Terra Cotta is often associated with indigenous cultures or school art programs. But actually, a quality terra cotta with a good glaze can produce ware of excellent durability.
Casting strong and durable terra cotta at cone 04
Terra cotta is being discovered by many potters (and hobbyists with interest in production) as being capable of much more strength and durability than they thought. Terra cotta clays fire much stronger than white-burning clays. Why? Because they are impure. But iron is not the only impurity, they are also loaded with K2O, Na2O, CaO, MgO, etc. These are all fluxes, they make the clay vitrify at a much lower temperature. These impurities are what make a terra cotta a potential stoneware that fires ten cones lower!
Low firing is so much easier and faster than for stoneware. And the colors available are much brighter than at high temperatures. It seems there are thousands of commercial glaze products for use at cone 06-04 (most can be fired considerably higher with no issues). And it is just as easy to mix your own glazes. We are working on a white engobe to fit this also. "Fit" means it has the same firing shrinkage at the target temperature.
Plainsman Clays makes a body named L215, we are attempting to make a casting body that fires to similar colors, of similar maturity and working with the same glazes. The L215 data sheet points people who want to do casting to this page.
But there are two big obstacles for those interested in production: The expense of the jars of glaze and the time-consuming process of painting them on. What if you could mix buckets of your own base clear and white glazes that rival or exceed the quality of commercial products? And dip bisque ware to get even coverage and quick drying? And at a fraction of the cost. Then use commercial paint-on products to decorate the ware? Then later, if you ramp up your production, you could buy stains and make your own decorating products.
Note that this page, unlike anything else you will find online, has the data. The charts at the bottoms of the columns show our measurements and the calculated fired shrinkages and porosities that come from them. These numbers indicate how vitrified a body is. They are important in matching the engobe. There are pictures to indicate what works and does not work. Pictures from firings at many different temperatures. To make terra cotta work as a functional body will not happen by accident, you will need to pay attention to technical issues much more than with stoneware. But the rewards are worth it!
*Note: This recipe is superseded by L4170B, a 90:10 blend of Redart/Ball Clay to get faster casting and adapt it as a plastic body if needed. However the notes below are still applicable.
I have tried many terra cotta casting recipes in the past but decided to start over with the simplest recipe I could think of: 100% RedArt! RedArt is a fabulous red-burning terra cotta widely available in North America, it is redder than any other clay I know and has been consistent for decades. And it matures at a very low temperature. As it turns out, it is not plastic enough to have the strength to pull away from the mold. So I added 20% ball clay. The result is something amazing! I can cast 3 mm thick in 10 minutes and have the piece out of the mold in another 10. A dry in front of the fan in another 20. The KT1-4 ball clay is not slowing down casting, yet adds so much to the leather hard handling strength and dry strength. If you do not have a large particle ball clay and have to use OM4, for example, then consider using a 90:10 ratio instead (otherwise casting might be too slow).
So far, this has proven to work well with my G1916Q clear glaze. I have used a simple firing schedule that ramps up to 1830 quickly, then slows down to 108F/hr to 1920, holds for 10 minutes, then shuts off (you would need to use a cone in your kiln to discover at what temperature it reaches cone 04). I will graduate this to a drop-and-hold firing schedule if it becomes necessary. I will also be testing more at cone 03. For thinner ware there is a tendency for some shapes to warp, so that is a signal that the body is firing to good strength there. But I am suspicious that cone 04 will be best production temperature for me.
The fired strength of this body at cone 04 is excellent. It is not a stoneware, but it far stronger than a white-firing body of the same temperature range. I think there would be very few utilitarian tasks this combination would not be up to the challenge for. And the warm red color is outstanding. The potential for volume production of ware with this body, the G3879 glaze and cone 04 firing is tremendous.
This recipe very easy to mix, I just added the Darvan to the water, poured in all the powder and it mixed in 20 seconds. Read the paragraph below, this recipe is a suggestion. "Listen to your slip" rather than this page if it is not deflocculating right.
