In order to examine the seawater-seafloor sediment interactions that influence the chemical composition of seawater through time, we examined hundreds of pore fluid geochemical analyses from 13 clay-rich sedimentary successions drilled by the ODP-IODP. Chemical trends such as monotonous increases in Ca2+, and decreases in Mg2+ and δ18O with depth are traditionally interpreted to result from water-rock interaction. In this view, the release of Ca2+ into fluids and the uptake of Mg2+ and 18O mainly results from the formation of low-temperature clays in the sediment and within underlying basalts. Chloride concentration profiles and isotopic compositions, however, suggest that different processes may influence pore water geochemistry. The data examined here show relatively constant chloride contents but with a systematic decrease in δ37Cl of chlorides with depth from 0 permil (the seawater value) down to -8.5 permil. The δ37Cl data are highly correlated with δ18O (with δ18O down to -5.7 permil).

The δ37Cl-depletions of pore fluid chlorides are found in all studied sedimentary piles regardless of tectonic or sedimentary history. These trends cannot be explained by water-rock exchange reactions because minerals formed at low temperature have Cl contents that are too low to compensate for δ37Cl depletions observed in pore fluids. Accordingly, we hypothesize that fluid-specific processes are responsible for the δ37Cl-depletions of the fluids and that δ37Cl-enriched chlorides were expelled out of the sediments into the ocean. After reviewing the fluid-specific processes that are known to change the chlorine isotope ratios in chlorides, we rule out diffusion and gravitational isotope fractionations of chlorides could generate this isotope pattern. The flow of a δ37Cl -depleted fluid from the underlying basaltic basement into the sediments could explain the δ37Cl data. But the mechanism that produces depletion in δ37Cl of the fluid remains unknown. It cannot be chloride exchanges between fluids and rocks.

Here we show that compaction-induced ion filtration of chlorides through clay-rich membranes can produce the observed pore fluid δ37Cl-depletions, with isotope fractionation factors ranging from 1.000 to 1.008 between the chlorides of the expelled fluid (the permeate) and those of the residual fluid (the retentate). We find that smectite-rich sediments are associated with higher isotopic fractionation factors, while illite/chlorite-rich sediments are associated with intermediate values and with clay-poor sediments associated with lower values. This suggests that chlorine isotope fractionation might be controlled by surface charge associated with specific clay minerals. Our calculations show that compaction-induced filtration has the capacity to produce 18O-depletion for oxygen isotope fractionation factors between the expelled fluid and the retentate ranging from 1.000 to 1.005. 18O-enrichment in the expelled fluid is in agreement with the experimental data of Haydon and Graf (1986). Overall, although further experimental work on both chlorine and oxygen isotopes is certainly needed, the results of this study indicate that ion-filtration should be considered as a potential mechanism for fractionating isotopic species in sediment pore waters, particularly for oxygen isotope ratios whose variations are often commonly attributed to water-rock exchange.