|Monthly Tech-Tip |
Questions and suggestions to help you reason out the real cause of ceramic glaze blistering and bubbling problems and work out a solution
Blisters are evident on the fired glaze surface as a 'moonscape' of craters, some with sharp edges and others rounded. These craters are the remnants of bubbles that have burst during final approach to temperature or early stages of cooling. In some cases, there will be some unburst bubbles with a fragile 'dome' than can be broken. Blisters can vary in size and tend to be larger where the glaze is thicker. This problem is so serious that entire production lines can shut down when it hits. Generally potters will encounter this problem much more than industrial producers (the latter use more balanced glazes and cleaner clay bodies).
Often glazes appear like the melt should have plenty of mobility to heal but this can be deceptive. A melt flow testing regimen is the only way to know for sure about how much your glaze flows and if flow is changing over time. Melt flow testers have a reservoir at the top of a steep incline, the glaze runs down a calibrated runway. Generally, a fluid glaze will heal blisters much better, but only if it has time and if they have broken.
I am about to tell you something that could fix a blistering problem that has nagged you for years! It seems logical that a glaze not healing its blisters does not have enough melt fluidity. But very often, the opposite is the actual truth. How is it possible that a highly fluid melt can form blisters that do not heal? Bubbles can only form in a high-fluidity high-surface tension melt (analogous to the way bubbles form with soapy water). Reactive glazes are often of this type, those having high percentages of boron (e.g. 0.7 or more molar in a cone 6 glaze) so they melt well but at the same time having high surface tension. At the same time, such glazes often contain colorants and other materials, having high LOIs (producing a lot of gases of decomposition that become bubbles in the melt). Such glazes that form bubbles can be soaked for hours at top temperature and the bubbles may not break. If the temperature is dropping rapidly they may not break at all, and if they do at some point, what remains of the melt may not have time heal itself. This can happen even with fine porcelain bodies. Here is the key: Melt-fluid glazes need to avoid high surface tension. If that is not possible they need to be cooled to the point where the increasing viscosity of the melt is enough to overcome the surface tension that enables the bubbles to form and hold (of course that temperature is different for each glaze, it is normally found by decreasing the drop-and-hold temperature over a series of firings). This breaks the bubbles and provides enough time for the remaining melt fluidity to level out the surface. Each glaze has a different point at which this happens and ideally, cooling the kiln to just below that temperature and soaking there is the best solution. Failing that knowledge you can experiment with slow cooling the kiln through various temperature windows to discover the best results. Be aware that this temperature could be lower than you think, be prepared for it to be 200 degrees F or more below the firing temperature. Of course, if you are able to adjust the chemistry to favor fluxes that have a lower surface tension that will also help a lot. There is a linked photo below that shows how you can recognize high surface tension on a flow tester.
Significant amounts of gases can be generated within the glaze itself due to the decomposition of some materials after melting has started (i.e. dolomite, whiting, manganese dioxide, clays, carbonate colorants, etc). Substitute these materials for others that melt cleanly. For example, use frits, supply CaO from wollastonite instead of whiting or dolomite, use cleaner clay materials, or use stains instead of metallic carbonates. If you are using organic additives be aware that some of these can generate considerable gases during decomposition; do tests without them, use an inorganic substitute or find way to disperse them better into the slurry.
You might be under estimating the amount of gases that are coming out. Are you holding the top temperature long enough? Perhaps a much longer than expected soak might be necessary (on very thick tile or sculptural pieces, for example, 24 hours might be needed). Could you do a test on a small piece to confirm this? It might also work to adjust the firing schedule to soak, decrease the temperature a little (so the glaze is still fluid enough to level) and soak it there.
Gas release from decomposing materials in the body can continue until 950C. Many glazes begin melting long before this.
All clays release gases from burning of carbon material and decomposition of other compounds. Some clays release sulphur compounds also. If the glaze is melting during release of these gases, they must bubble up through it. If the melt is stiff, the kiln is ramped up too quickly, cooled too rapidly, or the glaze melts too early, it will not have opportunity to heal properly.
