Mass balance vs Rietveld refinement to determine the modal composition of ultramafic rocks: The case study of mantle peridotites from Northern Victoria Land (Antarctica)

• Published in: Tectonophysics
• Main Authors: GENTILI, SILVIA, COMODI, Paola, Bonadiman, C., Coltorti, M.
• Other Authors: Gentili, Silvia, Comodi, Paola
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2015
• Subjects: Rietveld method; quantitative analysis; mantle xenoliths; Victoria Land; Antarc*; Antarctica
• Online Access: http://hdl.handle.net/11391/1354792; https://doi.org/10.1016/j.tecto.2015.01.024

This paper aims at applying the Rietveld refinement method to X-ray powder diffraction (XRPD) spectra in order to determine the mineralogical assemblages of ultramafic rocks. The results are compared to those obtained by mass balance (MB) calculations, a least squares method that reduces major element chemical analyses to the whole rock composition. This methodological work was carried out in five hydrous and anhydrous mantle xenoliths from Northern Victoria Land, Antarctica.The intrinsic goodness of the XRPD and MB results, evaluated by with Rwp and r2 respectively, shows similar values of modal compositions. Independent of the mineralogy (amphibole-bearing and amphibole-free) and textures (fine to coarse grained), good agreement (<2%) is observed for modal contents lower than 10%, whereas a discrepancy of up to 10% is recorded for phase abundances higher than 70%. The comparison of the two methods has allowed us to demonstrate that the Rietveld method is applicable even when limited amount of material (<1g) is available, and the "classical" chemical approaches (XRF and EMPA) cannot be applied. Moreover it not only provides information on rock mineral concentrations (wt.%), but also gives supplementary information on crystallographic data (i.e., mineral crystallite size, mineral lattice parameters, density, site occupancies).On the other hand, the MB procedure provides direct chemical information of both bulk rock and minerals, but requires a large amount of material (i.e., for XRF at least >. 1. g of material is needed). Dealing with MB procedure, problems can occur if accessory minerals cannot be identified in thin sections, which are, however, part of the whole-rock analysis. Besides, the strength of XRPD, if compared with all the suitable methods for the quantitative mineralogical identification, is to provide direct information about the physical properties, and mineral site occupancies that could indirectly give a mineral chemical composition. © 2015 Elsevier B.V.