2000g of powder produces about 1.8 liters. It is best to mix the slurry to a workable consistency first (not too thin). Then work with it until it gels and remix, adding a little more Darvan. You may need to do this on several occasions before it stabilizes. I used Darvan 811, it is supposed to be better for high iron slurries (to prevent gelling), but I am not convinced that it actually works better than regular Darvan.
G3879 transparent at cone 1 on L4170 terra cotta
This combination qualifies as a stoneware, the body has less that 2% porosity at this temperature.
G3879 Clear on L4170 TerraCotta Casting
The clear glaze is G3879. The white on the outside of the one on the left has 10% added zircopax. The overglaze colors are Spectrum Majolica colors.
L4170 TerraCotta with G3879 zircon white
The insides of both are done with a 10% zircopax-added version of G3879. I compared it with Spectrum Majolica white (on the outside of the one on the right, it had to be painted on), the potential to get a much more even coverage is there because we can dip-glaze this one.
The outside color tests on the left are Crysanthos Underglazes.
L4170 at cone 06, 02, 1 (bottom to top)
Cone 1 is stoneware strength. Cone 02 is the warmest color. But cone 04 (not shown) is also pretty good.
Fired Redart, Tuckers, M2 bars cone 02, 04
Top to bottom:
L4170 Redart casting body
L4183 Tuckers terra cotta
L4184 M2 throwing body
*This reduces the amount of ball clay in L4170 to speed up casting time. The casting behaviour is still good and release is quicker. This body is a real testament to the quality of Redart as a material.
We are testing a variety of glazes and underglazes on this with good success. We have found that typical low temperature glazes are working just fine to cone 2.
This is producing an amazing stoneware starting around cone 02. To make a plastic version we are adding 3% bentonite. It is much more thermally stable than typical terra cotta bodies. It does not bloat on over firing and fires to porcelain density by cone 4, producing remarkable strength. Most terra cottas turn an ugly brown when fired toward maturity, but with this one the color gets even richer. Remarkable. By cone 5 it is continuing to densify (reaching zero porosity), so that means it will survive to cone 6 and possibly beyond. We have never seen a terra cotta that can do this. This actually makes it look possible that one could make stoneware at cone 02 using standard commercial low fire glazes!
At a 1.79 S.G. this recipe will produce about 4 litres of usable slip (from a 5000 gram batch of powder).
Joe: Jan. 4/21
After 10 minutes mixing time.
S.G. was 1.78
Ford Cup was 58 seconds.
Next day: Ford Cup increased to 1:15 and 1:20. After 10 minutes in mold, slip surface began to cloud over at about 8 minute mark, but not total distortion before pouring out. Bowl released within 2 hours and had a nice smooth surface in and out. Cutting lip of bowl displayed minimal tearing of surface, this is good.
Joe: Repeat of above mix Feb.8/21
Ford Cup 1:02 after 15 minutes mixing time. Checked bucket 2 days later, and no water sitting on surface, mixed for abut 20 minutes and ford cup reading was 2:02, so has more or less doubled.
Casting wonderful as usual, a 25 minute cast seems to be quite nice.
Cast 2 mugs for 25 minutes and added handles made from L215 P6877 run about 2 -3 hours later and used slip to join handles to mugs. This slip adheres like super glue. Mugs dried slowly over the weekend with plastic/rags and there is no indication of cracking at any point where handle joins mug. This casting body also seems to be quite durable with high green strength?
Joe: Feb. 22/21 doubled recipe to 10 KG. powder quantity, mixed briefly and let sit overnight. Next morning no water sitting on surface of slip in 20 litre pail, so, even more surface area than approx. 7 litre Canadian Tire Bucket. This amount mixed quite well with lab mixer, but any more and the large mixer would be required, and was used for initial mixing first day.
After mixing approx. 15 minutes S.G. 1.79 and Ford Cup 2:01 so cast two terra cotta molds and slip is beginning to show light distortion on surface at 12 minute mark. Cast pieces 25 minutes and ware was out of molds within 75 minutes.
