|
|
|
8 June 2004
This article describes a simple line blend. The blend here was undertaken to arrive at a tinted raku crackle glaze. I was searching for a fiery orange colour through blending a red and a yellow stain, in a clear crackle base glaze.
A line blend is a method of blending two different substances in known proportions. Often one starts with 100% of substance A at one end and one progresses to 100% of substance B at the other end. In between are mixtures of both substances, with the proportion of one declining as the proportion of the other rises. The proportions are measured by weight.
A simple table shows a basic eleven mixture line blend.
| Blend # | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Substance A | 100 | 90 | 80 | 70 | 60 | 50 | 40 | 30 | 20 | 10 | 0 |
| Substance B | 0 | 10 | 20 | 30 | 40 | 50 | 60 | 70 | 80 | 90 | 100 |
In this example, a clear crackle base glaze is tinted with blends of two colorants. In practice, one could blend entire mixed glazes this way, or a material with a glaze, or a pair of materials.
The base glaze used is the standard 80/20 clear crackle.
| Material | Parts |
|---|---|
| Gerstley borate | 80 |
| Nepheline syenite | 20 |
The stains are Sherry's Western, #279496 Bright Red and C-416 Bright Yellow.
I was looking for a fiery orange tinted crackle. I chose to use Sherry's Western Ceramics stains and blend a yellow and a red stain together. I knew from earlier tests that somewhere around 10% stain would likely give enough intensity to the colour. So, in order to get 10% I needed to ensure that the sum of the colorants in each test sample was 10% of the raw material weight in the base glaze. You will find in doing tests like this, that it is easy to end up with a very small quantity of a material to measure. This can easily exceed the accuracy of the scale you have. Thus you must ensure that each batch is large enough so that you are not measuring very small amounts of colourants.
In this case, the initial intention was to use 75g of base glaze for each sample. This would need only 7.5g of colourant for each sample. Of course, this is not the smallest amount that would require weighing. In the 90/10, mixture 1/10th of 7.5g would need to be weighed out, 0.75g! Measurement inaccuracies would be unavoidable. Thus I settled on 125g of base glaze per sample, leading to 1.25g being the smallest measurement. This is likely still too small for real accuracy. Nonetheless, this is how the test was done and the results show a progression of colour from red to yellow. This would suggest that any inaccuracies were tolerably small. For a more representative test a larger batch would have to be made. Any fine tuning required could then be performed.
Below is the table of amounts to weigh out for each blend. Note also, that 100g of gerstley borate and 25g of nepheline syenite were weighed for each sample, as the base glaze.
| Blend # | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Bright Red (g) | 12.5 | 11.25 | 10 | 8.75 | 7.5 | 6.25 | 5 | 3.75 | 2.5 | 1.25 | 0 |
| Bright Yellow (g) | 0 | 1.25 | 2.5 | 3.75 | 5 | 6.25 | 7.5 | 8.75 | 10 | 11.25 | 12.5 |
Each blend, containing 100g of gerstley borate, 25g of nepheline syenite and 12.5g of colorants, is placed in a separate jar. Water is add and each is mixed thoroughly to the normal consistency of the raku clear crackle glaze. A test tile containing a space for each sample is then painted with each sample in order. The tile thus looks exactly like the table above. When fired one can place the sample along side the table and see exactly what each blend produced.
As this blend was aimed at deriving a glaze of specific colour, there was no need for a vertical tile. In other blends a vertical tile may be of use, particularly where flux and glaze blends are concerned.
As can be seen above, there is a clear progression of colour. At each end the sample colourant is 100% of either the red or yellow stain. There is slight variation in colour in each test due to thickness of application. I did not really get the fiery orange I was looking for, but I now know where to concentrate later tests. Somewhere in the 60 to 90 % yellow range would be good to start. These are the 5th through 2nd tests from the right.
It appears that 10% colour may have been too little. It would be worth trying up to 15% or maybe higher. It is also possible that a different stain would work better.
It is also interesting to note that the crackle is much finer in the 50 - 90% yellow range. There could be a number of factors here, but it would be worth checking further if the crackle size became too large or too small.
As this test was primarily about the colour, it was performed on a white stoneware. It would be worthwhile to do tests on the clay body which will be used in production, to see if the colour intensity is sufficient.
As has already been mentioned, there is a need to increase the intensity, and to test on another body. It is also true that the exact colour desired has not been achieved. Other stains might be tried instead of increasing the 10% concentration.
As already mentioned, it may be useful in this case to explore blends between 50 and 90 % yellow, in greater detail. One possible blend for doing this is presented below.
| Blend # | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
|---|---|---|---|---|---|---|---|---|---|
| Bright Red (g) | 9.375 | 8.4375 | 7.5 | 6.5625 | 5.625 | 4.6875 | 3.75 | 2.8125 | 1.875 |
| Bright Yellow (g) | 9.375 | 10.3125 | 11.25 | 12.1875 | 13.125 | 14.0625 | 15 | 15.9375 | 16.875 |
In this blend the colourant proportion has been raised to 15 % and, finer variations have been added between 50 and 90 % yellow. Other proportions are simply not included.
This table demonstrates several things worth noting. First, there is no sense in 4 decimal places for weighing grams. The kinds of scales we use simply do not offer that accuracy. Secondly, the increments between colourant weights is now very low. A larger batch size is therefore required, if the tests are to accurately show the effect of such small differences. 250g of the base glaze or even more would be helpful. Thirdly, a line blend does not require an specific number of mixtures. Nor does it need to start at 100% of each material at each end, if these combinations are not interesting.
One final note. It is not wise to change too many variables in one test. In the proposed scale above both the concentration of colourant and the ratios of colourants are changed from the previous test. Doing much more than this would be hard to interpret. In this case, these two changes are relatively easy to understand in light of the first test. If, however, the crazing were to increase dramatically, for example, another round of testing would be required to determine whether it was one colourant or the other, or simply the 15 % concentration rather than the 10 % concentration that made the difference. Be very wary of changing more than one variable at a time in tests.