New molecular compounds of the layer lattice type. II. Theory of formation of graphite compounds

Abstract

Experimental results given in Part I of this series (Croft 1956a) have been examined theoretically with particular attention given to finding reasons for the specificity shown by graphite towards various chlorides. This specificity cannot be accounted for in terms of "molecular sieving" or polar interaction. It is suggested that intercalation involves transfer of electrons from the conducting band of graphite to cations of reactive chlorides. Intercalation of transition metal chlorides can be correlated with possession of unsaturated penultimate electron shells by the cations concerned. However, correlation is only complete when the cations exhibit maximum electron affinities, as in their higher valence states. Reaction of Group IIIA chlorides appears due to accommodation of electron pairs from the π band of graphite in n(p) orbitals (or their hybrids) of the cations. The distinctive hydrolysis of these chlorides in situ indicates that graphite is a weaker donor than water. Similar consideration of unreactive chlorides indicates that they are not intercalated because their cations are incapable of electronic interaction with graphite. Several important conclusions, supported experimentally in Parts III and IV of this series (Croft 1956b, 1956c), have been drawn from these theoretical considerations. Firstly, since intercalation depends on electronic configurations of cations, any compound containing a suitable cation should be capable of intercalation in graphite irrespective of the anion associated with it. Secondly, provided they satisfy two conditions, other substances should be capable of intercalating foreign compounds in a similar manner to graphite. The necessary conditions are that a host substance is able to interact electronically with an intercalated substance, and that it must afford physical accommodation for the latter. The potential value of graphite reactions, in elucidating fundamental problems, is also discussed.