Ceramic glazes, like this GA6-B, are actually just glass. But they are not like bottle glass. The latter is formulated to work well in forming machines (harden quickly), melt and stiffen quickly, have low melt viscosity and resist milkiness and crystallization on solidification. The chemistries to accomplish this have adequate resistance to leaching and adequate durability for a few uses. A stoneware glaze melt needs to be much more viscous (to stay put on vertical surfaces). And, it must have a lower thermal expansion (to match common clay bodies). And, it must resist crystallization much more (since it cools slowly). Fortunately, meeting these needs brings along big benefits: Greater durability, hardness and resistance to leaching. Stoneware glazes and bottle glass share a common trait: They have about the same amount of SiO2. But the similarity ends there, stoneware glazes have:

-High Al2O3. Three to five times more! It is the key oxide to producing durable glass. And it stiffens the melt (that disqualifies high levels from bottle glass).
-The same fluxes (CaO, MgO, K2O, Na2O). But they distribute very differently (half the CaO, half to one third the KNaO, much more MgO). Other fluxes like SrO, Li2O are also common.
-Low KNaO (which they call R2O). In glazes, it produces crazing, 5% is a typical maximum. But bottle glass can have double or triple that (the high thermal expansion is not an issue, and its cheap source materials supply lots of melting power).
-B2O3 melter. It is expensive but can be justified because the glaze is just a thin layer. Glazes at the low end of the stoneware range have 5% or more boron.

Far right: A glass bottle. Left: Small test bottles made from dark and light burning stonewares. Third: A production ceramic bottle. Notice how much the dark body darkens the GA6-B glaze.

These videos from Eastfork Pottery demonstrate their use of thixotropic glaze slurries. Watch them to see how effective a highly gelled glaze is. It enables a quick dip, stays fluid while draining, gives even coverage and dries in seconds. These don't hard-pan or settle out in the bucket either. They work on porous or dense bisque. Almost any glaze can be thixotropic if you take the time to learn how to do it. The fast drying enables the use of twin running (or twin belt) foot wiper machines (best shown on these Instagram and Facebook videos).

This procedure of converting G1214M into G1214Z first appeared in the Digitalfire desktop Insight instruction manual 30 years ago. The process is beyond interesting if you want to know more about glazes, recipes and their chemistry. Watch me adjust the recipe of my high-calcium transparent cone 6 glaze to convert it to a calcium matte. In an Insight-live.com account, the process is easy enough for anyone. We'll cut the Si:Al ratio, increase the CaO, maintain the thermal expansion for glaze fit and make the recipe shrinkage-adjustable using a mix of calcined kaolin and raw kaolin. Watch the G1214Z video to see it all.

If your pottery glaze is doing on drying then it will crawl during firing. Wash it off, dry the ware. Then check the water content. If the glaze has worked fine in the past then it is likely going on too thick because the specific gravity is too high - just repeat cycles of adding a little water and dip testing (make it thixotropic if needed). But that was not the issue here. Glazes need clay to suspend and harden them, but too much clay means trouble. This was Ravenscrag Slip, a clay, being used pure as a cone 10R glaze. The glaze appeared to go in perfectly and it dried to the touch in ~20 seconds. But shrinkage continues after that, revealing after a couple of minutes. Fixing the issue was a matter of adding some roasted Ravencrag Slip to the bucket. That reduced the shrinkage and therefore the cracking. Any glaze containing excessive kaolin can be fixed the same way (trade some of the raw kaolin for calcined kaolin). Some glazes that contain plenty of clay also have bentonite - a simple fix for these is to simply remove the bentonite.

Although Nepheline Syenite and Custer Feldspar are used as effective body maturing agents and fluxes in glazes past cone 6, curiously, neither of them melt well by themselves. Thus, both of these come 6 melt fluidity tests add 20% Ferro Frit 3134 to get them flowing. This is a 2021 shipment of the feldspar and a 2022 shipment of the nepheline.

Most low-fire bodies contain talc. It is added for the express purpose of increasing thermal expansion. The natural quartz particles present do the same. These are good for glaze fit but bad for ware like this. There are also sudden volume changes associated with cristobalite, but it forms (from quartz) at stoneware temperatures so should not be a concern in terra cotta or a white low fire body. You could fiddle with the clay recipe or change bodies, but better to change the firing schedule. The quartz in stonewares goes through a sudden volume change between 950-1150F on the way down. Quartz particles in low fire bodies will do the same. A simple fix is to slow down the entire cooling cycle like this potter did. Or, learn to program your kiln to approach this range more slowly, then ease down through it. No electronic controller? Learn a switch-setting-schedule to approximate this down-ramp (buy a pyrometer if needed).

Potters love plastic clay. On the wheel it enables pulling larger, more overhang, thinner walled pieces. For beginners it can make the difference between success or a collapsed lump of mud. The downside is high drying shrinkage and danger of cracking. But potters know how to exercise care in drying to get success anyway.