This may be a vindication of suspicions that a highly fluid melt of high surface tension is supporting bubble formation and existence. On the second firing the melt will be even more fluid than the first. It is doubly important that you identify the on-the-way-down temperature at which the glaze melt is too viscous to support stretching a bubble but fluid enough to still level out. When you find it, soak the kiln there.
This is a cone 10 glossy glaze. It has the chemistry that suggests it should be crystal clear and smooth. But there are multiple issues with the materials supplying that chemistry: 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. A study to isolate which of these three materials is the problem might make it possible to adjust the firing to accommodate it. But probably not. The most obvious solution is to just use non-gassing sources of MgO, SrO and CaO (which will require some calculation). There is a good reason to do this: The glaze contains some boron frit, that is likely kick-starting melting much earlier than a standard raw-material-only cone 10 glazes. That fluid melt may not only be trapping gases from the body but creating a perfect environment to trap all the bubbles coming out of those carbonates and talc. The Aero chocolate bar of glazes!
This is also a common problem at low fire on earthenware clay (but can also appear on a buff stonewares). Those white spots you see on the beetle also cover the entire glaze surface (although not visible). They are sites of gas escaping (from particles decomposing in the body). The spots likely percolate during soaking at top temperate. Some of them, notably on the almost vertical inner walls of this bowl, having not smoothed over during cool down.
What can you do? Use the highest possible bisque temperature, even cone 02 (make the glaze thixotropic so it will hang on to the denser body, see the link below about this). Adjust the glaze chemistry to melt later after gassing has finished (more zinc, less boron). Apply a thinner glaze layer (more thixotropy and lower specific gravity will enable a more even coverage with less thickness). Instead of soaking at temperature, drop 100 degrees and soak there instead (gassing is much less and the increasing viscosity of the melt overcomes the surface tension). Use a body not having any large particles that decompose (and gas) on firing. Use cones to verify the temperature your electronic controller reports.
Low-fire glazes must be able to pass the bubbles they and the underlying bodies generate (or clouds of micro-bubbles will turn them white). This cone 04 flow tester makes it evident that 3825B has a higher melt fluidity (A has not even dripped onto the tile). And its higher surface tension is demonstrated by how the flow meets the runway at a perpendicular angle (it is also full of entrained micro-bubbles). Notice that A, by contrast, meanders down the runway, a broad, flat and relatively clear river. Low-fire glazes must pass many more bubbles than their high-temperature counterparts, the low surface tension of A aids that. A is Amaco LG-10. B is Crysanthos SG213 (Spectrum 700 behaves similarly, although flowing less). These two represent very different chemistry approaches to making a clear glaze. Which is better? Both have advantages and disadvantages.
Rutile blue glazes are difficult, blistering and pinholing are very common. You must get it right on the first firing because pinholes and blisters will likely invade on the second. On the second firing the melt fluidity increases, the glaze runs and creates thicker sections in which the bubbles percolate and just do not heal well during cooling (even if it is slow). When finishing leather hard or dried ware do not disturb thrown surfaces any more than necessary. Make sure that ware is dry before the glaze firing. Do not put the glaze on too thick. Limit the melt fluidity (so it does not pool too thickly in any section). Do not fire too high. Drop and hold firing schedules can help a lot (coupled with a slow cool if needed).
These are 10 gram balls of four different common cone 6 clear glazes fired to 1800F (bisque temperature). How dense are they? I measured the porosity (by weighing, soaking, weighing again): G2934 cone 6 matte - 21%. G2926B cone 6 glossy - 0%. G2916F cone 6 glossy - 8%. G1215U cone 6 low expansion glossy - 2%. The implications: G2926B is already sealing the surface at 1800F. If the gases of decomposing organics in the body have not been fully expelled, how are they going to get through it? Pressure will build and as soon as the glaze is fluid enough, they will enter it en masse. Or, they will concentrate at discontinuities and defects in the surface and create pinholes and blisters. Clearly, ware needs to be bisque fired higher than 1800F.