After casting about 2 litres of slip over the course of a week at a rate of 25 minutes in planter molds, checked S.G. once again and is constant 1.79 but ford cup would barely not empty at 2 minute mark so added 3 grams darvan to remaining 6 litres of slip and ford cup reading is now 2:07, but slip is slightly warm so this might be aiding in flow. Cast once again at these readings.
April 20/21 Checked slip and ford cup would not empty at 3 minute mark so added 3 grams Darvan and ford cup barely did not empty at 2 minute mark so left here as reluctant to add too much Darvan.
April 21/21 Joe:
Powder 5450 GM. (68.14%)
Water 2500 GM. (31.26 %)
Darvan #7 48 GM. (.60%)
S.G. 1.780 and Ford Cup 50 seconds.
Trying a slightly looser or thinner slip.
L4170B bars fired cone 06-8 (bottom to top)
No sign of bloating even at cone 6. But it is clearly melting by cone 8. The front-sides of the cone 5-7 bars have some solubles salts that are glossing edges.
G3879 transparent at cone 1 on L4170 terra cotta
This combination qualifies as a stoneware, the body has less that 2% porosity at this temperature.
Cast L4170B pieces fired at cone 04, 2
Very dense and strong at cone 2, it’s fired surface is almost velvety, very pleasant to touch.
L4170B terra cotta at cone 02 with G1916Q
Very strong piece, no sign of crazing a week later. It appears as pieces are fired to cone 02 or more, crazing ceases to be an issue.
G1916Q on L4170 cast pieces
G1916Q on L4170
Thin application is clearly the way to get the best transparent.
Handles are L215 P6877 pugged clay
Handles are L215 P6877 pugged clay attached to mugs with the L4170B slip which acts like glue, a minimal amount required to attach handles. No cracking evident anywhere around handles.
L4170B Terra cotta fired at cone 3 with G1916Q
The glaze is clouding and micro-bubbling but the body is showing no signs of bloating.
L4170B terra cotta fired to cone 4
Glaze is G1916Q. Cloudiness is worse that cone 3. This is clearly too high a temperate for glazed ware.
L4410G, L4170B cone 2 spectrum low fire glazes
Spectrum low temperature 753 yellow, 754 orange glazes are bubbling at cone 2.
*This is the product of a long development program that culminated in the realization that a super-white engobe requires using super-white kaolins and bentonites and plenty of zircopax. And that it is possible to dip and paint engobes on bisque if the slurry is gelled and conditioned with two gums. This eliminates the issues that arise with application to leather or dry hard ware, especially tiles.
This recipe looks so good because it promises to enable the strength of terra cotta with the whiteness of porcelain. This, in turn, promises to provide white surfaces for brush work and transparent overglazes that produce brilliant gloss and colors not achievable at high temperatures.
It appears that to be paintable on porous bisque a low specific gravity is required, that means lots of water, almost double the normal amount (about 100 water to 50 powder)! How to gel that watery slurry into a creamy consistency? 3% Veegum! To make it bond well with the bisque: CMC gum. These gums can of course be tuned to thicken or thin the slurry and make it dry faster or slower.
Mixing this is more difficult because both the Zircopax and Veegum want to agglomerate the slurry. You must mix the powder well first, shake it throughly in a bag. Then mix it into the water using a high speed mixer. It is absolutely necessary to sieve it (preferably at 80 mesh). As noted, just use the amount of water needed to get the viscosity desired. Use almost all the water needed when propeller-mixing and sieving, then add more if needed. You may even need to propeller-mix the sieved slurry to remove lumps again.
Having this high water content means that complete immersion dipping of thin-walled bisque ware will water-log it. I have found that pour-application on the insides of pieces works extremely well, the engobe drains to a perfectly even thickness and drip dries enough to stop runs fairly quickly. Pour glazing tile works equally well.
Engobe bonding with bisque depends a lot on matching the firing shrinkage with the body. Unlike any other engobe recipe, below you will find data on the fired shrinkage of this at ten temperatures! How do you use that data? An example: To fire at cone 2 the body needs to have around 3.6% fired shrinkage. We retested this engobe with 5% frit 3110, L3685Z4, that increased maturity more than expected. A 2.5% fritted version turned out to be good for the L4170B terra cotta.
