The mineral dissolution rate conundrum: Insights from reactive transport modeling of U isotopes and pore fluid chemistry in marine sediments

Pore water chemistry and U-234/U-238 activity ratios from fine-grained sediment cored by the Ocean Drilling Project at Site 984 in the North Atlantic were used as constraints in modeling in situ rates of plagioclase dissolution with the multicomponent reactive transport code Crunch. The reactive transport model includes a solid-solution formulation to enable the use of the U-234/U-238 U activity ratios in the solid and fluid as a tracer of mineral dissolution. The isotopic profiles are combined with profiles of the major element chemistry (especially alkalinity and calcium) to determine whether the apparent discrepancy between laboratory and field dissolution rates still exists when a mechanistic reactive transport model is used to interpret rates in a natural system. A suite of reactions, including sulfate reduction and methane production, anaerobic methane oxidation, CaCO3 precipitation, dissolution of plagioclase, and precipitation of secondary clay minerals, along with diffusive transport and fluid and solid burial, control the pore fluid chemistry in Site 984 sediments. The surface area of plagioclase in intimate contact with the pore fluid is estimated to be 6.9 m(2)/g based on both grain geometry and on the depletion of U-234/U-238 in the sediment via alpha-recoil loss. Various rate laws for plagioclase dissolution are considered in the modeling, including those based on (1) a linear transition state theory (TST) model, (2) a nonlinear dependence on the undersaturation of the pore water with respect to plagioclase, and (3) the effect of inhibition by dissolved aluminum. The major element and isotopic methods predict similar dissolution rate constants if additional lowering of the pore water U-234/U-238 U activity ratio is attributed to isotopic exchange via recrystallization of marine calcite, which makes up about 10-20% of the Site 984 sediment. The calculated dissolution rate for plagioclase corresponds to a rate constant that is about 10(2) to 10(5) times smaller than the laboratory-measured value, with the value depending primarily on the deviation from equilibrium. The reactive transport simulations demonstrate that the degree of undersaturation of the pore fluid with respect to plagioclase depends strongly on the rate of authigenic clay precipitation and the solubility of the clay minerals. The observed discrepancy is greatest for the linear TST model (10(5)), less substantial with the Al-inhibition formulation (10(3)), and decreases further if the clay minerals precipitate more slowly or as highly soluble precursor minerals (10(2)). However, even several orders of magnitude variation in either the clay solubility or clay precipitation rates cannot completely account for the entire discrepancy while still matching pore water aluminum and silica data, indicating that the mineral dissolution rate conundrum must be attributed in large part to the gradual loss of reactive sites on silicate surfaces with time. The results imply that methods of mineral surface characterization that provide direct measurements of the bulk surface reactivity are necessary to accurately predict natural dissolution rates. (c) 2005 Elsevier Inc. All rights reserved.