Extended Abstract: This paper deals with the dehydroxylation reactions of clay minerals and allied substances such as lamellar hydroxides, serpentine minerals, and amphiboles. Probably in all cases, these reactions occur topotaxially, that is, there is a three-dimensional similarity between the crystal structures of starting material and product, as a result of which each crystal of the starting material is changed into something approaching a single crystal of the product. Many investigators have studied the orientation relationships between starting materials and products, and their results have provided the main experimental foundation for hypotheses about the reaction mechanisms. These hypotheses have, however, rested also on assumptions, of which until recently three have usually been made. These are (i) loss of water occurs more or less uniformly from all parts of the crystal; (ii) the Si-O parts of the structure are more stable than the metaloxygen parts; and (iii) the silica-rich material often presumed to be formed as a subsidiary product is pure SiO2. Recent work suggests, however, that none of these assumptions is securely based, and the purpose of this paper is to suggest an alternative approach to the problem.
The high temperature reactions of clay minerals are perhaps best understood by treating the latter as assemblies of close packed, or nearly close packed, oxygen ions with cations in the interstices. From this point of view, one might expect that the relatively small cations, such as Si4+, Al3+, Mg2+, and especially H+ would move more easily than the oxide ions. One would not expect the oxygen atoms that are lost as water to be necessarily those present initially as OH− groups; they are more likely to be those that can most easily escape from the structure, e.g. from near external surfaces or grain boundaries. On this hypothesis, a typical dehydroxylation reaction involves three types of process: Loss of oxide ions from exposed or “donor” regions, and their combination with H+ ions from all regions, to form water. The donor regions are changed into pores. Cation migrations. These may be of two kinds. Firstly, all cations other than H+ migrate from the donor regions into the remaining or “acceptor” regions, and H+ ions migrate in the opposite direction. Secondly, within the acceptor regions themselves, further migrations may cause differentiation into Si-rich and Si-poor parts to occur. Changes in the type of oxygen packing in the aceeptor regions, to complete the formation of the crystalline products. These last changes may occur throughout the acceptor regions, or only in certain parts of them.