Crystal chemistry of the schorl-dravite series
Bosi, Ferdinando; Lucchesi, Sergio
European Journal of Mineralogy Volume 16 Number 2 (2004), p. 335 - 344
published: Mar 29, 2004
ArtNo. ESP147051602015, Price: 29.00 €
Nineteen tourmaline samples of various provenances and geological settings were studied by EMPA, SREF and MS to represent the schorl-dravite compositional field. All samples belong to the Alkali group (except one with an X-site vacancy content of 0.53 apfu) and to the Oxy- and Hydroxy-subgroups. Among divalent cations, the main substitution involves YMg for YFe2+, to produce the two end-members dravite and schorl. Site populations were determined by a new minimization procedure that simultaneously accounts for both structural and chemical data. Results show that the crystals are characterized by disordered cation distribution between Y and Z sites: Al populates both sites, with a marked preference for the smaller Z octahedron; Mg is often equally distributed between Y and Z. Both Fe2+ and Fe3+ populate both Y and Z sites, but show a strong preference for Y. Specific mean bond distances (Å) optimised for major elements are: YAl-O = 1.908, YMg-O = 2.084, YFe2+-O = 2.139, ZAl-O = 1.900, ZMg-O = 2.077 and ZFe2+-O = 2.131. In the schorl-dravite solid solution, structural variations appear to be primarily due to Y and Z interactions. These effects are conspicuous over the entire structure, as Y dimensions directly affect the a cell parameter, while Z is similarly correlated with c. The dimensions of Y and Z octahedra are determined by Al contents. Dimensional variations of Z are well described by its bond-distance variations, except for Z-O7D. Both octahedra reciprocally interact, influencing their distortions: inverse correlations exist between Y dimension vs. Z quadratic elongation and Z dimension vs. Y quadratic elongation. As a common feature, the effects of the octahedral second coordination sphere are only confined to polyhedral distortions instead of dimensional variations, which only depend on site populations.