The HT diffusion of hydrogen in riebeckite

Federico Galdenzi^1,2, Giancarlo Della Ventura^1,2, Umberto Susta¹, Francesco Radica¹, Augusto Marcelli^2,3

 

1 Dip. Scienze, Università di Roma Tre, L. S. Leonardo Murialdo 1, 00146, Rome

2 INFN-LNF, Via E. Fermi 40, Frascati 00044 (Rome)

3 Rome International Centre for Material Science Superstripes - RICMASS, Via dei Sabelli 119A, 00185 Rome, Italy

                                                 

In this work we address the diffusion of hydrogen at high temperature in a sample of riebeckite close to the end-member composition Na₂Fe³⁺₂ Fe²⁺₃Si₈O₂₂(OH)₂. We carried out isothermal experiments on both powders and single-crystals and monitored the behavior of the O-H stretching signal by FTIR spectroscopy. Two different sets of experiments were performed: in the first one we collected data on five doubly polished chips with the same thickness (85 µm) at different temperatures, in the range 520 to 560 °C. For the second set, we collected OH-stretching data at a constant T = 550 °C on six samples with thickness ranging between30 and 150 µm. In any case the target temperature was reached as fast as possible (90°C/min rate) and held constant while collecting FTIR spectra every 2 minutes, until the complete disappearance of the OH-signal. Preliminary spectra collected on amphibole powder embedded in KBr disks showed no OH loss even after prolonged heating, therefore the isothermal experiments were performed on pellets consisting of compressed pure powder. The integrated OH intensities as a function of time were fitted using the Avrami equation; for single-crystals, the data showed an initial intensity increase that was fitted testing two different procedures. The resulting parameters were plotted in the Arrhenius space to derive the activation energy (E_a) for the H⁺ diffusion in riebeckite. The final values are: 19.6±1.5 kJ/mol (from powder data), 26±3 kJ/mol or and 34±2 (from single-crystal data, depending on the fitting method). The activation energy for powders is lower than that obtained for single-crystals, and this result supports the model in which the oxidation of amphiboles occurs at the sample surface. Moreover, the E_a obtained here are considerably lower than the values reported in the literature (e.g. Ingrin and Blanchard, 2006) for pure diffusive processes of H₂ and H₂O through several different crystal matrixes. It is also consistently lower that all values reported so far for amphiboles (e.g. Johnson and Fegley, 2003). This can be related to the peculiar deprotonation mechanism in riebeckite where the OH - O²⁺ substitution at the anionic O3 site is coupled to ^M(1)Fe²⁺ - ^M(1)Fe³⁺ oxidation (e.g. Della Ventura et al., 2018, Galdenzi et al., 2018) and the transformation of the phase into an oxi-amphibole.

 

 

References cited

 

Della Ventura, G., Milahova, B., Susta, U., Cestelli Guidi, M., Marcelli, A., Schlüter, J., Oberti, R. (2018) Am. Min., 103, 1103-1111.

Galdenzi, F., Della Ventura, G., Cibin, G., Macis, S., Marcelli, A. (2018) Rad. Phys. Chem., 1, 1-4.

Ingrin and Blanchard (2006) Rev. Min. Geochem., 62, 291-320.

Johnson, N.M., Fegley B. (2006) Icarus, 164, 317-333.