ALBERT

Description: The presence of structural OH in amphiboles in excess of the usual two OH per formula has been debated for over 40 years (Gier et ah,1964; Leake et ah,1968). However, the reality of the excess-OH phenomenon is still an open question, because accurate water analyses of amphiboles are rarely available. In this study, we review the data available on the chemically simple synthetic system Na2O—MgO-SiO2-H2O (NMSH) and present new results from NMR, infrared spectroscopy, and X-ray-diffraction that allow re-interpretation of previous studies of NMSH amphiboles along the pseudobinary join between the two end-member compositions Na2Mg6Si8O22(OH)2 and Na3Mg5Si8O21(OH)3.We show that there is extensive solid solution involving excess H at 650—750°C, but also document the presence of a wide miscibility gap below 600°C. This miscibility gap is defined by amphiboles very close to the end-member composition Na3Mg5Si8O22(OH)3 coexisting with amphiboles with compositions near the ‘normal’ Na2Mg6Si8O22(OH)2 end member.We also report the characterization of triple-chain silicates (TCS) in the NMSH system and their phase relations with NMSH amphiboles. The upper thermal stability field of the key TCS Na2Mg4Si6O16(OH)2relative to its decomposition to two NMSH amphiboles with a combined equivalent composition has been determined and a pronounced backbend of the transformation boundary documented. Phase relations observed in synthesis experiments suggest that at 550—650°C all TCSs have compositions close to Na2Mg4Si6Oi6(OH)2. Infrared spectroscopy indicates that the TCS synthesized on this composition, studied in detail here, vary from end-member Na2Mg4Si6Oi6(OH)2 to binary solid solutions with less than ∼6 mol.% clinojimthompsonite component. No clear spectroscopic evidence for a ‘Drits’ component NaMg4Si6Oi5(OH)3 (Drits et al.,1975) has been found. Analysis of H2O by vacuum extraction and Karl-Fischer titration indicates large excesses of H2O in all the TCSs studied here that clearly exceed the amounts expected from (OH) groups alone. Thermogravimetric analysis (TGA) and differential thermal analysis (DTA) indicate that this excess H2O is structural. We propose that the excess H2O is likely to be molecular H2O located in the ^4-site channels. The observed backbend of the triple-chain decomposition curve is in agreement with a reaction involving dehydration and loss of this molecular H2O. However, the absolute amount of analysed molecular H2O exceeds that expected from the change in Clapeyron slope alone.While demonstrating the reality of excess OH in amphiboles, the evidence presented in this paper also points to interesting avenues for future research on both amphiboles and TCSs, such as understanding the dynamics and enhanced crystal chemistry of excess OH and molecular H2O in pyriboles.