The correlation between decreasing reaction rates of silicate minerals and increasing duration of chemical weathering was investigated for both experimental and field conditions. Column studies, using freshly prepared Panola Granite, produced ambient plagioclase weathering rates that decreased parabolically over 6 years to a final rate of 7.0×10⁻¹⁴ mol m⁻² s⁻¹. In contrast, the corresponding plagioclase reaction rate for partially kaolinized Panola Granite, after reaching steady-state weathering after 2 months of reactions, was significantly less (2.1×10⁻¹⁵ mol m⁻² s⁻¹). Both rates were normalized to plagioclase content and BET surface area. Extrapolation of decreasing rates for the fresh plagioclase with time indicated that several thousand years of reaction would be required to replicate the rate of the naturally weathered plagioclase under identical experimental conditions. Both rates would remain orders of magnitude faster than field weathering rates previously measured for a weathering profile in the Panola Granite.

Additional trends in weathering rates with time were established from a tabulation of previously reported experimental and field rates for plagioclase, K-feldspar, hornblende and biotite. Discrepancies in the literature, produced by normalization of weathering rates with respect to surface areas measured by gas absorption (BET) and geometric methods, were overcome by developing a time-dependent roughness factor. Regression curves through the corrected rates produced strong correlations with time that were similar for the four silicate minerals. The average silicate weathering rate R (mol m⁻² s⁻¹) was described by the power function

R=3.1×10^-13t^-0.61

which was similar to the relationship describing the decrease in the fresh Panola plagioclase with time and suggesting control by transport-limited reaction.

The time dependency of silicate weathering is discussed in terms of processes intrinsic to the silicate mineral and those extrinsic to the weathering environment. Intrinsic surface area, which increases with the duration of weathering, was shown to account for a third of the exponential decrease in the weathering rate shown by the above equation. Other factors, including progressive depletion of energetically reactive surfaces and accumulation of leached layers and secondary precipitates, must explain differences for fresh and weathered plagioclase reacting under identical experimental conditions. Extrinsic controls, including low permeability, high mineral/fluid ratios and increased solute concentrations, produce weathering reactions close to thermodynamic equilibrium under field conditions compared to highly unsaturated conditions during experimental reaction of fresh and weathered plagioclase. These differences explain the time-dependent difference in field and lab rates.