Chlorine/Fluid Cycling in Subduction Zones: Evidence from Chloride Concentrations and Chlorine Stable Isotopes

Abstract

Pore waters with chloride concentrations less than seawater are a characteristic feature of accretionary complexes in convergent margins. One well-documented case where such freshened fluids have been documented is in the Nankai Trough, Japan, Ocean Drilling Project (ODP) Site 808. The enigmatic profile of pore water chloride concentration versus depth in these sediments is characterized by an extensive section of lower than seawater chloride concentrations between 560 and -1200 mbsf, with a broad minimum of about 450 mM (~20 % seawater dilution) at ~1,100 mbsf (Fig. 1). Newly obtained chlorine stable isotopic ratios (37C1/35C1) in the pore fluids at the Nankai drill site are highly fractionated (-7.4%0) compared to seawater (0%0), with the most extreme fractionation (Fig. 2) coinciding with the lowest pore water chloride concentration (Fig. 1). Chlorine stable isotopes in subduction zone settings can be fractionated by water-rock exchange when 37C1 is preferentially incorporated over 35C1 into the structural (OH) site of diagenetic or metamorphic hydrous minerals (e.g. smectite, illite, chlorite, serpentine, talc, and amphibole) and by diffusion. In the case of mineral uptake, lower temperature minerals fractionate C1 isotopes more highly than those formed at higher temperatures; but their maximum C1 contents are generally an order of magnitude lower (e.g. < 100 ppm in smectite versus several parts per thousand in amphibole. The source of the freshened fluids at Nankai is presently being debated. On the basis of pore water chloride data alone, previous studies have suggested that clay mineral reactions, in particular those involving smectite, are primarily responsible for the freshening observed. These studies attribute the increase in pore water chloride to ~560 mbsf (Fig. 1) to the in situ hydration of volcanic glass and formation of clays and zeolites. The decrease in chloride between 560 and 1100 mbsf is proposed to result from clay mineral (and zeolite) dehydration reactions. Below 1100 mbsf, the formation of hydrous minerals in the underlying volcaniclastic section is invoked. Data from C1 stable isotopes now permit us to evaluate these hypotheses. In the case of simple loss of H20 from smectite interlayers or zeolites, the CI stable isotope ratios of associated pore waters are unaffected. However, reactions that involve smectite or other Cl-bearing hydrous silicate minerals will affect the C1 stable isotopic signature of the pore fluids. Below -560 mbsf, the two profiles are decoupled. The decrease in pore water chloride to ~1100 mbsf is accompanied by a strong decrease in pore water (~37C1; the opposite of what would be expected if this isotopic signature was the result of clay mineral dehydration that involved the destruction of smectite and/or other low temperature hydrous layer silicates. Such a reaction would tend to release 37C1 into the pore waters, not remove it. Likewise, decoupling is evident between the two profiles below 1100 mbsf where hydration reactions are thought to be occurring. The decoupling of pore water chloride and the chlorine stable isotope data between 560 and 1100 mbsf provides important evidence that H20 in freshened fluids at Nankai are coming from mineral reactions occurring in the seismogenic zone. Calculations demonstrate that contributions to the observed freshening in Nankai, due to the release of H20 from smectite interlayers and zeolites in the section drilled can account for only 4 % freshening of seawater, at best, instead of the 20% observed. Neither is there any physical evidence for recent fluid flow along the drcollement (-960 mbsf). Porosity in sediments below the d+collement is higher (40% vs 30% in the overlying sediments) than in sediments above; and St, He, and O isotope data, as well as pore water chloride and sulphate, suggest that freshened pore waters at Nankai originate from repeated transient episodes of fluid incursion from the high