On the origin of alkali feldspar megacrysts in granitoids: the case against textural coarsening

Abstract

Alkali feldspar megacrysts (up to 20 cm in length) are common in granitoids, but their origin has been a matter of continued debate. In particular, the relative role of various magmatic and subsolidus processes has proven difficult to assess given the protracted crystallization history of these megacrysts. Textural coarsening (also known as Ostwald ripening) has been invoked to explain the extraordinarily large sizes of these feldspar megacrysts. While coarsening is a thermodynamically spontaneous process, it is important to evaluate the timescales in which it operates during magmatic evolution. In this work, we consider theories of coarsening of a second phase in the melt in binary systems to evaluate if coarsening can explain the origin of alkali feldspar megacrysts in granitoids. The theories explored are all approximate and somewhat incomplete; however, all of them predict the same basic kinetic equation for the coarsening process. Calculation of the fundamental rate constant controlling the coarsening processes requires determination of parameters that are not particularly well constrained for minerals in silicate melts. Nonetheless, we demonstrate that diffusion rates are the main control on the rate constant for coarsening. Using a range of rate constants calculated using suitable parameters for silicate magmatic systems, we show that coarsening in silicate magmas cannot lead to crystals larger than tens of µm on timescales relevant for igneous processes. The timescales required to generate millimeter to centimeter crystals by coarsening are so prohibitively large that the weaknesses of the existing kinetic theories becomes unimportant, and the idea of generating megacrysts via textural coarsening can be conclusively abandoned. The origin of feldspar megacrysts probably results from the interplay between nucleation and growth rates in magmas, a problem that has not been dealt in detail in the literature.