Simplistically, LOI is the percentage of weight a material loses on firing. Assuming firing to a typical stoneware temperature of 1200C, the amount of weight loss can be surprising. Kaolins, for example, lose around 12% (mainly crystal-bound water). Ball clays lose about half of that (a combination of crystal water, organic carbon and sulfur). Many raw stoneware, earthenware and fireclays can have high sulfur content (high enough to create a strong odor during firing). Carbonates decompose and gas off CO2. Dolomite and calcium carbonate lose over 40% of their weight and copper carbonate is not far behind. Barium carbonate loses 20% plus and lithium carbonate a whopping 60%. Gerstley Borate and Colemanite are other high weight losers (20%+). Even feldspars lose some weight (usually less than 1%). LOI can also be other things (e.g. chlorine, oxygen, fluorine; the latter might break down quickly in one use but linger in a melting glaze or frit).
LOI-Adjusting an Analysis
If the original already had specified a 2% LOI, for example, then the formula to calculate the multiplier would be:
98 - 95 = 3 / 100 = 0.97
What material makes the tiny bubbles? The big bubbles?
These are two 10 gram balls of Worthington Clear glaze fired at cone 03 on terra cotta tiles (55 Gerstley Borate, 30 kaolin, 20 silica). On the left it contains raw kaolin, on the right calcined kaolin. The clouds of finer bubbles (on the left) are gone from the glaze on the right. That means the kaolin is generating them and the Gerstley Borate the larger bubbles. These are a bane of the terra cotta process. One secret of getting more transparent glazes is to fire to temperature and soak only long enough to even out the temperature, then drop 100F and soak there (I hold it half an hour).
How much gas escapes firing from cone 03 & 04?
These were fired to cone 03 (upper) and 04 (lower). At cone 03 the loss in weight is 4.54%, at 04 it is 4.45%. That is 0.08% difference. If a mug weighs 250 grams, that is only 0.21 grams. Does not sound like much. But wait. Air weighs 0.001225g/cc. While this is not the exact weight of the gases escaping during firing it suggests that around 170cc of gases need to bubble up through the glaze if the piece was bisque fired at cone 04 and glaze fired to cone 03.
Examples of DFAC disk, SHAB bar, LOI bar for clay testing
By preparing these three tests you can measure many properties of a clay body. These include drying shrinkage, fired shrinkage, porosity, drying performance, soluble salts content, water content and LOI.
Imposing an LOI in INSIGHT
Desktop Insight can calculate the LOI of a recipe based on the LOIs it knows of the individual materials in the recipe. But sometimes you need to impose an LOI to force a calculated analysis to match an actual measured LOI in the lab.
Desktop INSIGHT MDT dialog showing kaolin LOI
The LOI appears below the material name and alternative names (beside the weight). The formula that goes with that LOI is the bold numbers in the blanks beside the oxide names on the right.
Magnesium carbonate vs. oxide: One big difference
Here is a screenshot of side-by-side recipes in my account at insight-live.com. It takes 120 mag carb to source the same amount of MgO as 50 mag ox. I just made the two recipes, went into calculation mode and kept bumping up the magcarb by 5 until the chemistry was the same. Note the LOI of the magcarb version is 40. This one would certainly crawl very badly.
LOI is not important? Think again!
This chart compares the gassing behavior of 6 materials (5 of which are very common in ceramic glazes) as they are fired from 500-1700F. It is a reminder that some late gassers overlap early melters. The LOI (loss on ignition) of these materials can affect your glazes (e.g. bubbles, blisters, pinholes, crawling). Notice that talc is not finished until after 1650F (many glazes have already begin melting by then).
Why is that transparent glaze firing cloudy? The balls test us.
G1916Q and J low fire ultra-clear glazes (contain Ferro Frit 3195, 3110 and EPK) fired across the range of 1650 to 2000F (these were 10 gram balls that melted and flattened as they fired). Notice how they soften over a wide range, starting below cone 010 (1700F)! At the early stages carbon material is still visible (even though the glaze has lost 2% of its weight to this point), it is likely the source of the micro-bubbles that completely opacify the matrix even at 1950F (cone 04). This is an 85% fritted glaze, yet it still has carbon; think of what a raw glaze might have! Of course, these specimens test a very thick layer, so the bubbles are expected. But they still can be an issue, even in a thin glaze layer on a piece of ware. So to get the most transparent possible result it is wise to fire tests to find the point where the glaze starts to soften (in this case 1450F), then soak the kiln just below that (on the way up) to fire away as much of the carbon as possible. Of course, the glaze must have a low enough surface tension to release the bubbles, that is a separate issue.
