The Behaviour of Iron in Periclase at Non-ambient Conditions

Massimo Boiocchi (boiocchi@crystal.unipv.it), Franca Caucia, Marcello Merli, Danilo Prella & Luciano Ungaretti

Dipartimento di Scienze della Terra, Università di Pavia, Centro di Studio per la Cristallochimica e la Cristallografia, C.N.R., Pavia, via Ferrata 1, 27100 Pavia, Italy

An experimental study has been carried out on samples of synthetic MgO and of natural periclase (Fe content<0.05 apfu). By means of high resolution crystal-structure refinements, microchemical analyses, heating treatments of periclase crystals in air and dipped into hematite powder, it has been established that: 1) the iron present in any natural periclase crystals is always in part trivalent and in part divalent, the former occupying the tetrahedral cavity; as a consequence, four octahedral vacancies are formed around each occupied tetrahedral site, and this corresponds to a syntaxial overgrowth of a spinel phase (magnesioferrite): 2) heating experiments (T<1100°C) allow to increase the amount of trivalent iron at the expense of the octahedral Fe²⁺ and to follow the spinel growth by measuring the intensity of the reflections exclusive of the spinel structure: 3) heating experiments (T=1500°C) on MgO single-crystal dipped into hematite powder produce the formation of a thick rim of a spinel phase; X-ray refinements have shown that spinel is a Mg-bearing magnetite; there is no clear boundary between MgO and spinel, being the oxygen framework almost identical in the two structures: 4) heating the samples made by the two syntaxial phases at T>1500°C, the spinel fraction melts and all the iron forms an alloy with the platinum of the crucible. The high difficulty to form MgO-FeO solid solutions within the lithosphere has been explained on the basis of the high number (12) of very short M-M distances which are present around each cation in this structure. Only very high pressure can contrast the electrostatic Fe-Fe repulsion, allowing the formation of MgO-FeO solid solutions with Fe>1/12 of the octahedral cations. This is in agreement with the hypothesized presence of MgO-FeO solid solutions into the lower mantle (Agee, 1998; Bina, 1998).

Agee CB, Min. Soc. Am. Rev. in Mineral, 37, 165-204, (1998).

Bina CR, Min. Soc. Am. Rev. in Mineral, 37, 205-240, (1998).

Micro-XANES Experiments on Iron K-Edge in Glass Inclusions on ID22 Beamline at the ESRF

Michelle Bonnin-Mosbah (mosbah@drecam.cea.fr)¹, Alexandre Simionovici (simion@esrf.fr)², Nicole Métrich (metrich@drecam.cea.fr)¹, Jean-Paul Duraud (duraud@drecam.cea.fr)¹, Philippe Dillmann (dilmann@drecam.cea.fr)¹, Dominique Massare (massare@drecam.cea.fr)¹ & Anatoly Snigirev (snigirev@esrf.fr)²

¹ Laboratoire Pierre Süed, CEA-CNRS, Saclay, 91191 Gif-Sur-Yvette, France

² ESRF, BP 220, 38043 Grenoble, France

Third generation of X-ray synchrotrons enables to focus photons at the micrometer scale with a high brilliance. X-ray micro-fluorescence analysis and mapping are now currently used. Furthermore oxidation state, binding distance and geometry site can be obtained on small objects by using µ-XANES with X-ray synchrotron radiation. In this work, we present the results obtained for the study of iron in glass inclusions (some µm³) trapped in minerals of volcanological interest. The knowledge of iron oxidation states in these glasses allows to constrain the chemico-physical conditions of the magma in depth.

µ-XANES absorption spectra were acquired both by transmission and detection of X-ray fluorescence at the K-iron edge (around 7.1 keV). Fresnel zone plates are used for the focusing, a pinhole is located just before the sample in order to suppress the diffused X-rays. The beam size turns is about 1.5 x 15 µm² with a fluence of 10⁸ photons/s using the high resolution Si 311 reflex of the monochromator. In these conditions, the pre-edge and main edge shifts permit to perfectly identify the presence of Fe²⁺ and Fe³⁺ as well as their mixing in the glass inclusions.

Our aim is to compare experimental and theoretical results. Indeed, the experimental results seem to show a correlation between the oxidation state ratio and the ratio of the Fe³⁺ and Fe²⁺ areas. Considerations on the coordination number and spin states must be carefully detailed in order to corroborate this assumption. Theoretical calculations aiming to model these features are now underway in our group and will be compared to the experimental ones.