In zinc silicate macrocrystalline glazes, nickel oxide is well known for producing blue crystals set on an orange ground. There is a belief circulating in the crystalline glaze community that the blue colour is caused by cobalt contamination of the nickel oxide rather than directly by nickel oxide itself. To test this, I obtained a sample of high purity nickel oxide with a specified maximum cobalt content of <10ppm. I compared the colours from it with regular technical grade nickel oxide and also with a test, free from nickel, but containing a carefully controlled amount of cobalt. The tests showed the characteristic blue in both nickel series and a lack of blue in the cobalt series even though the amount of cobalt exceeded the amount that could have been contributed by the <10ppm present as a contaminent in the high purity nickel oxide sample.

The nickel oxide was supplied by PI-KEM, the analysis is shown below.

I chose a glaze from my early tests during the year 2002. It containes 3% black nickel oxide as the colourant and is unorthodox in that it contains barium as the only alkaline earth flux. It is realised using my own custom frits, I include the molar analysis to show the oxide content. To establish the correct balance between primary crystal and ground, for each series, I produced a line blend trading Na2O against ZnO.

The glaze with 3% high purity nickel oxide contains at most 0.3ppm by mass of cobalt, expressed as the element. It is well beyond the capability of my weighing equipment to add that sort of level directly. Instead, I added 1g of cobalt oxide, as Co3O4 to 999g of ground silica and mixed them in a ball mill with 1kg of water for 4 hours. After dewatering, I took the powder and screened it through 200s mesh. The cobalt oxide was supplied by Potterycrafts and the analysis shows 0.71 – 0.72 cobalt by mass. This gives a cobalt doped silica containing 735ppm of cobalt, expressed as the element, with an uncertainty of about +/-40ppm. In the nickel free test series, I substituted 4g (in 100g total of dry glaze powder) of this doped silica for non-doped silica giving a final cobalt concentration of ~29ppm, about a factor of 95 times the maximum amount that could be in the high purity nickel glaze (my target was 100 times, but 95 is close enough – about 2 orders of magnitude greater).

For each of the 6 test glazes, I prepared a 50g batch and mixed it in a ball mill with 16g of water. The mill time was 1 hour in each case. For each series, I covered 8 test faces with a glaze loading of 1200g/m2 (dry basis) by a linear wet blend of the the endmembers. The test material was Royale Porcelain. All were fired in close proximity on the same kiln shelf to cone 10 followed by an 8 hour crystal growing cycle in the temperature range 950°C – 1095°C.

Setting aside the differences in crystal nucleation, the series with nickel oxide show little, if any, difference in colour: the crystals are blue and the ground, though partially obscured by prolific secondary crystallisation, is orange/yellow. The tests without nickel, containing 735ppm of cobalt, show only the faintest blue and that shows most strongly in the crystals rather than the glassy ground. The conclusion is that nickel alone is responsible for the blue of the primary crystals. That cobalt produces a detectable colour at only 735ppm shows just how powerful a colurant it is. It certainly seems plausible that cobalt contamination of nickel oxide at the percentage level would influence the overall colour, but the assertion that cobalt contamination is solely responsible for the blue colour is not sustained by these results.

Nickel offers a wide range of colours depending on the glaze chemistry. The colour produced by nickel, and other colourants, also depends on whether the glaze remains glassy or forms crystalline phases. In the glassy phase, nickel is present as the Ni2+ ion. The colour depends on the chemistry of the rest of the glassy phase. With a sufficient concentration of low field strength cations (bearing in mind that Al3+ out-competes Ni2+ for compensation) , the most useful from the potter’s point of view being potassium as K+ (and even more so with rubidium and caesium), Ni2+ exists in fourfold coordination with oxygen and is linked into the main tetrahedral silica network. The characteristic colour is purple. With lower field strength cations such as sodium and lithium, the nickel exists in a distorted trigonal bipyramidal coordination with oxygen and the characteristic colour tends towards brown. This is the ground colour in zinc silicate glazes.

In crystalline phases the colours are different. The blue colour demonstarted here is caused by Ni2+ substituting for Zn2+ in the willemite structure of the primary crystals. It is well-established in the literature [1] that under these conditions the characteristic colour from Ni2+ is blue. There are other fascinating possibilities. Below is a series of tests I carried out several years ago, exploring the chemistry of the popular ‘Nickel Pink’ glaze attributed to Emmanuel Cooper [2]. It is a five way line blend keeping most of the glaze chemistry fixed but systematically varying the proportions of BaO and ZnO. These tests are over Audry Blackman porcelain, fired to cone 8 followed by 6 hours in the temperature range 1000°C-1080°C. The range of colours is fascinating! From left to right, first a slightly dingy purple followed by a vivid purple, characteristic of nickel in tetrahedral coordination. The red is the familiar colour present in Emmanuel Cooper’s glaze as is the following blue. The final blue is characteristic of nickel in zinc silicate crystalline glazes.

To support my research into nickel colours, I have produced a series of frits. These are necessary to introduce high levels of sodium and potassium using insoluble sources that are otherwise unavailable. The frits provide suffificent monovalent compensating cations to charge compensate the nickel allowing it to participate as a glass former in the tetrahedral silica network. To give low solubility the frits also incorporate divalent cations form the alkaline earth series: barium for the lower field strength and calcium or magnesium for the higher field strength. The raw materials for the frit are fully melted at  1200°C an form a black glass that not even a powerful loight source can reveil any colour in. It is only after grinding to a fine powder that the colour is revealed: purple for the K/Ba low field strength and brown for the Na/Ca intermediate field strength. Placing the NiO in the second column recognises that, like alumina, it participates as a glass former when charge balanced. I found that 0.3 mols of NiO (base unity) is close to the limit of incorporation under these conditions. With 0.4 mols and above, instead of a unifom glass, there are two layers: the black glass and a green layer of nickel (II) oxide.

[1] W.A. Weyl; ‘Coloured Glasses’; Society of Glass Technology; p199.

[2] Emmanuel Cooper; ‘The Complete Potter: Glazes’; B. T. Batsford; p84.