• journal_title：American Mineralogist
• Contributor：Manuel Muñoz ; Olivier Vidal ; Clément Marcaillou ; Sakura Pascarelli ; Olivier Mathon ; François Farges
• Publisher：Mineralogical Society of America
• Date：2013-07-01
• Format：text/html
• Language：en
• Identifier：10.2138/am.2013.4289
• journal_abbrev：American Mineralogist
• issn：0003-004X
• volume：98
• issue：7
• firstpage：1187
• section：Articles

We investigated the influence of linear polarization of the synchrotron X-ray beam on the determination of iron oxidation state in phyllosilicates. Fe K-edge XANES spectra and pre-edge peaks have been recorded for various orientations of single crystals of biotite (Bt), clinochlore (chlorite group; Cli), talc (Tc), and antigorite (serpentine group; Ant). Ab initio XANES calculations, performed for 6 orientations of the biotite structure, support the experimental results. Depending on crystal orientation, the experimental results show, (1) important changes both for XANES and pre-edge peaks, (2) characteristic changes of spectral signatures regardless of the mineral species, (3) uncorrelated changes between XANES and pre-edge peaks, and (4) important changes of the energy position of pre-edge peaks, but with no significant change for their integrated areas. Regarding the crystal orientation, the pre-edge peak centroid varies in energy by ~0.4 eV in the case of Bt and Tc, and by ~0.2 eV for Cli and Ant. Such variations correspond to XFe³⁺ (i.e., Fe³⁺/Fe_total) of 0.22 and 0.15, respectively. Comparison with the analysis of powdered samples show that Fe-redox can be generally framed as follows: −(2/3)XFe³⁺ < XFe³⁺_(powder) < +(1/3)XFe³⁺. In good agreement with the so-called “magic angle” theorem for powdered samples, we propose an ideal orientation of the single crystals that provide similar pre-edge peaks as for powdered samples. With the wavevector (i.e., beam direction) perpendicular to (100) or (010), measurements should be done with an angle of 35° between the electric-field vector and the [001] crystal direction. Moreover, measurements performed with the wavevector perpendicular to (001) systematically result in an overestimation of XFe³⁺ up to 0.07. Finally, we show that the appropriate positioning of single crystals reduces the XFe³⁺ uncertainty to the intrinsic error of pre-peak measurements. This approach opens possibilities for in situ analyses of Fe-bearing phyllosilicates in thin sections that are potentially relevant for scientific fields such as hydrothermal/metallogenic, metamorphic, meteoritic or environmental mineralogy.