ABOUT A POSSIBLE RELATIONSHIP BETWEEN THE STRUCTURE OF TWIN BOUNDARIES AND REE CONTENT IN FLUORITE CRYSTALS
Shkurskiy B.B.
Moscow State University, Department of Geology, Russia
In this work several problems of altervalent substitutions in imperfect crystals of some nonstoichiometric fluorides are considered. Fluorite-like compounds with the general formula M1-xRxF2+x, where M = Ca, Sr, Ba; R = Y, La and other REE, have been theoretically studied. A new scheme of substitutions at twin boundaries is proposed.
In accordance with Willis' model [1] a defect structure of fluoride compounds contains an excess anions of F' in 48i (0.5,u,u), 32f (v,v,v) Wyckoff positions and F-vacancies in 8c (1/4,1/4,1/4) positions. Pauling's ABC-notation can be applied to M2+ and R3+ arrangements. Let us regard interpenetrate twin on (111) of fluorite. Planes' and vectors' symbols correspond to point group 6/mmm of the twin, e.g. (001) will correspond to (100) of m3m.
Suppositional atomic structure near (001) twin boundary can be represented as (...A,BCB,A...) sequence of 36 M-layers. This model was obtained from cubic Fm3m structure described as (...ABCABC...)-sequence by reflection in (001) plane coinciding in with C-layer.
Flat hexagonal domain (..,BCB,..) one can regard as a three-layer framework of M- and F- atoms. It is suitable choice of origin for P63/mmc cell at 2c(2/3,1/3,3/4), then M-site is 2c (0, 0, 0) and F-atoms' site is 4f (1/3, 2/3, z). These F-atoms centering adjacent FM4-tetrahedra form a wavy 63-layers localized between (..,BCB,..) M-layers. As least distance F-F are too small, structure should be characterized by strong steric strains.
There are two ways of structure relaxation, which carried out togheser. The first way consists in movement of nearest F-atoms from their sites along threefold axes towards opposite vertices of adjacent FCa4-tetrahedra, and the second one is some dilation (det(M)= approx 1.1 for tysonite) along the same direction. Parameters of these transformations was calculated.
The pair of empty reflect-related F8-cubes gives a polyhedron of a trigonal symmetry. It has 14 vertices (one of 8 atoms per each cubes is loosed). M-atoms have the coordination numbers 8 (base-centered trigonal prisms). Presence of the F-vacancies makes rather possible altervalent REE3+ --> M2+ substitutions. Excess anions can preferentially fill in M3-coordinated sites, that is similar to ones due to Willis' scheme. In spite of redundancy number of vacancies in 4e (1/3,2/3,z), which are images of 32f (v,v,v), its can be reduced to 2b (1/3,2/3,0).
Summing up we can say that F-atoms occupy two point systems: 4f (2/3,1/3,z) and 2b (1/3,2/3,0) of P63/mmc. M- or REE-atoms site in 2c (0,0,0) and the origin is there.
If all of M-atoms of C-layer are substituted with REE, C-layer and its F-neighboring is not other that planar fragment of the tysonite-type structure, e.g. (Ce,La)F3 - a mineral fluocerite-(Ce), SmF3, LaF3 etc. REE-atoms are 11-coordinated with F-atoms, which form all-faces-centered trigonal prisms around REE-cations as well as it takes place in tysonite-type structures. These planar tysonite-like clusters localizes exactly at twin boundaries.
If only even number of stacking-falls (..ABCBA..) are realized, more large along c-axis fragments of tysonite-like REE-fluorides can be concluded in fluorite crystals without twinning. If these fragments form three-dimensional cluster and its areas and thickness are small (only a few of atoms), its can form tetrahedral or octahedral clusters early proposed [2].
Y and REE of it's family can probably chow another substitutional preference. The twin boundaries (112) for Fm3m ( (100) for P63/mmc) seem be having higher Y-capasity with comparison to (111) or (001) correspondely.
Every attempt to observing of these
tysonite-like REE-fluorides occuring as thin fragments in
fluorite-like crystals will be desirable.
f
ff