First batch I mixed this 900:500 water:powder to get about 65% H2O. Second batch, 950:500.
Cone 1-6 (bottom to top), cone 01, 02 at top
L3685Z1 and L3685Z2 at cone 04
Z2, this improved version, is much whiter. On Plainsman L210. Fit is amazing for Z2.
And it is incredibly white.
L4170B with L3685Z2 engobe
With the gum additions, it applied very evenly to the bisque (using a pour-in pour-our technique). The layer was very thin yet covers very well.
Cone 04 L3685Z2 white engobe with clear glazes
The colour as amazingly white. And this engobe appears to be fitting very well even though we are dipping pieces, like a glaze, to bisque ware.
L215 with L3685Z2 engobe at cone 04
Although a very thin layer it completely covers the dark red body.
*This recipe provides the greatest thermal expansion adjustability we have seen in a low fire clear glaze. It combines a middle-of-the-road thermal expansion frit with a very low and very high expansion frit (they cancel each other out but increase gloss of the otherwise silky Frit 3195).
This melts well from cone 05 and is stable to cone 2 or higher.
Do not assume the glaze fits because a piece emerges from the kiln without crazing or shivering. Thermally stress it (by two minute boiling-water and ice-water immersion cycles). We refer to this as the 65:10:10:15 recipe.
If this crazes, try 65:20:15 (3195:3249:EPK).
If it shivers, use 65:20:15 (3195:3110:EPK).
The ball clay also imparts good working properties and it has a lower LOI than EPK (to generate less micro-bubbles in the glaze).
The 2% iron oxide is only needed on dark burning bodies. It enhances the red color of terra cotta (when used as a transparent) and acts as a fining agent to remove the microbubbles they produce. If you are adding stains or using the glaze on a white burning body, remove the iron.
We found about 3800 water for 4000 water to get 1.44SG. The slurry gels more than with EPK.
G1916Q at Cone 01 on 3D+iron
Very nice results on L3724E red body at cone 01. Piece is very strong.
1916J and Q fired to 1450F
These glazes are starting to melt, the surface having reached the consistency of a porcelain and have densitfied to very low porosity. Notice the iron in the ball clay really shows up at this stage (it will be less evident later).
G1916Q Cone 04 using 04DSSC schedule
Counterclockwise: L212, Raku, Buffstone, L213, L210, L215
Crazing out-of-the-kiln on Raku and buffstone.
G1916Q and J fired 1650-2000F
Ten-gram balls melted and flattened as they fired. They soften over a wide range, starting well below cone 010! At 1650F 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, this is a thick layer, so the bubbles are expected. But they still can be an issue 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 (1450F in this case), then soak the kiln just below that (on the way up) to fire away as much of the carbon as possible.
Success with cyrstal clear glaze cone 03
Uses Cone 03 soak-soak-slow cool schedule.
Left: P6282 with 3685U slip and 1916Q. Clear and very good. Glaze is thicker than the other two. Shivering on lip, the slip is not fitting the body.
Center: P6282 with G2931F Ulexite clear. Better clarity even though it is applied very thin. Shivering on lip, the slip is not fitting the body.
Right: L3724F with 3685U slip. No shivering. Very good coverage of the glaze, very clear, the best I have seen yet!
G1916Q on L215, L212, L210, L213, Buffstone at cone 03
All exited from the klin without crazing. The L215, L213, L210 and L212 samples subsequently survived a 300F/Icewater test without crazing, but the Buffstone did not. The L213 would not likely survive a cold-to-hot test without shivering.
1916Q cone 04, 03
Both were slow cooled. While the cone 04 version is glassy and ultra-gloss, it has significant clouding of micro-bubbles. The cone 03 version, right, is completely transparent.
G1916Q on L210 fired at cone 04
G1916Q+2%Iron on L212 talc body fired at cone 05
This will likely shiver over time. But the speckle that happens on white bodies is clearly visible.