Abiotic hydrocarbon formation in serpentinization experiments: network analysis on a comprehensive dataset

Abstract

Serpentinization refers to the hydrothermal alteration of ultramafic rocks (e.g. peridotite) in which ferromagnesian silicate minerals react with water to dominantly precipitate serpentine and (hydr)oxides while producing hydrogen (H2). In natural systems, various amounts of methane (CH4) and other reduced carbon compounds are also observed in fluids issuing from serpentinization environments that range from on-land ophiolites to mid-ocean ridges and subduction zones. Formation of the light hydrocarbons (HC) at least, has been attributed to the abiotic reduction of oxidized carbon sources (mainly CO2) by H2 in those systems. Such abiotic organic reactions has implications for global carbon cycling and may offer the most favorable environments for life origin and development on water-bearing planets. Therefore, a large number of experiments has been performed over the past decades to reproduce H2, CH4, and light HC formation in order to identify controlling parameters. If H2 was easily formed, CH4 and light HC were hardly produced or displayed highly variable concentrations that preclude systematic conclusions. Numerous experimental parameters that vary from one experiment to another render the task even more difficult. Therefore, we built up a dataset that includes experimental setups, conditions, reactants, and products from 30 peer-reviewed papers. Then we applied Pearson correlation algorithm and network analysis to sort out the main experimental parameters controlling HC production during serpentinization. This first application of novel analytical methods on this large dataset allowed the identification of 3 main similarity groups of experiments primarily controlled by T and so by the experimental protocol too. It was then possible to show that CH4 and light HC concentrations in experimental fluids are not correlated with H2 concentrations but are positively correlated with temperature, and to a lesser extent with pressure. Some minor effects of catalysts are also identified for both CH4 and H2 production but still need further experimental investigations. Our results also highlight contamination in some experiments. Continuously adding new experimental data to our available dataset will provide useful constraints on these reactions that could, in turn, be used to guide future experiments.