Pottery: Iron Glazes

Summary

The more calcium there is in an iron glaze, the more difficult it is to obtain a good red, however some calcium is required to get red from iron. To get a tomato red also requires phosphorus; the red becomes more orange when magnesium is added to these two. The best iron reds in oxidation firing are obtained with about 0.10 molar CaO, 0.03-0.06 MgO, 0.01-0.02 P2O5 and 0.08-0.12 FeO

Recent work suggests that the surface of Fe₂O₃ crystals is actually Fe₁₁O₁₆, a much more chemically active form. This may have something to do with the variances we see when using iron glazes.

Literature

Thickness of application
This is often mentioned as a major determining factor in colour of red iron glazes. In my own work, I have found that a minumum thickness of about 50 µm is required.

Cooling temperature
Cooling schedule is known to be important. Refiring to cone 04 is often recommended to improve the red. Murrow (Ceramics Monthly Sept 2001) found that shino glazes turned red only below 982°C. Marians (Ceramics Monthly June/July 2007) found that temperature holds between 980°C and 870°C during the cooling cycle were critical to the development of red in her iron glaze. She saw three components, one believed to be iron-rich and two silica-rich, at low magnification. My own work confirms these findings, that holds at 950°C for an hour during cooling produce the best reds.

Colour analysis
Murrow&Vandiver (noted in Ceramics Monthly Sept 2001) found that the red colour of shino glazes came from a surface layer of ferric microcrystals about 20 µm thick. Under this layer, the glaze was white. My collaboration with John Stirling found that in iron glazes the surface is iron sesquioxide (Fe₂O₃) while deeper iron is iron monoxide (FeO). Park&Lee (J.Ceram.Soc.Japan 113(1314):161-165 2005) found that in high magnesia glazes the red colour is magnesioferrite MgO.Fe₂O₃.

Calcia
Cardew (Pioneer Pottery) states that alkali earths must be minimised for an iron red, that even 0.2 calcia will turn an iron glaze brown, and proposes a special frit to this end. My work confirms this for calcia.

Magnesia
Park&Lee (op.cit.) found by X-ray diffraction that in their glazes magnesia forms a red colour as magnesioferrite MgO.Fe₂O₃, and that magnesioferrite crystal formation is closely related to the presence of whitlockite-type crystals Ca.9(Mg,Fe).(PO₄).6(PO₃OH). Phosphorus seems to crystallize as whitlockite at 1000-1050°C, magnesioferrite at 900-950°C. Edouard Bastarache (unpublished) considers that the presence of dolomite (Ca,Mg)CO₃ in iron red glazes makes better reds.

Soda
Hamer&Hamer (The Potter's Dictionary) mention that soda combined with small amounts of iron produce blue, and that soda encourages red with large amounts of iron, but give no details.

Phosphorus
Bone ash is reported by many to make more reliable reds; some use ferric phosphate instead of red iron oxide and the bone ash (which contains a lot of calcia).

Boron
Hamer&Hamer (The Potter's Dictionary) mention that boron combined with small amounts of iron produce blue, but give no details. Rhodes (Glazes for the Potter) mentions blue from iron and boron; again no details. Obstler (Out of the Earth, Into the Fire) mentions that boron increases green in iron celadons; no details given. Hesselberth&Roy's Waterfall Brown (Mastering Cone 6 Glazes) obtains green from iron; it has 50% more boron than usual formulations, also more soda. True boron iron greens seem to require the near-absence of calcia, which can only be achieved with frits since natural borates are half calcia.

Titania
Weyl (Coloured Glasses) notes that iron is green when it is a network-modifier (equivalent of interstitial atoms in crystals), and that titania moves it to brown by shifting it to a network former (the equivalent of taking part in a crystal lattice). 4% titania achieved this in my experiments with the calcium iron glaze below.

Migration of iron
There are two unpublished reports of sub-surface iron migrating to the surface of glazes. Ron Roy has noticed this under strong reduction, but not in oxidation under otherwise identical conditions. Hank Murrow believes that fluorine in a glaze assists migration of iron to the surface and uses a percent or less cryolite in his glazes to achieve this.

Sankey Iron Red Custer feldspar 44g silica 16.5g bone ash 14g red iron oxide 11g EPK 10.5g talc 10g lithium carbonate 3 Bentonite 2g COE: 6.8x10-6/K calcia: 10% molar Stoneware (Tucker Smooth White) Source: Kevin Baldwin, adapted to local clays Painted on bisque, fired cone 6 electric, one hour hold at 950°C. Vase is 7 cm high. A very even colour as long as it is thicker than 50 µm, red crystals on a black ground with visual depth. It tends to black where it is thin. By far the most reliable of the iron reds I've tried, and the one I've chosen for my own dinner set. The expansion is high, but it works perfectly on my clay, which is fairly low expansion (6.64x10-6/K), probably due to the high potash content, which increases both elasticity and tensile strength. Microphotos are about 60x. X-ray analysis shows that all the crystals are pure iron oxide; Fe2O3 is the red, FeO the black. Red crystals were analysed at three depths. The data shows that the result of the 950°C soak is primarily to oxidize the iron crystals on the surface to red Fe2O3 while the buried iron remains black FeO. This may be done by movement of oxygen to the surface under chemical-strength forces or by surface oxidation. The glaze is far too viscous at 950°C to permit physical sorting of crystals.