This industrial jiggering machine has the opposite priority: Ability to hold shape immediately after forming and to dry crack-free quickly. The secret is low plasticity stiff clay (notice how it splits around the edges when flattened). Notice, in the video, how much water is used yet it does not stick to the heated metal mold. Note also how the machine avoids tearing it by applying pressure slowly right to the end. Even then, the vertical splitting on the outer belly and the crumbly way it cuts verify its poor plasticity.

A restoration project faced a tile-matching challenge. At installation in a bathroom 90 years ago, the tiles were not crazed. But between then and now it happened (shown inset upper right). Now, a restoration specialist is tasked with duplicating the aged effect (one unsuccessful attempt is shown here). The shade, opacity, degree of matteness, bubble-free matrix and surface character of the original are all real challenges. Duplicating the crazing is even more difficult. Why? Matching "time-crazing" with a crackle glaze pattern will be temporary (it will craze much more after installation).

The reason why functional mattes seldom craze can be seen in the chemistry. This chart compares the thermal expansions of the oxides that combine to form the fired glaze matrix. ~80%+ of the makeup of almost all common base glazes (without colorants, opacifiers) is SiO2 and Al2O3 (orange bars). Mattes almost always need a low Si:Al ratio (e.g. below 6:1). The rest is fluxing oxides to melt them (the blue bars + B2O3). Here is the problem with making a crazing matte: Almost all crazing is caused by high levels of K2O and/or Na2O (the top two bars on the graph). But they produce high gloss (as can be seen in this test tile). The main matting fluxes and agents are MgO, CaO, SrO, BaO; they have a low COE (and don't craze glazes). Further, both zircon and tin oxide, the opacifiers needed, also have low thermal expansions!

Other possibilities of making crazed matte:
-A matte glaze can have a high SiO2:Al2O3 ratio and craze if it is very melt fluid (containing lots of KNaO) and cooled slowly so that micro-crystals cover the surface. The downside is unpleasantness to the touch.
-Glossy glazes can be matted by the addition of micron-fine alumina (e.g. 800 mesh, this is done in the tile industry).
-A low expansion body with no ball clay or silica (e.g. just kaolin and feldspar with enough bentonite to get the needed plasticity) will craze most glazes. Adding pyrophyllite will further lower its COE.
-Print the lines on the tile (using ceramic transfers) and use a translucent matte glaze (like G2934).

Super white translucent porcelains are expensive, approaching $200/box in some countries! Even so, variation in properties is common (certainly not good for a "tipping point body" that is difficult to make). The idea of making your own clay body is actually feasible here. Yet Kirsty Kash, maker of this mug, told me that this amount of DIY was something she had never really considered, thinking it would be too complicated without guidance. But ongoing issues with the commercial clay gave her the motivation to give it a try (using a recipe similar to L3778D). She weighed out the materials, slurried up the mix using a propeller mixer, finished by blender mixing and then dewatered on a super-clean plaster bat.

By the third batch, something good happened: The porcelain blistered (tiny bumps on the surface). Subsequent correspondence brought the realization that this type of body is "walking a recipe tightrope" that requires control of the percentage of the key ingredient: Feldspar. It determines the maturity of the fired product. Too much brings blisters, too little and translucency is lost. Simple testing is all that is needed to determine the needed amount. More good news: The change enabled increasing the kaolin percentage by the same amount. That, in turn, enabled reducing the percentage of veegum (reducing the cost).

Her comment a few days ago was inspiring: "I'm getting to know my material so intimately. I have been learning SO much." When I suggested she might end up buying commercial again she responded: "I just bought large bags of all the materials and plan to keep going. I like having the control and being less reliant on the boxed clay. You've converted me!".

This is one of the things Gerstley Borate commonly does (when its percentage is high enough). It is also highly thixotropic - this can be stirred vigorously to thin it, yet within seconds it turns back to jelly again. This is part of the reason it is often referred to as “ghastly borate”. Side effects of this include high water content, slow drying, excessive shrinkage on drying, cracking and crawling. To attempt a fix, I deflocculated it with Darvan. It was stable enough to dip bisque ware, with difficulty (but pieces dried very slowly). But overnight it has turned back into a gel. What can be done with a mess like this? Start over, with three options:
1. Identify the mechanism to isolate the base recipe and substitute another.
2. Replace the Gerstley Borate with an equivalent that does not do this (Gillespie Borate).
3. Do a little chemistry to source the B2O3 from a frit instead (e.g. in an account at insight-live.com).

There is a lot of magic, Canadian magic, in this picture. Pretty well every single potter working at mid-temperature needs rutile blue, gloss black, honey amber and transparent glazes (even multiple versions of each). And almost all need a base slip (or engobe). Here they are.

Upper left: GA6-C and GA6-B on light and dark burning bodies.
Upper right: GR6-M and GA6-C on M340 (with black engobe L3954B).
Lower left: GA6-C and GA6-B on M340 (with black engobe).
Lower right: GR6-M, G3914A, G2926BL on slow and fast cooled mugs.

