A series of laboratory experiments were conducted to examine how partitioning of Fe among solid reaction products and rates of H2 generation vary as a function of temperature during serpentinization of olivine. Individual experiments were conducted at temperatures ranging from 200 to 320 °C, with reaction times spanning a few days to over a year. The extent of reaction ranged from <1% to ~23%. Inferred rates for serpentinization of olivine during the experiments were 50-80 times slower than older studies had reported but are consistent with more recent results, indicating that serpentinization may proceed more slowly than previously thought. Reaction products were dominated by chrysotile, brucite, and magnetite, with minor amounts of magnesite, dolomite, and iowaite. The chrysotile contained only small amounts of Fe (XFe = 0.03-0.05, with ~25% present as ferric Fe in octahedral sites), and displayed little variation in composition with reaction temperature. Conversely, the Fe contents of brucite (XFe = 0.01-0.09) increased steadily with decreasing reaction temperature. Analysis of the reaction products indicated that the stoichiometry of the serpentinization reactions varied with temperature, but remained constant with increasing reaction progress at a given temperature. The observed distribution of Fe among the reaction products does not appear to be entirely consistent with existing equilibrium models of Fe partitioning during serpentinization, suggesting improved models that include kinetic factors or multiple reaction steps need to be developed. Rates of H2 generation increased steeply from 200 to 300 °C, but dropped off at higher temperatures. This trend in H2 generation rates is attributable to a combination of the overall rate of serpentinization reactions and increased partitioning of Fe into brucite rather than magnetite at lower temperatures. The results suggest that millimolal concentration of H2 could be attained in moderately hot hydrothermal systems like Lost City during fluid circulation on timescales of a few years.
Bibliographical noteFunding Information:
This research was supported by the U. S. National Science Foundation Marine Geology and Geophysics program though grant #NSF-OCE 0927744 to T.M.M. and A.S.T and #NSF-OCE 1059534 to FK. Additional support to T.M.M. from the Hanse Wissenschaftskolleg during the project is gratefully acknowledged. N.J. was funded through DFG-Research Center/Excellence Cluster “The Ocean in the Earth System” . The IRM is supported by the Instruments and Facilities Program of the NSF Division of Earth Science . The authors greatly appreciate the helpful comments of Jeff Alt, Dionysis Foustoukos, Benjamin Malvoisin, and an anonymous reviewer on the manuscript. This is IRM contribution 1502.
© 2016 Elsevier Ltd.