one hour hold at 950°C no hold scanning electron microscope photo of Fe2O3 crystals

Borate Iron Red: Gerstley borate 32g silica 30g Custer feldspar 20g red iron oxide 15g talc 14g EPK 5g bone ash 6g Bentonite 2g COE: 5.6x10-6/K calcia: 14% molar Stoneware (Tucker Smooth White), thrown and trimmed. Source: published many times under many names Dipped on bisque, prefire thickness 0.47 mm. Fired cone 6 electric, two hour rise to maximum temperature (1220°C), held there for 10 min, kiln off until soak temperature reached (typically 30 min.), held there for a soak period, kiln off (5 hr to reach 200°C). I did a series of runs with the same bowl, changing only the soak temperatures. The first involved a long soak to fully develop all crystals that might be formed. The microphoto (about 50x) shows that there are two types of components to this glaze. One group forms very small crystals or crusts on the surface; it forms a rust colour with long soak times. The other component oozes out to the surface without crystalizing and is often yellow, sometimes bright. It is probably ferrosilite (see below). Under certain temperature regimes the ferrosilate is coloured a dull red by the iron. Bright colours always seem associated with surface components; the two other microphotos shown are typical of the variety of colours seen. X-ray analyses show that the surface crusts are thinner than 4 µm, the effective penetration depth of the 20 kV electrons used; reliable analyses could not be obtained. Black gradually took over the bowl with repeated firings (the 980°C at right was the eighth in the series). So, if you don't get the colour you wish with this glaze, try refirings, but not too many. I obtained the most interesting colours with moderate-length soaks in the 900-980°C range. X-ray analysis of the final glaze (the 980°C photo) showed considerable fine and medium-scale differentiation in the surface. There were large patches of nearly pure silica (87%). Other patches had double the concentration of iron as the total glaze formulation; all these were low in calcia. There was also evidence of calcium silicate. It was not possible to match the X-ray image to an optically-visible feature of the glaze. In contrast to some reports, I found this to be a very well-mannered glaze as I mixed it, not at all runny, or even droopy when laid on thickly, as you can see from the lack of problems with the sharp horizontal edges of the 8 cm diameter sugar bowl.

5 hr hold at 870°C 30 min hold at 940°C 30 min hold at 980°C glaze painted thickly on the outside, thinly on the inside, 1 hr hold at 920°C scanning electron microscope photo of a 0.7 mm square portion of the surface, showing a typical crust-like differentiation.

Calcium Iron: Wollastonite 28g EPK 28g Fusion F2 frit 23g silica 17g red iron oxide 7g nepheline syenite 4g Bentonite 2g cobalt carbonate 2g COE: 5.5x10-6/K calcia: 17% molar Porcelain (Tucker 6-50), thrown and trimmed. Dipped on bisque, fired cone 6 electric. This began as an attempt at a matte lustre black, but turned out to be a microcrystalline glaze. With the same bowl, a fast cool of this glaze gives a glossy black, then a refire followed by a slow cool converts it to a mottled black and green semi-matte, and vice versa. Green crystal size is dependent upon cooling rate. Cooling from 1200°C at 50C/hr to 800°C produced crystals typically 2 mm in diameter. At 80C/hr the crystals were about 1 mm, at 100C/hr, ½ mm. 150C/hr is required to get a smooth surface, but then it's close to glossy. Unless fired glossy, iron in excess of 7% comes out of solution to form a metallic layer between the green crystals. Bowl is 10 cm diameter. Besides the individual oxides, possible mineral compositions include Andalusite Al2O3.SiO2, Anorthite CaO.Al2O3.2(SiO2), Wollastonite CaO.SiO2, Fayalite 2(FeO).SiO2, Hedenbergite CaFe.2(SiO3) and Ferrosilite FeO.SiO2. [mineral photos] The right half of the bowl is the result of a 10 hr soak at 1120°C after firing to 1220°C. X-ray analysis shows that the crystals are FeO; as shown in the microphoto (about 100x), they are mostly long rhombs. The surface is rough to the touch. The left half of the bowl had a 10 hr soak at 870°C after firing to 1220°C. This produces a surface layer of greenish-yellow crystal needles that grow in six rays from a central point. The colour could match either Hedenbergite or Ferrosilite, but the crystal form is most consistent with Hedenbergite. The crystals are so thin that their X-ray output was mixed with background material output, but it also indicates Hedenbergite. Not an attractive or useful glaze.

I thank John Stirling of Natural Resources Canada's Geological Survey for making the scanning electron microscope guided X-ray analyses.

