Variation of plagioclase shape with size in intermediate magmas: a window into incipient plagioclase crystallisation

Abstract

Volcanic rocks commonly display complex textures acquired both in the magma reservoir and during ascent to the surface. While variations in mineral compositions, sizes and number densities are routinely analysed to reconstruct pre-eruptive magmatic histories, crystal shapes are often assumed to be constant, despite experimental evidence for the sensitivity of crystal habit to magmatic conditions. Here, we develop a new program (ShapeCalc) to calculate 3D shapes from 2D crystal intersection data and apply it to study variations of crystal shape with size for plagioclase microlites (l < 100 µm) in intermediate volcanic rocks. The smallest crystals tend to exhibit prismatic 3D shapes, whereas larger crystals (l > 5–10 µm) show progressively more tabular habits. Crystal growth modelling and experimental constraints indicate that this trend reflects shape evolution during plagioclase growth, with initial growth as prismatic rods and subsequent preferential overgrowth of the intermediate dimension to form tabular shapes. Because overgrowth of very small crystals can strongly affect the external morphology, plagioclase microlite shapes are dependent on the available growth volume per crystal, which decreases during decompression-driven crystallisation as crystal number density increases. Our proposed growth model suggests that the range of crystal shapes developed in a magma is controlled by the temporal evolution of undercooling and total crystal numbers, i.e., distinct cooling/decompression paths. For example, in cases of slow to moderate magma ascent rates and quasi-continuous nucleation, early-formed crystals grow larger and develop tabular shapes, whereas late-stage nucleation produces smaller, prismatic crystals. In contrast, rapid magma ascent may suppress nucleation entirely or, if stalled at shallow depth, may produce a single nucleation burst associated with tabular crystal shapes. Such variation in crystal shapes have diagnostic value and are also an important factor to consider when constructing CSDs and models involving magma rheology.