Melting range is mainly about boron content
Fired at 1850. Notice that Frit 3195 is melting earlier. By 1950F, they appear much more similar. Melting earlier can be a disadvantage, it means that gases still escaping as materials in the body and glaze decompose get trapped in the glass matrix. But if the glaze melts later, these have more time to burn away. Glazes that have a lower B2O3 content will melt later, frit 3195 has 23% while Frit 3124 only has 14%).
Does this terra cotta clay have an LOI higher than kaolin? No.
These two samples demonstrate how different the LOI can be between different clays. The top one is mainly Redart (with a little bentonite and frit), it loses only 4% of its weight when fired to cone 02. The bottom one is New Zealand kaolin, it loses 14% when fired to the same temperature! The top one is vitrified, the bottom one will not vitrify for another 15 cones.
A body containing manganese bubbles the glaze
Laguna Barnard Slip substitute fired at cone 03 with a Ferro Frit 3195 clear glaze. The very high bubble content is likely because they are adding manganese dioxide to match the MnO in the chemistry of Barnard (it gases alot during firing).
G2931F Ulexite-based transparent bubbles, G2931K frit-based version does not
I melted these two 9 gram balls on tiles to compare their melting (the chemistry of these is identical, the recipes are different). The Ulexite in the G2931F (left) drives the LOI to more than 14%. That means the while the ulexite is decomposing during melting it is creating gases that are creating bubbles in the glass. Notice the size of the F is greater (because it is full of bubbles). While this seems like a serious problem, in practice the F fires crystal-clear on most ware.
Decomposing manganese granular particles in a buff stoneware causing it to bloat
A cone 6 stoneware with 0.3% 60/80 mesh manganese granular (Plainsman M340). Fired from cone 4 (bottom) to cone 8 (top). It is normally stable to cone 8, with the manganese it begins to bloat at cone 7. The particles of manganese generate gases as they decompose and melt, these produce volumes and pressures sufficiently suddenly that closing channels within the maturing body are unable to vent them out.
More carbon needs to burn out than you might think!
Hard to believe, but this carbon is on ten-gram balls of low fire glazes having 85% frit. Yes, this is an extreme test because glazes are applied in thin layers, but glazes sit atop bodies much higher in carbon bearing materials. And the carbon is sticking around at temperatures much higher than it is supposed to (not yet burned away at 1500F)! The lower row is G1916J, the upper is G1916Q. These balls were fired to determine the point at which the glazes densify enough that they will not pass gases being burned from the body below (around 1450F). Our firings of these glazes now soak at 1400F (on the way up). Not surpisingly, industrial manufacturers seek low carbon content materials.
Orange-peel or pebbly glaze surface. Why?
This is a cone 10 glossy glaze. It should be crystal clear and smooth. But it contains strontium carbonate, talc and calcium carbonate. They produce gases as they decompose, if that gas needs to come out at the wrong time it turns the glaze into a Swiss cheeze of micro bubbles. One solution is to use non-gassing sources of MgO, SrO and CaO. Or, better, do a study to isolate which of these three materials is the problem and it might be possible to adjust the firing to accommodate it. Or, an adjustment could be make to the chemistry of the glaze such that the melting happened later and more vigorously (rather than earlier and more slowly). The latter is actually the likely cause, this glaze contains a small amount of boron frit. Boron melts very early so the glaze is likely already fluid while gases that normally escape before other cone 10 glazes even get started melting are being trapped by this one.
Glaze melt fluidity comparison between G2931F and fritted G2931K show the effect of LOI
These two glazes have the same chemistry but different recipes. The F gets its boron from Ulexite, and Ulexite has a high LOI (it generates gases during firing, notice that these gases have affected the downward flow during melting). The frit-based version on the right flows cleanly and contains almost no bubbles. At high and medium temperatures potters seldom have bubble issues with glazes. This is not because they do not occur, it is because the appearance of typical glaze types are not affected by bubbles (and infact are often enhanced by them). But at low temperatures potters usually want to achieve good clarity in transparents and brilliance in a colors, so they find themselves in the same territory as the ceramic industry. An important way to do this is by using more frits (and the right firing schedules).
Insight-Live comparing a glossy and matte cone 6 base glaze recipe
Insight-live is calculating the unity formula and mole% formula for the two glazes. Notice how different the formula and mole% are for each (the former compares relative numbers of molecules, the latter their weights). The predominant oxides are very different. The calculation is accurate because all materials in the recipe are linked (clickable to view to the right). Notice the Si:Al Ratio: The matte is much lower. Notice the calculated thermal expansion: The matte is much lower because of its high levels of MgO (low expansion) and low levels of KNaO (high expansion). Notice the LOI: The matte is much higher because it contains significant dolomite.
Out Bound Links
In Bound Links