Science Page 7
Chromium
Image: Lead chromate
At the moment I'm looking at yellow glazes, and one way to achieve them is through chromium (not chrome - that just refers to shiny plated metal). Whilst yellow chromium glaze recipes are not uncommon, the underlying chemistry seems unclear, so this post is poking around at that. I'm not looking at pigments containin Cr, as these are largely inert.
Chromium is most commonly found in the green trivalent form, Cr2O3 or chromium oxide. From the formula it looks like it'll be an intermediate, and this is reinforced by Cr and Al being of very similar size. And this is what happens, except the solubility of Cr2O3 is very low, around 1% molar, so not much enters the glaze. Any excess remains as Cr2O3 crystals, opacifying the glaze.
It is also fairly common in the hexavalent form as the yellow chromates or orange dichromates. The best known of these is chrome yellow, PbCrO4. But barium, calcium, iron, magnesium, nickel, potassium, sodium, strontium and zinc can all form chromates, though it isn't clear if all can be formed in glazes, nor what the requirements are. Where they can, you'll see the unfired green glaze come out a shade of yellow after firing, as it changes from trivalent to hexavalent. Then there's the red potassium and sodium dichromates as well. All of these have the same low solubility as Cr2O3, and generally perform well as seeds for crystal growth.
The other form of note is CrO3. This (and not Cr2O3) is the volatile form of chromium. It may form from either direct oxidation of Cr2O3 (above 1000C), or from other hexavalent compounds.
Hexavalent chromium compounds are all highly toxic and carcinogenic, so no yellow or red chromium glaze should be used in food contact areas.
The hexavalent forms will only form in an oxidation firing. In reduction above 1160C CrO forms, turning the glaze blue; at lower temperatures things get complicated with pyroxene crystals forming. Other oxides with other valencies have also been reported, though not studied in depth yet.
All a bit more complex than just a simple green colourant!
Another Yellow
Continuing with chromium, another yellow. This is based on perovskite, CaTiO3 (but may work if you substitute the Ti with Sn). Adding <0.1% molar of Cr2O3 to the CaTiO3 will result in a yellow, with the strongest yellow at about 0.03%. Practically speaking, as any excess Ca will just act as a standard alkali earth, the Cr2O3 needs to be 0.03% of the TiO2.
I was asked about chrome tin pinks and reds. These come from the sphene CaSnSiO5 into which some Cr2O3 is introduced, replacing Sn. Here the Cr has a valency of +4, so this is another colour that needs oxidation to form. The colour forms when heated to above 1200C, and a hotter firing or longer hold increases the colour intensity. Note that there can be a number of variations on this theme; the most common of which is replacing some or all of the Sn with Ti. The sphene structure can also be made using Ba or Sr instead of Ca (I haven't found reference to Mg); I don't know if these can be formed in the kiln, or what the colour would be like. Also, it would be interesting to play with other +4 ions to replace the Cr - of the materials we use, this could be any of the transition metals except Zn and Cd. As well as the charge, the size of the ion will be significant; in the top row of the transition metals, size decreases as we go from left to right, so Mn and V look particularly interesting. Something to explore!
When making stains in the glaze like this, the stain components are obviously not available to participate in the glaze so, strictly speaking, they should probably be moved to the Other column in the UMF.
Other routes to red with Cr are corundum (Al2O3 with Cr3+, aka ruby) and spinel (MnAl2O4.Cr3+). In corundum, some of the Al is replaced with Cr (about 1%); the bigger size and different charge structure of Cr distorts the crystal and gives it a red colour. Spinels are a large group of crystals with the basic structure A B2 X4, where A has a charge of +2, B has +3, and for us X is oxygen with a charge of -2. Here, the Al is replaced by Cr. This spinel isn't too stable in a glaze, but better in a clay body.
Colourants are a huge topic, and after this little diversion I'll return to glaze structures.
Copper
Copper glaze, varying Cu content and level of reduction
1 - in oxidation
In oxidation, copper is generally a modifier, as blue-green Cu+ or colourless Cu2+. Glazes generally have a mix of both. As would be expected, Cu+ acts in the same way as the alkaline metals, whereas Cu2+ acts in the same way as the alkaline earths. Note it also destabilizes many glazes - in lead glazes, Cu increases leaching of lead by a factor of 10, and observation of other lead-free glazes shows it often has a similar effect.
Increasing K, Li, Na or Sn, or firing higher, increases Cu+. Or increase Cu2+ by increasing Al, B, Ba, Ca, Mg, Pb, Sr, Ti, or adding more Cu to the glaze. Firing hotter also converts more copper to Cu+. The different colours in oxidation are primarily driven by the ratio of Cu2+:Cu+, ranging from blue when Cu+ is high, to green with high Cu2+.