Iron Glaze Chemistry

To investigate the interaction of calcia, magnesia and phosphorus with iron in oxidising fired glazes, a series of mixes were made. A base glaze contained none of any of the three oxides under study, and one glaze each was made up similar to the composition of the base glaze, but with a large quantity of each of the oxides in turn. The target analysis for each was 3.5 Seger SiO₂, 0.4 Al₂O₃ and 0.25 B₂O₃ (B₂O₃ and FeO omitted from Seger ratios).

mix	recipe	molar Seger oxide
base	28 silica 28 potassium carbonate 24 kaolin,EPK 10 iron oxide,red 10 frit,Fusion 367	0.602  3.392  SiO2 0.166 0.935 K2O 0.102 0.575 FeO 0.010 0.059 Na2O 0.072 0.404 Al2O3 0.045 0.253 B2O3
magnesia	30 silica 25 magnesium sulphate 25 kaolin,EPK 10 iron oxide,red 10 frit,Fusion 367	0.611  3.497  SiO2 0.163 0.935 MgO 0.098 0.562 FeO 0.072 0.411 Al2O3 0.043 0.248 B2O3 0.010 0.058 Na2O
calcia	32 silica 26 kaolin,EPK 22 calcium carbonate 10 iron oxide,red 10 frit,Fusion 367	0.615  3.485  SiO2 0.166 0.939 CaO 0.094 0.533 FeO 0.071 0.405 Al2O3 0.041 0.234 B2O3 0.010 0.055 Na2O
phosphorus	27 silica 26 potassium carbonate 23 kaolin,EPK 15 iron phosphate 9 frit,Fusion 367	0.587  3.494  SiO2 0.157 0.936 K2O 0.100 0.596 FeO 0.070 0.418 Al2O3 0.041 0.246 B2O3 0.033 0.197 P2O5 0.010 0.057 Na2O

Since the potassium and magnesium salts are hygroscopic (magnesium sulphate particularly so), each of these was baked at 220°C until anhydrous before weighing. Each glaze mix was then ground in a ball mill until uniformly fine. An attempt was made to use a 200 mesh seive on the dry mixes, but the hygroscopic-powered clumping of the soluble salts made that impractical - lumps had to be taken out after mixing with the oil.

Each tile glaze was made with four 1/4 tsp portions (1 tsp=5 ml) of dry material, using a precision stainless steel measure, then mixed with 3/4 tsp corn oil. (Water could not be used due to the soluble salts required to separate the oxides.) They were fired to 1290°C (cone 10), cooled to 950°C, then held there for 1 hr.

Since the best results were obtained with the maximum amount of phosphorus available from these mixes, and calcium was shown to be required, a supplementary mix was made up using bone ash (which contains calcium) to obtain larger amounts of phosphorus. Finally, the ratio that gave the best colour response was tested with 0-25% added red iron oxide. Some of the 54 test tiles are shown at right. Although not exactly decorative, they show several useful results for oxidation firing above 1000°C:

1. Although excess calcium is known to turn iron glazes brown, some calcium is required to get red from iron.
2. To get an intense red also requires phosphorus, but too much phosphorus turns the glaze brown.
3. The red becomes more orange when magnesium is added to these two, but too much magnesium turns the glaze grey.
4. The best iron colour responses in this series were obtained with 0.10 molar CaO, 0.03-0.06 MgO, 0.01-0.02 P2O5.
5. Using a base with these proportions fired to cone 6, the best reds were obtained with 9-14% added red iron oxide.
6. 1% cryolite (fluorine) may help to brighten the red on the surface a bit, but gives foaming to spitting problems during firing if used even slightly in excess or if insufficiently finely divided.

I have located only one phosphorus-iron mineral that is red: Simferite Li(Mg,Fe⁺⁺⁺,Mn⁺⁺⁺)2(PO₄), which is not a player since none of my test series contained lithia. Park&Lee identified a magnesium-iron compound as their red. However, they seem to have concluded that the role of phosphorus is to tie up calcia so it doesn't interfere with the formation of magnesioferrite. It can't be as simple as that, since I got zero red from any glaze that did not contain calcia, and only a trace of very dark red on glazes that contained calcia but no phosphorus.

My current working hypothesis is that a calcium-phorphorus compound acts as a promotor (perhaps catalyst) for the oxidation of FeO to Fe₂O₃. It is possible that the two instead tie up something else that inhibits the oxidation, but this seems less likely.

Notes:
The tests were originally planned for cone 6 and boron added that is usually sufficient for it. Although the magnesium mix melted at cone 6, all mixtures with it turned viscous and foamy as soon as they melted. Cone 10, the maximum for my kiln, helped with some, but was still inadequate for a few.
Every stray cat in the neighbourhood was attracted to the kiln outside air vent by the smell of the corn oil!

Microphotos Here is the setup I use to take my microphotos. My microscope is a basic full-size frame - no condenser, fine focus, stage movement or anything else expensive. The objective is a 5x 0.1 NA - low NA gives large depth of field, important for viewing solid objects. The eyepiece is a 10x periplan, selected from a dozen brands in the store for clearest view over the field with this objective. A point&shoot camera is placed on the eyepiece, moved from side to side until all the circle of light from the microscope is visible on the screen, then zoomed until the edge of the circular field just touches the edges of the view. That standardizes the magnification of the photos so only one session with measuring scales is required. The microscope is focussed until the image is sharp on the LCD, then the shutter pressed. John Sankey other notes on pottery other microphoto methods