Abstract
This paper presents the results of 4 reactive percolation experiments set up for investigating the impact of flow rate on serpentinization reaction paths for conditions relevant of the oceanic peridotite sub-seafloor during the initial stages of its hydrothermal alteration. The experiments consisted in injecting artificial seawater into porous compressed olivine powder cores at constant flow rates Q: 0.24, 0.48, 1.14 and 5.21 mL·h−1. The experiments were conducted at constant temperature (170 °C) and pressure (25 MPa) and lasted 11 to 28 days. At the end of the experiments, the outlet fluids composition displayed similar compositions, buffered by the formation of serpentine (aMg2+/a(H+)2 = 9.7–10; aSiO2 = −3.9 to −5.2; pH in situ = 6.1). These values were achieved in a few to up to 300 h for the high flow rate experiment suggesting that they corresponded to a steady-state regime of mass transfer which depended on flow rate. Differences in the composition of fluid versus time and in the structure of reacted samples during and after the four reactive percolation experiments suggested also various incipient serpentinization reaction paths. The low Q experiments produced SiO2(aq) enriched outlet fluids and nodular aggregates were identified covering the reacted olivine surfaces. During high Q experiments, fibrous filaments of proto-serpentine were formed on the olivine surfaces and the fluids progressively achieved steady state compositions similar to the other experiments. These results together with those of previously published reactive percolation experiments lead us to propose two end-member reaction paths for incipient serpentinization of olivine-dominated permeable rocks infiltrated by seawater derived hydrothermal fluids: (1) a transport-controlled reaction path occurring in diffusion dominated zones is characterized by transient brucite precipitation, which produces Mg trapping and Si release in solution, followed by serpentine precipitation and (2) a kinetics-controlled reaction path occurring in advection dominated zones where transport conditions are favorable to Mg leaching and where serpentine precipitates first. The occurrence of these two end-member reaction paths is determined locally by the composition of the fluid, which varies along flow paths. Thus, both reaction paths can coexist in the sample depending on the local pore geometry. Our study shows that the interplay between fluid transport and reaction kinetics controls the chemical fluxes between the mineral surface and the bulk solution, and the incipient serpentinization reaction paths. In natural systems, the scale and distribution of these reaction domains will depend on the complex structure of the ultramafic basement. Our results suggest that the precipitation of serpentine and silica rich phases will be favored in fluid focusing zones such as faults and fractures, whilst formation of brucite will preferentially occur as part of pervasive background serpentinization.