Based upon dissolution of feldspars under controlled laboratory conditions, we conclude that Sr release, at pH 3, is neither consistently stoichiometric nor constant for the feldspars measured. Bytownite, microcline, and albite all initially release Sr at rates which are 5 (bytownite) to 160 (microcline) times faster than steady-state release rates. The Sr/Si ratios in the early effluents are significantly elevated compared to the bulk mineral values. The 87Sr/ 86Sr measured in effluent early in dissolution is higher than the bulk mineral 87Sr/ 86Sr for bytownite, but lower than bulk mineral ratios for microcline and albite. 87Sr/ 86Sr ratios for the feldspar powders also changed markedly during dissolution of the three phases. In part, nonstoichiometric release of Sr can be explained by the presence of secondary phases (exsolution lamellae or minute quantities of accessory phases) or by surface leaching. Although we infer that these feldspars eventually release Sr with isotopic composition roughly equal to that of the bulk mineral at steady-state, the feldspars dissolve at extremely different rates (bytownite releases Sr at a steady-state rate ∼10 2 to 10 3 times faster than albite and microcline, at pH 3). Therefore, a mixture of these feldspars, or of other minerals exhibiting vast differences in dissolution rate, will release 87Sr/ 86Sr ratios distinctly different from the bulk whole rock. In addition, initial Sr release rates of the minerals (bytownite > microcline > albite) differ from steady-state release rates (bytownite > albite > microcline), complicating analysis of weathering solutions. Log (rate constants) for bytownite, albite, and microcline decrease from −13.5 to −16.4 to −17.2 (mol Sr cm −2 s −1). Interpretation of catchment scale riverine 87Sr/ 86Sr ratios on the basis of whole-rock Sr isotopes is, therefore, problematic at best, and would require normalization of bulk isotopic ratios by relative rates of dissolution of Sr-contributing phases. We also argue that abraded feldspar particles formed naturally, for example, during glaciation, will show this initial transient nonstoichiometric release. However, once the transient release is completed (perhaps 10 2 to 10 3 yr after abrasion), as long as the solution chemistry remains relatively constant, stoichiometric release of cations from feldspars, including Sr, is expected. The most likely way, therefore, to increase riverine fluxes of major cations or radiogenic 87Sr/ 86Sr is to create highly flushed (and, therefore, far from equilibrium) water-rock systems such as glacial sediments and soils, with reactive minerals (e.g., carbonate, plagioclase, or biotite) containing significant radiogenic strontium.
Read full abstract