Isotopic geochemistry and origin of water, carbon and sulfur in volcanic gases: rift zones, continental margins and island arcs

The origin of water vapour, carbon and sulfur in high temperature gases released from sub-aerial volcanoes in different tectonic areas has been investigated from both their chemical proportions and their isotopic ratios. Knowing the actual source(s) of these three main components of volcanic gas emissions may indeed provide valuable information upon the exchanges of volatile matter between the mantle and the Earth's surface, the history of the atmosphere or the genesis of subduction zone magmatism. More than 200 volcanic gas samples were analysed for that study. These were collected at high temperature from different types of volcanism: tholeiitic along the sub-aerial rift axes in Afar, alkaline continental in Italy (Etna) and Antarctica (Erebus), calc-alkaline in subduction zones (Central America, Lesser Antilles, Japan, Indonesia, Kamchatka). The great number of analyses carried out has improved our knowledge of the chemistry of high temperature volcanic fluids. Together with other data previously obtained, the results show that the chemical proportions of water, carbon and sulfur in these fluids vary significantly with the nature of the corresponding magmas: andesitic or dacitic volcanic gases are usually richer in water and poorer in sulfur (H₂O: 90 to 99 mole %, C/S ratio: ∼ 2 to 5) than their basaltic counterparts (H₂O: ∼ 80 mole %, C/S ratio: ∼ 1.5 ± 0.5). Such a trend is coherent with available data on the abundance and the solubility of H₂O, CO₂ and S in magmas of different composition. The identification of their ultimate source(s) can only be achieved using isotopic tracers: ¹⁸O/¹⁶O, ¹³C/¹²C, ³⁴S/³²S, and ⁴He/³He ratios. In most cases the volcanic water vapors appear to derive from recycled crustal waters, of dominantly meteoric origin. Their ¹⁸O/¹⁶O ratio is indicative of a high temperature isotopic equilibrium with the magmas (δ¹⁸O ∼ +5 to +8 o/oo vs SMOW), their D/H ratio commonly lies close to that of local rainwaters and outside the isotopic range presumed typical of mantle hydrogen (δD ∼-80 to -60 o/oo vs SMOW). Moreover, there exists a broad relationship between the deuterium content of the vapors and the latitude of the corresponding volcanic sites. This gives evidence of a strong influence of climatic parameters on the isotopic composition of the volcanic waters, which would not occur if they had a deep-seated origin. A local dilution by sea water is possible at a few insular volcanoes; otherwise, the contribution of mantle-derived water is likely for only two of the volcanoes studied: Erta Ale (Afar) and Tolbachik (Kamchatka). In the latter case the vapor is both similar to water in mid-ocean ridge basalts and distinct from the local meteoric waters. The ¹³C/¹²C and ³⁴S/³²S ratios display more restricted variations and are more closely related to the nature of the tectonic and geologic environment of volcanoes. The carbon and sulfur released by tholeitic volcanoes of the Afar rift system have a typical mantle isotopic signature. They are very similar to the carbon and sulfur produced by submarine volcanism and hydrothermal circulation at mid-ocean ridges. Their isotopic ratio in the gas phase is well representative of their initial ratio in the melt - in spite of significant fractionation effects. So, the similarity of the δ¹³C and δ³⁴S values for sub-aerial volcanism in Afar and submarine volcanism at mid-ocean ridges supports the idea that the distribution of carbon and sulfur isotopes in upper mantle beneath seafloor spreading axes is rather homogeneous. The results obtained on subduction zone volcanoes or alkaline continental volcanoes are more complex. It seems that heavy volcanic CO₂ could originate from the dilution of mantle carbon by ¹³C- and ⁴He-rich crustal fluids, the more so large as the crust is thick (or the travel of magmas through that crust is long). The presence of carbonate xenoliths in lavas from volcanoes which emit the heavy CO₂ supports that model. A contribution from carbon in subducted sediments cannot be excluded; but from data presently available a contamination from the crustal basement of volcanoes seems more likely. In any case, a significant fraction of the carbon emitted by island arc volcanoes would be of crustal origin. This conclusion also applies to sulfur, although in a less systematic way