present in clay bodies including the very common iron oxide (Fe2O3) that those who fire in reduction depend on to tint and variegate their wares.

Now, let's change the firing atmosphere of the heating cycle in the prior example to a reducing one and assume that the cooling cycle is in oxidation, as would be found in most typical "reduction" firings.

On the outer skin layer of the glaze there now will be a composition similar to that predicted from calculations with all of the oxides calculated in their fully oxidized form. Because of the exposure to air during the cooling cycle, all of the reducible oxides, such as iron, will be in the most stable fully oxidized state. So on this surface skin, iron will tend to be in the red state, not the grey or black state. The exact thickness of this outer oxidized skin will be determined by the gas permeability of the glass melt to available atmospheric oxygen, and typically is very thin.

Just below this oxidized skin layer, the glaze materials will remain in the reduced state that was developed during the reducing part of the heating cycle as the glaze surface sintered and fused over. Therefore, certain elements will act in the melt different from those on the skin layer and based on what their RO configuration would predict: ie. any iron present would tend to be a flux acting on silica.

The interface layer we looked at above would still be present as before, but compounds containing reducible materials in the body will also have been affected by the firing atmosphere before the glaze sealed the penetration of CO and H gases. This will greatly affect the bleeding of coloring materials such as reduced iron oxide out of the clay body into the glaze as the FeO becomes involved in the melting of interface layer silica material.

Now, let's add some soluble soda ash into the source glaze batch to supply all of the Sodium Oxide in this formula, keeping the overall glaze composition exactly constant chemically. This change dramatically shows the impact of the raw material sourcing of the oxides on the final melt. Soda ash is completely soluble in room temperature water in the concentrations typically found in most cone nine glazes. So all of the Na2O will be in solution, not suspension in the wet raw glaze batch even though it is evenly distributed into the mixture.

Now, we dip a piece of bisqued pottery into the raw glaze batch and deposit a layer of glaze powder on the piece. The suspended glaze materials form an evenly distributed layer of the percentage composition of the glaze batch on the ware....... except for the soda ash content. The soda content is dissolved in the water and tends to go with it where it goes. So the soda ions are absorbed into the body wit h the water, and are not evenly distributed in the glaze powder layer like all the other insoluble materials. These soda ions that are now in the water inside the pores of the clay body could affect the melt of the body, however in practice they typically do not remain inside the clay.

The water used in glazing has to eventually evaporate out of the pot either before the firing, or during the very early pre- 212F stage of the firing. As the water migrates out of the clay toward the surface in order to evaporate, the soda ions go along with it. However they cannot stay with the water as it evaporates off the surfaces of the piece. So the soda is left behind in and on the outermost surfaces of the dry glaze layer.

If you were to now scrape off the outermost layer of this glaze and analyze it with very sensitive equipment, you would find that the content of sodium in this sample is far greater than that you would find in another sample you have taken from just below this outermost layer. The soda content of both of these samples would differ greatly from that indicated through a molecular calculation of the total raw glaze batch. Each of these areas of what we often think of a one single glaze would have a completely different molecular formula, and when compared to typical limit formulas, probably fall into different cone ranges. In this particular case wit the high soda content, the outermost layer would be the lowest melting glaze.

So in this case, the outermost layer is beginning to melt at a much lower cone than the rest of the glaze due to the high soda content there. This outer layer is, like the interface layer, not distinct, but is a zone. This outer layer will react with the kiln atmospheric conditions as you would expect for a glaze of that particular cone range. So the surface layer can become gas impermeable to CO and H much lower then the rest of the lower reaches of the glaze. And the surface layer can do things like trap free carbon particles in the rapidly melting surface. This differing composition can affect the reactions of the outermost surface on cooling conditions also. Many of the cone nine-ten Shino glazes are a good example of this phenomena, and can appear as quite shiny when looked at in the kiln at very low temperatures. Only the very outermost layer is shiny, the rest of the glaze is still far from melted.

I know of no real way to accurately predict all of this or take it into account in a calculated formula. If you know you have soluble materials, you know that this type of action will happen. The exact extent can only be established by testing, as far as I know. No available software has yet cracked the sourcing issues just mentioned here.

It is complicated to predict these kinds of behaviors in any exacting way. For example, to do this you need to know how much of a slightly soluble material actually will go into solution. That amount depends on variables such as the particle size of the raw material, if any milling of the batch is being done, the PH of the water used to suspend the batch, the time the batch spends mixed in the water, and the effect of other glaze components or suspenders on the soluble material. Whew!

In the case of my own Shino glaze which contains a high level of soluble soda compounds, I have done extensive testing and also know the formula of the batch with the soluble soda content completely removed, and with it all concentrated into one quarter of the total batch. This information helps me develop firing cycles, but it is really only an approximation utilized to assist my general understanding. It is not scientifically exact. Nothing takes the place of real world testing work.

So as you approach getting a better understanding of glazes through studying molecular calculation, keep in mind that you have some added variables that affect the final outcomes. Molecular calculation is a crucial tool to technical understanding, but you also need to understand a bit about the raw materials themselves and the action of the fire. Glaze does not exist in isolation. All things for you to consider in glaze development and in tracking down defects.