T212 The Impact of Crystal Chemistry in the Earth Sciences II: A Tribute to Charles T. Prewitt, Recipient of the 2003 Roebling Medal of the Mineralogical Society of America

Wednesday, 5 November 2003
8:00 AM - 12:00 PM

STEREOCHEMISTRY OF [3]B-SITE AS AN AID IN TOURMALINE NORMALIZATION PROCEDURES

Paper number: T212-12
Program: Technical Sessions
Time: 10:30 AM – 10:45 AM

Authors

Abstract

Chemical analyses of tourmaline are usually incomplete due to difficulties in the analysis of light elements, valence states of transition metals, and problems with the heterogeneity of crystals. When there are undetermined components, assumptions must be made in order to develop a structural formula. One common assumption, all iron is ferrous and OH + F=4 apfu, can result in the misidentification of buergerite as schorl. This can be seen by the exchange vector: Fe3+OFe2+-1(OH)-1, as schorl is defined as Na Fe2+3 Al6 (Si6O18) (BO3)3 (OH)3 (OH) and buergerite as Na Fe3+3 Al6 (Si6O18) (BO3)3 O3 F. Bond-lengths at the Y- and Z-sites have been used to differentiate cation occupancies or valences. However, as multiple species are commonly assigned to the Y- and Z-sites, the distinction between ferrous and ferric iron is not always recognizable based on bond-lengths at these sites. Although the stereochemistry of the triangular (BO3) groups have commonly been assumed to be relatively constant in tourmaline, data indicate that this is not the case. A study into a large suite of tourmalines indicates that the B-O2 and B-O8 bond lengths vary due to the chemical composition of the tourmaline, though the average B-O distance and the amount of B remain constant. Average B-O2 and B-O8 bond lengths for schorl are 1.354 and 1.384 Å and for buergerite are 1.380 and 1.374 Å, respectively. This difference is related to the occupancies of the X-, Y- and Z- sites, with the largest contribution from the Z'-O8 bond. Because of the significant difference in bond-lengths, crystal-structure data related to the B-site of tourmalines in question can be used to verify whether the assumption made during normalization are reasonable, especially in regards to the buergerite vs.schorl issue.