Every glaze company makes multiple variations of each of these, especially rutile blues (or floating blues). Unfortunately they often do not fit Plainsman Clays. But these do, in fact, they are adjustable (and better in other ways, as well as less expensive). Unfortunately, even though Plainsman Clays mines and makes most of the raw materials and gives out these recipes, it has not been making them, forcing customers to use the American-made products. This is a missed opportunity, even a responsibility. This will hopefully be rectified soon.

Fit? It has to stick well. And stay stuck during drying (and shrinking). And the bond has to survive shrinkage that happens during firing. This potter is doing thick applications of each slip (actually that makes them engobes). She uses stains, that's wise, metal oxides bring baggage when used to color slips (e.g. their decomposition can affect the bond, they can gel the slurry, flux the fired product thereby increasing the firing shrinkage of the slip). Stains are better because they affect slurry and fired properties less. But there are still enough issues that each colored slip deserves testing. This potter first slaked B-mix as a slip (it is highly plastic), using it at a runny yogurt consistency. But it bubbled when fired hotter than a cool cone 6. A switch to porcelain slip (which is non-plastic) is shown here. It flaked off as it dried (even in a damp box for 24 hours), also after bisquing to cone 08, and sometimes even after firing to cone 6. This signalled a drying mismatch between body and slip, the bond that managed to survive drying was weakened enough to fail on firing.

The solution was an engobe recipe that is super plastic and sticky. The popular Fish Sauce recipe is an example, it contains 10% bentonite and is unbelievably sticky. L3954B was formulated with this in mind. It adheres well to leather-hard clay and doesn't flake off (misfit is instead evident by surface cracking if its shrinks more than the body). And it is not highly vitreous, keeping its fired shrinkage low enough to match stoneware bodies. Mixing your own recipe also enables compensating the amount of feldspar if the stain affects the slip's degree of vitrification, and therefore fired shrinkage (e.g. blues, oranges, yellows).

These (right) were made individually in the factory during the 1930s and 1940s (the insides have pronounced throwing rings and slip drips). The potters were able to make up to 500 per day, even though they took the time to smooth the outside using a rib! The inside base of this one is bowl-shaped (the walls near the base are very thick), this helps explain how they were able to throw them so quickly.

Perhaps most surprising is how much whiter and speck-free the bottle is even though it is fired four cones higher than the crock (Plainsman M340 at cone 6). Both pieces have porosities above 2%. Why? First, they got their clay from further east in Saskatchewan (near Willows), where the cleanest clays are much lower in iron contamination (likely the H0009 body). The whiteness is better even though they would have had to add some ball clay to make the clays more wheel-throwable. Second, they employed a wet process to refine the clay (slaking, blunging, sieving and filter pressing), this enabled them to sieve out the iron pyrite particles. Fortunately, modern dry grinding and air separation equipment is greener and able to accomplish without water.

Notice also the transparent G1129 glaze on the beer bottle (the upper section is likely the same glaze stained using iron oxide): After almost 100 years it has not crazed. This is both a testament to the ease of glaze fit these natural materials offer (because of the high quartz content) and the skill of the engineers of the time at matching the thermal expansion of glaze and body.

3D print this, pour in plaster to make a slip casting mold! My previous work on this project assumed a smaller 3D printer (making it necessary to print flanged PLA mold sections that clip together). But larger 3D printers are now common, making the CAD work much easier. This OnShape drawing is parametric for height, body diameter, wall and plaster thickness, and neck height (for the full bottle set body=160mm, neck=96). This uses my standard natch system. Neck vertices are proportional to height, so resizing works well. The top end is filleted to permit the longest possible mold on the print bed (diagonally). The bottom inside perimeter is chamfered, strengthening the default 0.8mm side wall junction to the base (that being said, be careful not to flex it too much when removing it from the print bed).

Doing this smaller size is for prototyping and testing. Casting plaster on a 3D print creates artifacts; these were not an issue. This PLA mold prints quickly, it has a hollow back side, permitting easy removal with a heat gun. There is no spare, it employs a pour spout, making the mold shorter and producing a better lip.

Need a stoneware slip casting recipe? L4768E or L4768H are a good choice. A glaze recipe? How about GA6-B (or similar)? Go full DIY with this, you will never turn back.

This is the L3954B engobe. 15% Mason 6600 black body stain has been added (instead of the normal 10% Zircopax used for white). Of course, a cover glaze is needed for a functional surface. A lot of development work went into producing a recipe fits this body, M340. It works even when thickly applied because it has the same fired maturity as the body. Lots of information is available on using L3954B (including mixing and adjustment instructions). Engobes are tricky to use, follow the links below to learn more. L3954B is designed to work on regular Plainsman M340 (this piece), M390 and Coffee Clay. Most important, adjusting its maturity, and thus firing shrinkage, are documented. These bodies dry better than porcelains and are much less expensive, so coating them with an engobe to get a surface like this makes a lot of sense. Ed Phillipson discovered this 80 years ago, enabling selling pieces made from these clays as white hotel ware.