Incommensurate superlattice ordering in priderite

The mineral priderite has been examined by electron microprobe analysis and electron diffraction. The composition was approximately (Ko.97Bao.aa)(FeH7 Ti6.60Mg0.23)O16, which is in line with the observed superlattice periodicity of m = 5.88, there being intergrowths of rn = 5 and m = 6 superstructures. A surprisingly high degree of lateral correlation between tunnel cations was noted. RECENTLY considerable interest has been shown in the crystal chemistry of compounds with the hollandite structure. One of the reasons for the interest is the proposal by Ringwood et al. (1979) to employ a synthetic rock, 'synroc', in the disposal of radioactive waste, one of the key components of this being the hollandite phase BaA12Ti6016, which is intended as a host material for radioactive barium and cesium. The chemistry and structure of this phase is very similar to that of the mineral priderite, (Ba,K)l.a3(Fe,Ti)8016 , which was discovered in the leucite-bearing rocks of the West Kimberley region of Western Australia and first described by Norrish (1951). The essential details of the hollandite structure are quite simple and were described originally by Bystrom and Bystrom (1950) for the mineral hollandite, of composition BaMn8016. Subsequent structural refinements have now been completed on a wide variety of chemically distinct hollandite phases, including that of the synroc hollandite structure (Sinclair et al., 1980) and of the mineral priderite (Sinclair and McLaughlin, 1982; Post et al., 1982). The hollandite structure, depicted schematically in fig. 1, consists of an oxide framework, which is not close-packed, with Fe and Ti in octahedral co-ordination. Octahedra are linked by edge-sharing to form double strings parallel to the c-axis. These strings of octahedra share corners with each other to form a three-dimensional framework containing tunnels of approximately square cross-section running parallel to the c-axis. Ideally the crystal symmetry is tetragonal, but a monoclinic distortion has been observed in some hollandites (Cad6e and Verschoor, 1978). The large cations (Ba and K in the case of priderite) are accommodated in sites within the tunnels, these (~ Copyright the Mineralogical Society FIG. 1. [001] projection of the hoUandite structure. [MO6] octahedra share edges along the c-axis, forming strings which are linked by corner-sharing and edgesharing to form square tunnels which contain the large cations (Ba2+,K+). sites being eightfold co-ordinated in a nominally cubic geometry. Diffuse superlattice reflections indicating cationic ordering within the tunnels was first reported by Dryden and Wadsley (1958) for the compound BaxMgxTia_xO16 , with x in the range 0.67 < x < 1.14. They proposed a structural model based upon sequences of alternating Ba cations and vacant tunnel sites. More detailed measurement of these diffuse reflections, which generally have irrational repeats in reciprocal space (Bursill and Grzinic, 1980) has shown that this picture is an oversimplification of the situation. These latter workers established that incommensurate superlattice ordering occurred in this compound, and that the exact positions of the diffuse reflections was a function of composition. They therefore proposed a structural model based upon complex intergrowths of phases with doubled, tripled, and 66 A. P R I N G A N D D. A. quintupled c-axis repeats, involving considerable short-range order. The possible importance of 'synroc' and hence of the 'synroc' hollandite has made a full understanding of this and related phases of vital importance in optimizing the stability of the material. Priderite, being a natural analogue of this synroc hollandite, provides a means of investigating the structure of a stable hollandite containing a mixture of cations within the tunnel sites. An examination of the superlattice ordering in priderite also proves a useful test for the Bursill and Grzinic model under geological conditions of crystallization.