The clay is terra cotta. It is volatile between 04 and 02, becoming dramatically more dense and darker colored. The electronic controller was set for cone 03, but it has fired flat (right-most cone), that means the kiln went to 02 or higher. At 02 the body has around 1% porosity, pushing into the range where the terra cotta begins to decompose (which means lots of gas expulsion). The outsides of some pieces blistered badly while the insides were perfect (because they got the extra radiant heat). The firing was slow, seven hours. Too long, better to make sure ware is dry and fire fast. Another problem: The potter was in a rush and the pieces were not dry. These factors taken together spelled disaster.
The glaze is running down on the inside, so it has a high melt fluidity. "High melt fluidity" is another way of saying that it is being overfired to get the visual effect. It is percolating at top temperature (during the temperature-hold period), forming bubbles. There is enough surface tension to maintain them all the way down to the body, and for as long as the temperature is held. To break the bubbles and heal up after them the kiln needs to be cooled to a point where decreasing melt fluidity can overcome the surface tension. Where is that? Only experimentation will demonstrate, try dropping a little more (e.g. 25 degrees) over a series of firings to find a sweet spot. The hold temperature needs to be high enough that the glaze is still fluid enough to run in and heal the residual craters. A typical drop temperature is 100F.
The buff stoneware mug on the right was bisque fired at cone 02, the one on the left at cone 06. The cone 02 mug was immersed in the clear glaze for 1 second and allowed to dry. The other was glazed on the inside first, allowed to dry, then glazed on the outside with a 1 second dip. Of course, the cone 02 one took longer to dry. In spite of this, the glaze is thicker and more even on the one bisque fired to cone 02. How is the possible? The secret is the thixotropy of the glaze. When that is right, a one second dip will give the same thickness and evenness whether dry or bisque, 06 or 02. Why bisque fire to cone 02? To get a glazed surface free of pinholes on some stoneware clays.
When electric kilns, especially large ones are tightly packed with heavy ware, the shady or undersides of the pots simply will never reach the temperature of the element side, no matter how long you soak. In this example, the inside of this clear glazed cone 6 bowl has a flawless surface. The base is pinholed and crawling a little and the surface of one side (the shady side), the remnants of healing disruptions in the melt (from escaping gases) have not smoothed over. The element side is largely flawless like the inside, however it is not as smooth on the area immediately outside the foot (because this is less element-facing). Industrial gas kilns have draft and subject ware to heat-work by convection, so all sides are much more evenly matured.
An example of how calcium carbonate can cause blistering as it decomposes during firing. This is a cone 6 Ferro Frit 3249 based transparent (G2867) with 15% calcium carbonate added (there is no blistering without it). Calcium carbonate has a very high loss on ignition (LOI) and for this glaze, the gases of its decomposition are coming out at the wrong time. While there likely exists a firing schedule that takes this into account and could mature it to a perfect surface, the glaze is high in MgO, it has a high surface tension. That is likely enabling bubbles to form and hold better.
An example of how a carbonate can cause blistering. Carbonates produce gases during decomposition. This glaze (G2415B) contains 10% lithium carbonate, which likely pushes the initial melting temperature down toward the most active decomposition temperatures.
First, the layer is very thick. Second, the body was only bisque fired to cone 06 and it is a raw brown burning stoneware with lots of coarser particles that generate gases as they are heated. Third, the glaze contains zircopax, it stiffens the melt and makes it less able to heal disruptions in the surface. Fourth, the glaze is high in B2O3, so it starts melting early (around 1450F) and seals the surface so the gases must bubble up through. Fifth, the firing was soaked at the end rather than dropping the temperature a little first (e.g. 100F) and soaking there instead.
A closeup of a cone 10R rutile blue (it is highlighted in the video: A Broken Glaze Meets Insight-Live and a Magic Material). Beautiful glazes like this, especially rutile blues, often have serious issues (like blistering, crazing), but they can be fixed.
