Metasomatism Within the Ocean Crust

Abstract

From ridge to trench, the ocean crust undergoes extensive chemical exchange with seawater, which is critical in setting the chemical and isotopic composition of the oceans and their rocky foundation. Although the overall exchange fluxes are great, the first-order metasomatic changes of crustal rocks are generally minor (usually <10% relative change in major element concentrations). Drastic fluid-induced metasomatic mass transfers are limited to areas of very high fluid flux such as hydrothermal upflow zones. Epidotization, chloritization, and serizitization are common in these upflow zones, and they often feature replacive sulfide mineralization, forming significant metal accumulations below hydrothermal vent areas. Diffusional metasomatism is subordinate in layered (gabbroic-doleritic-basaltic) crust, because the chemical potential differences between the different lithologies are minor. In heterogeneous crust (mixed mafic-ultramafic lithologies), however, diffusional mass transfers between basaltic lithologies and peridotite are very common. These processes include rodingitization of gabbroic dikes in the lithospheric mantle and steatitization of serpentinites in contact to gabbroic intrusions. Drivers of these metasomatic changes are strong across-contact differences in the activities of major solutes in the intergranular fluids. Most of these processes take place under greenschist-facies conditions, where the differences in silica and proton activities in the fluids are most pronounced. Simple geochemical reaction path models provide a powerful tool for investigating these processes. Because the oceanic crust is hydrologically active throughout much of its lifetime, the diffusional metasomatic zones are commonly also affected by fluid flow, so that a clear distinction between fluid-induced and lithology-driven metasomatism is not always possible. Heterogeneous crust is common along slow and ultraslow spreading ridges, were much of the extension is accommodated by faulting (normal faults and detachment faults). Mafic-ultramafic contacts hydrate to greater extents and at higher temperatures than uniform mafic or ultramafic masses of rock. Hence, these lithologic contacts turn mechanically weak at great lithopheric depth and are prone to capture much of the strain during exhumation and uplift of oceanic core complexes. Metasomatism therefore plays a critical role in setting rheological properties of oceanic lithosphere along slow oceanic spreading centers, which – by length – comprise half of the global mid-ocean ridge system.

Appendix

Thermodynamic calculations were used to predict phase relations in ocean floor metasomatism. In fluid-rock interactions, factors controlling the stable mineral assemblages, besides temperature and pressure, are the primary rock composition, the composition of the fluid entering the system, and the mass flux of that fluid through the system. The dependencies of the metasomatic assemblages on fluid flux and rock composition can be examined using simple isothermal and isobaric geochemical titration models, in which rock is added to a constant amount of H₂O. In oceanic crust metasomatism, the composition of the fluid entering the system (oceanic crust) is seawater. The composition of seawater is from Klein et al. (2009) as is the composition of mantle peridotite. Average mid-ocean ridge basalt from the mid-Atlantic Ridge (Klein 2004) was used as a starting composition for the basalt. In the models of diffusional metasomatism, the fluid into which the rock is titrated is seawater that has reacted with one rock type, before it is allowed to react with the other one. For instance, in the steatitization model, seawater was first equilibrated with basalt and the resulting fluid was reacted with serpentinite.

Geochemical reaction path modeling was conducted using EQ3/6 (Wolery and Jarek 2003). The database was compiled for a pressure of 100 MPa using SUPCRT92 (Johnson et al. 1991). Specifics about the database and calculations (solid solutions, activity models, etc.) are provided in Bach and Klein (2009), McCollom and Bach (2009), and Klein et al. (2009).

In all diagrams displaying results of the reaction path model calculations, we plot predicted modes versus a reaction progress number (ξ). ξ is scaled to the amount of rock titrated into the system. This must not be confused with physical water-to-rock ratios and we thus purposely do not report the mass of rock added in the models. The rock mass was chosen merely as an internal parameter to have the model predict the full range of a metasomatic sequence. In all models, ξ =1 corresponds to approximately equal masses of rock and water. In other words, at ξ =1, the intergranular fluid is entirely controlled by the rock added. As ξ approaches zero, the rock is more and more controlled by the composition of the externally buffered fluid (seawater, intergranular fluids in basalt or in peridotite, etc.). In the models of diffusional metasomatism, one can look at ξ as a measure of distance to the interface between lithologies. Small ξ means close to the contact. With increasing ξ we move away from the contact and into the lithology that is being metasomatized under the influence of an externally buffered fluid.