This recipe, G2826A, a base transparent recipe having 50% Gerstley Borate plus 20% kaolin, is "jelly city". Even with 2.5g of Darvan deflocculant in this jar it is still thick enough to require pushing this tile down into it! Even then, it needs 5 seconds to build up enough thickness. And then does not even cover properly. People have suffered with this popular fluid-melt recipe for 50 years or more just to get the surface variegation it produces. They add all manner of colorants and opacifiers to it. And endure its incessant running onto kiln shelves, bubbling and clouding. It is time to just stop this "Dr. Jekyll and Mr. Hyde" of ceramic materials! And use different base transparents that employ frits to source the boron (B2O3). Same chemistry, just a better recipe: G2826A1. Then just add rutile or titanium to restore the variegation. Are you a masochist and still want to be punished? Then at least use the G2826A2 recipe with Gillespie Borate instead.
G2931 Worthington Clear is a popular low to medium-fire transparent glaze recipe. It contains 55% Gerstley Borate (GB) plus 30% kaolin (GB melts at a very low temperature). GB is also very plastic, like a clay. I have thrown a pot from this glaze recipe! This explains why high Gerstley Borate glazes often dry so slowly and shrink and crack during drying. When recipes also contain a plastic clay like this one the shrinkage is even worse. GB is also slightly soluble, over time it gels glaze slurries even in smaller percentages. Countless potters struggle with Gerstley Borate recipes.
This is the G2826A 50:30:20 GB:kaolin:silica base clear recipe. It is been used for decades as a base for all kinds of glazes. It starts melting early enough for use on low-temperature earthenware and is widely used in the raku process. Yet it is also common at middle temperatures (obviously care must be taken or it will run off ware onto kiln shelves when fired to cone 5-6). These tests were fired to cone 6 using the PLC6DS schedule.
The samples on the left use Gerstley Borate, on the right Gillespie Borate. The GBMF test tiles (lower left and right) reveal how much off-gassing is still happening on both when melting starts (they are full of bubbles). The GLFL test (centre) shows the melt flow of the two glazes, it is very similar (normal glazes do not run off the end of the runway like this). The two porcelain test tiles show it to fire crystal clear (there is some pooling since these were applied too thick). There is thus good reason to believe that Gillespie Borate will work well in this class of recipes.
This is one of the things Gerstley Borate does to glazes (when its percentage is high enough). This slurry has a high water content and should be far more watery. It is also highly thixotropic - this can be stirred vigorously to thin it and yet within seconds it turns back to jelly again. This was deflocculated with Darvan yesterday and it was stable enough to dip bisque ware - but overnight salts have gone into solution and it is no longer dispersed. It also dries very slowly on bisque ware. What can be done with a mess like this? Replace the Gerstley Borate with something else. Gerstley Borate sources B2O3, it can be supplied using frits or Ulexite (glaze chemistry calculations are needed to juggle the recipe), this can be done in an account at insight-live.com.
These are various different terra cotta clays fired to cone 04 with a recipe I developed that sources the same chemistry as the popular G2931 Worthington clear (50:30:20 GB:Kaolin:Silica) but from a different set of materials. The key change was that instead of getting the B2O3 from Gerstley Borate I sourced it first from Ulexite (G2931B) and then from a mix of frits (G2931K). All pieces were fired with a drop-and-hold firing schedule C03DRH. Fit was good on many terra cottas I tried (pieces even surviving boiling:icewater stressing). Where it did not fit I had thermal expansion adjustability because more than one frit was sourcing the boron. Frits are so much better for sourcing B2O3 than Gerstley Borate (the latter is notorious for turning glaze slurries into jelly!). Of course, a little glaze chemistry is needed to figure out how to convert a recipe from Gerstley Borate bondage to frit freedom, but there is lots of information here on how to do that.
On the left is G2826A3, a cone 6 transparent glaze (an improvement on the 50:30:20 classic Gerstley Borate base transparent recipe substituting Gillespie Borate, reducing its percentage and increasing SiO2). Despite the improvements it exhibits this strange cracking and crawling. The G2826A1 on the right uses a frit to source the boron instead, clearly a better idea. These tiles were fired to 1700F. The problem is likely the ulexite mineral in the Gillespie Borate - it is known for this behavior of suddenly shrinking and then suddenly melting (the latter of which is just starting). Since Gillespie Borate is plastic and suspends slurries well we thought calcined kaolin would be better than raw kaolin in the G2826A3 recipe (to minimize drying shrinkage). However, it did not improve the situation. All of this being said, this recipe is still working reasonably well at cone 6 (likely stopping and holding it at 1700F exaggerates the problem).
Replacing the Gerstley Borate in recipes containing 50% or more of it
How to fix the 50:30:20 Gerstley Borate cone 6 base recipe
Gerstley Borate was a natural source of boron for ceramic glazes. It was plastic and melted clear at 1750F. Now we need to replace it. How?