This is Polar Ice cone 6 porcelain that has been over fired. The electric kiln was set to do its standard cone 6 fast fire schedule, but a cone in the kiln demonstrates that it fired much higher (perhaps to cone 7 judging by the bend on the cone). This is a translucent frit-fluxed porcelain that demands accurate firing, the over fire has produced tiny bubbles and surface dimples in the glaze. The mug rim has also warped to oval shape. The lesson: If you are firing ware that is sensitive to schedule or temperature, use large cones and adjust if needed. If it fires too hot like this, then program to fire to cone 5 with a longer soak, or cone 5.5 (if possible). Or, program all the steps yourself; that is definitely our preference.
This chart compares the decompositional off-gassing (Loss on Ignition) behavior of six materials used in ceramic glazes as they are heated through the range 500-1700F. It is amazing that some can lose 40%, or even 50% of their weight on firing. For example, 100 grams of calcium carbonate will generate 45 grams of CO2! This chart is a reminder that some late gassers overlap early melters. That is a problem. The LOI (% weight loss) of these materials can affect glazes (causing bubbles, blisters, pinholes, crawling). Notice talc: It is not finished gassing until 1650F, yet many glazes have already begun melting by then (especially fritted ones). Even Gerstley Borate, a raw material, is beginning to melt while talc is barely finished gassing. And, there are lots of others that also create gases as they decompose during glaze melting (e.g. clays, carbonates, dioxides).
There is a direct relationship between the way ceramic glazes fire and their chemistry. These green panels in my Insight-live account compare two glaze recipes: A glossy and matte. Grasping their simple chemistry mechanisms is a first step to getting control of your glazes. To fixing problems like crazing, blistering, pinholing, settling, gelling, clouding, leaching, crawling, marking, scratching, powdering. To substituting frits or incorporating available, better or cheaper materials while maintaining the same chemistry. To adjusting melting temperature, gloss, surface character, color. And identifying weaknesses in glazes to avoid problems. And to creating and optimizing base glazes to work with difficult colors or stains and for special effects dependent on opacification, crystallization or variegation. And even to creating glazes from scratch and using your own native materials in the highest possible percentage.
An extreme example of blistering in a piece fired at cone 03. The glaze is Ferro Frits 3195 and 3110 with 15% ball clay applied to a bisque piece. Is LOI the issue? No, this glaze has a low LOI. Low bisque? No, it was bisqued at cone 04. Thick glaze layer? Yes, partly. Holding the firing longer at temperature? No, I could hold this all night and the glaze would just percolate the whole time. Slow cooling? Close, but not quite. The secret I found to fix this was to apply the glaze in a thinner layer and drop-and-hold the temperature for 30 minutes at 100F below cone 03. Doing that increased the viscosity of the glaze melt to the point that it could break the blisters (held by surface tension) while still being fluid enough to smooth out the surface.
A Broken Glaze Meets Insight-Live and a Magic Material
Use Insight-Live.com to do major surgery on a feldspar saturated cone 10R glaze recipe with multiple issues: blistering, pinholing, crazing, settling, dusting and possibly leaching!
Manually program your kiln or suffer glaze defects!
To do a drop-and-hold firing you must manually program your kiln controller. It is the secret to surfaces without pinholes and blisters.
In ceramics, surface tension is discussed in two contexts: The glaze melt and the glaze suspension. In both, the quality of the glaze surface is impacted.
Blistering is a common surface defect that occurs with ceramic glazes. The problem emerges from the kiln and can occur erratically in production. And be difficult to solve.
Ceramic glazes melt and flow according to their chemistry, particle size and mineralogy. Observing and measuring the nature and amount of flow is important in understanding them.
Cone 6 Drop-and-Soak Firing Schedule
350F/hr to 2100F, 108/hr to 2200, hold 10 minutes, freefall to 2100, hold 30 minutes, free fall
Glaze Pinholes, Pitting
Analyze the causes of ceramic glaze pinholing and pitting so your fix is dealing with the real issues, not a symptom.
Ceramic Glazes Today
Todd Barson of Ferro Corp. overviews the glaze formulations being using in various ceramics industry sectors. He discusses fast fire, glaze materials, development and trouble shooting.
|By Tony Hansen|
Follow me on