The blue Cu+ is in an octahedron, whereas Cu2+ takes on a linear shape. Whilst many put the green colour down to the ratio of the two ions, we now know that green occurs when the Cu+ forms a tetrahedron. This enables the Cu to switch role from a modifier to a glass former.
For wanting more chemistry, transition metal colours arise because the d-orbitals are not all filled and are not all at the same energy. Thus a photon can hit an electron and be absorbed by it, moving to a higher energy d-orbital, but absorbing the wavelength of light from the photon. Losing red light gives us a green colour, or losing orange gives us blue. The colour is not a property of the metal ion on its own, but depends on the form the metal takes due to the influences of the surrounding glaze, so changes in the composition of the rest of the glaze affects the colour we perceive. For more on this, google Ligand field theory.
A high Cu content (>10% molar) results in crystallisation of CuO, but increasing B or P delays crystal formation. Increasing Cu (below crystalisation levels) weakens Si glazes, but strengthens P and (I would guess) B. Cu also bonds with P or B in preference to Si.
As Cu increases above its solubility level, it forms blobs of CuO on the surface.
Above 1225C copper starts volatilising, so some of the copper in the raw glaze will be lost.
2 - Reduction (I)
Image is an Egyptian red glass amulet, 1300BCE
Continuing with copper, things get more complicated in reduction with the creation of copper reds. This requires more than just the reduction of the copper, so has spilled over to four posts!
Whilst the chemistry of copper reds is complex, copper red opaque glasses were produced from 3000BCE in Mesopotamia, and Chinese copper red glazes originated in the 9th century CE, before being replicated in Europe at the end of the 18th century. This is a useful reminder as to how the scientific approach is not the only way to gain knowledge and mastery of something.
On delving into the chemistry of copper reds, it is complex, and no single source has a full description. This is pulled together from multiple books and papers to give what I believe is a true, if brief, picture.
In reduction, the CuO or Cu2O is reduced to metallic copper. This may be achieved by firing in moderate reduction, or addition of fine SiC to the glaze, and both are helped by addition of a reducing agent such as antimony or molybdenum sulfides. To ensure all the Cu is reduced, this wants to occur before the glaze melts and so seals off the copper oxide particles. Get it right and you get red, but too much reduction gives muddy brown, whilst not enough gives a clear glaze.
Copper reds can be created in an electric kiln by adding fine silicon carbide for in-glaze reduction. One test had 0.6% by weight copper oxide in the dry glaze, and the best red was achieved with 0.03% SiC.
Note that to get the brightest reds you need the smallest amounts of copper, but these are also the most pernickety to fire. Low fired glazes require less Cu, as less is vaporised during the firing. Note that some sources say Cu2O crystals are responsible for the red colour, but I believe that the evidence for it being Cu crystals is considerably stronger.
Reduction (II)
Small amounts of the reduced Cu metal can dissolve in the silica glaze - we need to dissolve all of our copper, as blobs of undissolved metal won't give us red. This is why copper reds actually contain very little Cu (down to about 0.1% molar). But on cooling the Cu will precipitate out at about 1083C (the melting point of Cu), and it forms small crystals of copper, atoms moving by diffusion through the semi-solidified glaze. These crystals are about 10 to 150nm diameter, mostly below 50nm. Like so many crystalline glazes, copper reds tend to low levels of Al. This reduces the viscosity of the glaze, making it easier for the Cu atoms to move through the setting glaze - but it also makes the glaze run, and thinning of the glaze around the rim may result in a clear glaze as the copper reoxidizes on cooling.
A section through a test tile, and a slice through the red layer
If we take a slice through the glaze, we may see several layers. Going from the surface down, we have a clear layer, possibly a thin yellow layer, the red layer, possibly a blue layer, and probably a clear layer.
The top clear layer is probably formed by oxygen from the air diffusing into the glaze and oxidising the Cu to Cu2O (after the firing has returned to oxidation), which dissolves into the glaze rather than staying as a crystal; note that too much oxygen may take copper close to the surface through to green CuO. The bottom clear layer is also due to oxidation, probably due to components of the clay body dissolving into the glaze and reducing, thereby oxidising the Cu; if the clay body is very pure, and/or the firing is low and fast, this may not develop.
