The enhancement of the reduction rate of hematite to magnetite in the presence of potassium oxide is well known. On the other hand, the nucleation of magnetite and its first growth steps have recently been investigated. It was therefore of interest to study, on the same scale, the influence that potassium may have on these stages and to understand the reasons. A series of samples were therefore prepared with reaction times ranging from 0.5 to 45 min during the reduction of hematite single crystals (prepared by chemical vapour transport) at 600 ° C with CO-CO2 (2:98). Also the morphology and growth rate of samples with and without potassium were compared with a scanning electron microscope.

Potassium is known for having a favourable role in the reduction of hematite, either pure or as ore and a detrimental action on the crystals’ mechanical properties. However, several different interpretations have been put forward by previous authors: softening of the gangue, modification of the sintering process, consequences of the alkali penetration into the oxide lattices. Hence new data are desirable and have been recorded, with pure hematite crystals as well as with an hematite ore and with two doping techniques: immersion in a K2CO3 solution, or introduction of potassium components in the reducing gas. Microstructure investigation shows that potassium favours porous magnetite growth rather than lamellar growth and concentrates in a sublayer of magnetite at the inner interface. In fully reduced crystals, potassium lies mainly in the core of the particle. When the interface has a so‐called topochemical configuration, the shrinking core model provides the rate constants as a function of temperature. Arrhenius plots lead to the conclusion that potassium significantly lowers the activation energy. The proposed interpretation is based on the transient formation of KFe11O17, which is revealed by 3 different observations. It may act as a nucleation catalyst, thanks to easy epitaxy with Fe2O3. This is consistent with the change to porous rather than lamellar domains when potassium is brought into play and with the increased crystal fracturation. Hematite ore is less sensitive to potassium because its silica gangue behaves as a trap for it, as shown by treatment with HF before reduction.

In view of the numerous previous interpretations of the role of potassium in the reduction of wustite, experiments have been carried out under conditions that allow a more straightforward conclusion: instead of working with a diphasic solid (wustite partially reduced to iron), wustite single crystals were submitted to a series of redox cycles in CO-CO2 at 890°C, inside its homogeneity range, with and without potassium.

We have previously shown that, in the redox interaction between wustite and a CO-CO2 atmosphere, K+ accelerates the oxygen exchange in both directions. In view of the numerous interpretations suggested by previous workers as regards the role of potassium in the reduction of iron oxides and ores, the action of K+, Ba2+ and Li+ has been compared in order to establish the role of size and charge of the additive ions.