Glaze clouding is a universal issue in ceramics. Terra cotta bodies demonstrate this best. Pretty well all transparent glazes, even commercially available ones, can cloud. This example is G2931K, it can be beautifully crystal clear. But the thickness of application is the key to achieving that (as thickness increases this happens). We ball milled it to see if that would help, but as you can see, that has not impacted the problem. This is a dipping version so that is part of the reason why it is easy to get it on too thick. One of the advantages of brushing glazes is the ability to carefully control thickness,
This is a buff stoneware body, Plainsman M340. A L3954F black engobe was applied inside and upper outside at leather hard. The piece was fired at cone 6 using the PLC6DS schedule. The inside, totally clouded glaze is G2926B. Outside is GA6-B Alberta Slip amber clear. Normally that inside glaze is crystal-clear on other bodies, but this black engobe is generating tiny gas bubbles at the exact wrong time during the firing and the melt is unable to pass them. The black stain thus seems implicated as the gas generator or catalyst. The outside glaze, although amber rather than completely transparent, demonstrates here its ability to clear micro-bubble clouding in spite of this issue.
At the leather hard stage the sides of these two L4410P low temperature dolomite body pieces were coated with AMACO velvet underglazes. Both were bisque fired and finished with a layer with the same transparent glaze. But the difference is the thickness of that glaze and the method of application: The one on the left got three thin layers of a brushing glaze. The one on the right was quickly dipped in a base coat version of that same glaze. Evidently there is a thickness threshold, which, when exceeded results in clouding. We have observed that this happens with pretty well any clear glaze.
These are fired in cone 6 oxidation. They are all the same clay body (Plainsman M390). The center mug is clear-glazed with G2926B (and is full of bubble clouds). This dark body is exposed inside and out (the other two mugs have a white engobe inside and midway down the outside). G2926B clear glaze is an early-melter (starting around cone 02) so it is susceptible to dark-burning bodies that generate more gases of decomposition - they produce the micro-bubble clouding. That being said, the other two glazes here are also early melters, yet they did not bubble. Left: G2926B plus 4% iron oxide. That turns it into an amber color but the iron particles vacuum up the bubbles! Right: Alberta Slip GA6-A using Ferro Frit 3195 as the melter. It also fires as an amber-coloured glass, but on a dark body, this is an asset.
The mug on the left is a brush-on version of a boron-based clear glaze for cone 05. Three coats were applied and the often-encountered clouding occurred. The one on the right is an 85:15 lead bisilicate:kaolin mix. Three coats were also applied. It is an absolutely "knock your socks off" crystal-clear hyper-glossy surface that transmits the terra cotta color beautifully. And my lead testing kit passes it with no detectable lead release. Underglaze brushwork here we come! I have sought this effect for decades, this is it! Recent realizations about the slipware tradition in the UK (and their standard use of this same glaze) motivated me to get some of the frit. All I could get was a sample of the frit shards, but these milled down to a powder easily in our ball mill.
Of course, to be safe, I would still glaze the insides of pieces with a boron clear, likely as thin a layer as possible of G1916Q (with 2% added iron as a fining agent for the micro-bubbles). And, I will obviously fire these lead glazed pieces with the kiln exhaust system turned on.
Clouding in Ceramic Glazes
There a many factors to deal with in your ceramic process to achieve transparent glazes that actually fire to a crystal-clear glass
Many ceramic glaze benefits and issues are closely related to the thickness with which the glaze is applied. Many glazes are very sensitive to thickness, so control is needed.