Reduction (III)
Whilst we don't normally see Cu as red, at nanoparticle scale things work differently. The small particles (<50nm) just absorb light, whilst the larger ones (>100nm) mostly scatter light. Modelling the passage of light through all the glaze layers tells us that:
- as the ratio of large:small crystals increases, the glaze reflects more light and so is brighter, but the extra reflected light covers the whole spectrum, so the red is less intense. As the proportion of large crystals heads towards zero, the glaze heads towards being black
- a high proportion of small crystals gives a stronger red, whereas more large crystals shift the colour to purple
- the brightness decreases but colour saturation increases as the proportion of copper crystals in the glaze increases up to a certain threshold value, beyond which it is pretty constant
- if the large crystals get much above 150nm, the glaze may become more reflective and shift back towards red
- if the glaze is too thin, then there may not be a layer of red, resulting in a clear glaze, but if too thick the colour becomes more livery.
If the copper crystals are covered with a layer of Cu2O, as may occur if they are being oxidised, then the colour ranges from red at 100% Cu, through deeper red to purple at 75% Cu, blue at 50%, through green to yellow at over 90% Cu2O. I believe that the yellow and blue layers could form due to incomplete oxidation and dissolution of the Cu to Cu2O.
Being a bit more specific about the thickness of the glaze layer, one series of tests changed the (unfired) glaze thickness from 0.2 to 1.0mm in 0.2mm steps. The best red was at 0.4mm, with a white glaze at 0.2, and the red giving way to all cyan by 0.8, and a darker blue at 1mm. The thinnest glaze has cristobalite and tridymite crystals, but these reduced with thickness, and had gone by 0.8. All thicknesses had a red layer at the bottom. The topmost layer went from milky white at 0.2mm, through to cyan at 0.6, then blue-green and finally blue-purple. Between these there is a transparent layer up to 0.4mm, replaced by a purple layer from mixing of the red and blue layers in thicker glazes.
Reduction (IV)
Copper red in an electric kiln, using silicon carbide for reduction, by Mark Egan - you'll find this in Glazy
Carrying on with considering the thickness of copper red glazes:
At all glaze thicknesses, the modifiers were strongest on the surface and decreasing towards the bottom, whereas the Si and Al showed the opposite trend - this may be due to the modifiers diffusing from the bottom of the glaze into the clay body. Up to 0.6mm, the Cu and Sn were fairly uniform with depth, but in the thicker glazes there was noticeably less at the bottom.
Because the white and red colours are formed by light falling on the Cu crystals, it is not greatly dependent on the chemistry of the base glaze. What helps is a high solubility of copper into the glaze, and lead glazes increase this significantly, though too much Cu favours the growth of Cu2O crystals instead of Cu. If way too much Cu is added then it forms as blobs, outside of the glaze. Also, we want to minimise the time that the glaze is in oxidation, so less Cu is oxidised, and using frits means that the glaze will become homogenous more quickly. The colour will probably also vary if other glass formers are introduced.
Of course, as we're growing crystals the firing schedule is important, in particular when we drop the temperature enough for the crystals to grow. Tests on soda-lime glasses showed that red formation was not too dependent on temperature, but green from oxidation was a minimum around 610-762C, increasing either side; also the amount of green increased as the hold increased beyond 20 minutes. Too high a temperature and the red disappeared; too low and it took longer and longer to form, if at all. Obviously the optimal temperature will vary with the glaze viscosity, and with glazes being more viscous than glass we'd expect higher temperatures, but this does show how critical the firing schedule can be.
I'd hoped to finish on copper reds this week, but the post length limit has beaten my tendency to waffle on, so next week we'll finish by looking at tin and a couple of other things.
Reduction (V)
Ming dynasty copper red bowl
Tin is very important, though its mechanism isn't fully understood. Primarily, it is reduced from SnO2 to SnO during the reduction firing, and then when firing returns to oxidation it reoxidises, reducing any Cu2O or CuO it finds in the glaze to the metal. Other claims for it are that it helps the copper stay in solution in the glaze at higher temperatures, and then bonds with the growing crystals, preventing them from becoming too large. The glaze needs at least the same amount of Sn as Cu; one paper found that at a Sn:Cu ratio of 1:1 only half of the Cu was reduced to the metal, and some recipes have rations of up to 5:1. Insufficient Sn favours the growth of Cu2O crystals over Cu. Note that tin is unnecessary in lead copper red glazes, as the lead acts in a similar way to tin. But other polyvalent oxides are worth trying, e.g. glass makers have used As or Sb.
A small amount of Fe has been found to be beneficial, though the mechanism isn't yet understood, other than it not acting as a reducing agent.
And what happens if we grow larger copper crystals? First we get an opaque glaze, similar to sealing wax and called hematinone by glass workers, and then we get an aventurine with larger but fewer crystals - both have more Cu than copper reds, up to about 5%.
As an aside, this same approach of producing colour may be used with other metals - gold and silver are the best known, and also different glass formers result in different colours, if you want to explore. Also, a similar layering of glazes may be found in temmoku and other glazes that require a reduced colourant for their success.
Page 6 < > Page 8