Crystal size distribution analysis of plagioclase in experimentally decompressed hydrous rhyodacite magma

This study presents crystal size distributions (CSD) of plagioclase forming during decompression experiments of hydrous rhyodacite magma. Samples were annealed at 130 MPa, subjected to continuous decompression at either 2 MPa h−1 or 0.5 MPa h−1, and then quenched along their respective decompression paths at ~ 20 MPa intervals. Samples display concave-up curved CSD plots which result from a combination of two crystal populations: those which formed during the anneal period and those which formed during decompression. The CSDs also display a decrease in crystal number density at 3–7 μm, a size range that is easily resolvable with the imaging strategy employed. The downturn at small sizes is presumably due to insufficient compensation for the intersection probability effect in converting 2D measurements to 3D size distributions. Crystal nucleation and growth rates derived from CSDs using standard assumptions are compared with values obtained using 2D measurements of bulk crystal populations (batch methods). CSD-calculated nucleation rates are substantially low relative to the batch values; in fact, CSD-derived volumetric nucleation rates may underestimate actual nucleation rates by up to two orders of magnitude. In contrast, growth rates from CSDs are consistently higher than batch rates. Although plagioclase growth rates are relatively constant during decompression at a given rate, the average growth rate in the rapidly decompressed series is approximately five times faster than the crystal growth rate in the slowly decompressed series. Because crystal growth rate depends on decompression rate, CSDs are incapable of revealing decompression timescales or magma ascent rates without independent knowledge of crystal growth rate.

Research Highlights
► CSD- and batch-methods do not generate the same nucleation and growth rates.
► Feldspar growth rates remain relatively constant during constant-rate decompression.
► Crystal growth rates are dependent on decompression rate.
► CSDs cannot resolve ascent rate without independent knowledge of crystal growth rate.