Differences in equilibration rates among crystals of different sizes may be used to deduce paleofluid changes over time if the crystal-growth mechanism is known. To explore isotopic equilibration rates as a function of illite growth, we studied B-isotope changes during illitization of smectite. Montmorillonite (<2.0 μm SWy-1, K saturated) was reacted with aqueous boric acid (1000 ppm B) at 300°C, 100 MPa in sealed Au capsules (1:1 fluid:mineral ratio). The initial fluid was 0‰ (NBS 951 standard) but after R1 ordering occurred (65 days of reaction) the fluid was changed to −7‰ in order to examine the rate of isotopic re-equilibration. Samples were taken intermittently throughout the experiment. Each aliquot was NH4 exchanged and size separated into fine (<0.2 μm), medium (0.2–2.0 μm) and coarse (>2.0 μm) fractions. The isotopic composition of B in the tetrahedral sheet was then measured for comparison with the predicted equilibrium values.
The fine fraction showed equilibrium isotope ratios within 10 days, indicating that small, newly nucleated crystals precipitate in equilibrium with the fluid under supersaturated, closed conditions. These fine-fraction minerals did not re-equilibrate when the fluid was changed. The medium fraction gradually equilibrated with the initial fluid as illite grew to values >50%, but did not re-equilibrate with the later fluid. The coarse fraction was slow to begin recrystallization, perhaps due to dissolution kinetics of large crystals or the presence of detrital contaminants. However, it showed the fastest rate of isotopic change with crystal growth after R1 ordering. We conclude that at 300°C, the initial B–O bonds formed in illite are stable, and isotopic re-equilibration only occurs on new crystal growth. Therefore, different isotope ratios are preserved in different crystal size fractions due to different rates of crystal growth. Large crystals may reflect equilibrium with recent fluid while smaller crystals may retain isotope compositions reflecting equilibrium with earlier fluids.