Structure of High Cristobalite

Abstract

THREE structures have been proposed for high cristobalite, which exists for T⪞273° C in pure, well-ordered material. These have been described by Wyckoff¹ and by Sosman² who suggests that all three might be correct for different specimens. The first investigation, by Wyckoff^3,4 himself, from powder photographs taken at 290 and 430° C on a specimen prepared by devitrifying vitreous silica at 1,560° C, placed the structure in space group Fd3m, with 8[SiO₂] per unit cell. The Si atoms form a diamond lattice with the oxygens lying along the lines-of-centres between adjacent silicas. This structure is often still described in textbooks as the structure of real, or sometimes of “ideal”, cristobalite (see refs 5-7). It cannot be correct, however, because it requires an Si-O-Si angle of 180° and an Si-O bond length of 1.54 Å, both of which differ considerably from values established for other forms of silica, where Si-O = 1.61 ± 0.02 Å and , and for silicates, where usually Si-O > 1.58 Å and .