Hydrogen isotope determination by TC/EA technique in application to volcanic glass as a window into secondary hydration

International audienceThe use of volcanic glass as recorder of paleoenvironmental conditions has existed for 30 years. In this paper we investigate the methodological aspects of the determination of water content, isotopic composition, and water speciation in volcanic glass using the High Temperature Conversion/Elemental Analyzer (TCEA) mass spectrometer system on milligram quantities of glass concentrates. It is shown here that the precision and the reproducibility of this method is comparable to off-line conventional methods that require 100 times greater amount of material (δD ± 3‰; [H2O]tot ± 10relative% if 1 wt%) but is quicker and permits easy replication. This method extracts 100% of the water as verified by FTIR measurements. Finally, this study confirms the interest of DRIFT spectroscopy in the NIR range for the study of porous samples such as volcanic pumices and tephra, to determine the water speciation (H2O/OH). It may complement conventional FTIR transmission measurements in the MIR or NIR range that usually require homogeneous transparent sections or high degree of sample dilution in a non-absorbing matrix. Using these methods, we attempt to discriminate residual magmatic from secondary meteoric water in volcanic glass. Using mafic to differentiated samples from different geological settings and different climatic conditions, we show that the H-isotope composition and water content of volcanic glass alone are not always sufficient to provide clear distinction between magmatic and meteoric origin. However if the magma is known to have a δD between - 90‰ and - 40‰ (- 60‰ for MORB mantle source), it is quite easy to resolve the δD evolution during magmatic degassing from post-depositional rehydration by meteoric water with δD - 20‰. Water speciation measurements may provide additional information. In most cases, isotopic and total water measurements should be complemented by characterization of water speciation. During magmatic degassing (from 6 wt% to 0.1 wt% water) the H2O/OH is expected to decrease from 2 to close to 0. However, our dataset suggests that during secondary glass hydration (from 0.1 wt% to 6 wt% water) the H2O/OH ratio decreases from 5 to 2, which is the complete opposite. Overall our results support the use of H-isotopes of volcanic glass to discuss the composition of meteoric waters and paleo-climate within a specific region. To this purpose, the volcanic glass has to be almost fully rehydrated in order to fingerprint the isotopic composition of the ambient environmental water. As rehydration is exponentially faster with increasing temperature, efficient rehydration taking months to years, may occur in a cooling volcanic deposits that are meters-thick and thus can remain at a few hundred °C for a years to hundreds of years after the eruption. Such deposits could then provide a snap-shot view of climatic conditions at the time of the studied eruption