The Modified Description

It cannot be expected that a structural model derived purely from X-ray powder data would provide a complete and reliable description of the actual structure. One, and probably the most important feature of the model of Keggin and Miles, however, seems to be beyond any doubt namely the linear arrangement M- N—C—M —C—N—M- - along the edge of the unit cell. The unit cell constants of a wide variety of Prussian blue analogs have been determined. All the lattice constants measured so far are between 9.9 and 10.9 A. Since the C—N distance is known to be close to 1.14 A (16), the differences in the cell constants directly reflect the differences in the distances N and M —C. 

A very simple but highly significant way of testing a structural h thesis is, of course, the determination of the density to verify the 

The assumption of randomly distributed interstitial, i.e. uncoordinated, metal ions at position 8 c is discarded. Only the positions 4 a and 4 b, the two fourfold positions of the cubic face-centered unit cell, are considered as possible sites for the two metal atoms. Consequently, the occupancies of these two sites depend on the composition of the cyano compound M [M (CN)e]r H20, more specifically, on the stoichiometric ratio k 1. Positions 4 a and 4 b are both completely occupied only when k = l—l. For all other stoichiometries, the positions assigned to M , C, and N are assumed to have occupancies less than 1, whereas position 4 a is stiU occupied by the complete set of four 

The results obtained by applying these general principles to structures with different stoichiometries are summarized in Table 2. A comparison with Table 1 shows that the original model and the modified 

The validity of the modified model has been tested by several complete crystal-structure determinations. Up to the present, almost all the single-crystal studies have been carried out with compoimds of the stoichiometry Ms [MB(CN)6]2(M- = Mn2+, Cd +j M = Co +, Cr3+, Ir +). The compound CsaLiCo(CN)6 studied by Wolberg 50) is not considered as a pol5muclear cyanide in this context, being neither a transition metal nor a metal of class B 34). Despite great structural similarities with the Prussian blue analogs, it is rather an alkaline salt of the mononuclear cyanocobaltate III). This distinction is, of course, arbitrary. 

---