Volatile deficiency in amphiboles: a crystal-chemical perspective

Systematic studies on amphiboles done combining microchemical (EMP+SIMS), structural (SREF) and spectroscopic (FTIR) analysis provided reliable crystal-chemical models to detect and interpret deviations from the nominal amphibole stoichiometry of 2 (OH,F,Cl) pfu. Substitution of H by F increases amphibole stability at both high-T and high-P conditions, but the crucial issue to understand high-T stability is the crystal-chemical role of the oxo component (O2- at the O3 site). Its occurrence must be considered whenever expected from the T and/or fH2O conditions of crystallization. The oxo component is not confined to the pargasite-kaersutite series, and is more frequent than expected in all amphibole compositions. For instance, three end-members of sodic oxo-amphiboles (O2- > 1 apfu) have been identified, of which two crystallised from (Ti-free or Ti-bearing) Mn metasediments and one from late-stage hydrothermal fluids associated with recent volcanism. The oxo component may occur since crystallization or increase during post-crystallization processes, e.g. during eruption. At the short-range scale, it couples with the incorporation of either Ti at the M1 site or Fe3+ and Mn3+ at the M1 and M3 sites, the oxidation of Fe being the main mechanism observed for post-crystallization dehydrogenation processes. The correct exchange vectors and the site partitioning of the relevant cations can now be determined based on structural parameters, and the oxo component can now be straightforwardly identified and accurately quantified using regression equations based on structure modelling. In this way, correct unit formulae with reliable Fe3+/Fetot ratios can be calculated. In magmatic amphiboles (either calcic or sodic-calcic), measured amphibole/liquid partition coefficients (D) for H are rather constant, suggesting a strong control from the water content of the melt, and oxo-amphiboles invariably have higher Ds for HFSE