Crystal settling is ubiquitous, particularly in magmatic reservoirs supplied by mafic magma. A variety of mechanisms have been proposed for the formation of the crystal layers that are thought to accumulate at the bottom of a magma chamber (growth or reprecipitation of olivine crystals while in suspension with a melt, followed by mechanical settling, compaction through pressure dissolution at grain contacts and expulsion of melt).
In this work (Mourey et al., in prep), X-ray microcomputer tomography (X-μCT) was used to investigate the 3D distribution of melt inclusions, vapor bubbles and spinel in several Hawaiian olivine grains.
We collected high-resolution X-μCT scans of different populations of olivine crystals (with different morphologies) from the 1820 eruption at Kīlauea Volcano (Hawai‘i). We document the internal complexities (melt inclusion, vapor bubble and spinel fractions) of well-faceted, polyhedral crystals and more skeletal crystals and interpret those complexities to understand crystal growth conditions.
This study test how different olivine morphologies affect settling velocities and understand the timescales for the formation of olivine-rich mush and dunite cumulates. We performed numerical simulations coupling computational fluid dynamics and a discrete element method in order to investigate the effect of crystal morphology on settling rate.
Figure modified from Mourey et al., 2024: Numerical simulation of the flow of melt around an olivine crystal. The black box corresponds to the limit of the computational domain. The two gray surfaces are the mass inflow and pressure outflow boundary conditions. The crystal is located at the center of the domain. The curves located within the domain are the streamlines of the liquid and their color depends on the magnitude of the flow velocity.
We find that the settling velocity is higher when the elongated crystallographic axis is aligned with the flow of the melt. We evaluate the time of the olivine crystals suspended in a reservoir undergoing vigorous convection. Suspension times vary with the initial particle volume fraction, but in general, olivine crystals have short (<5 years) suspension times in a convective layer.
Figure modified from Mourey et al., 2024:
Time evolution of the particle volume fraction (ϕ) in suspension in the convective melt layer as a function of the crystal morphology. We consider a convective layer (in blue), in which convection occurs and with particle volume fraction ϕ = 0–0.5, and a locked layer (in red), in which no convection occurs (with ϕ > 0.5). The black circles represent the particles. The red arrows indicate the convection. The green dashed line indicates the limit between the convective and locked domain. The suspension time (in years) of olivine crystals in the convective layer is a function of the particle volume fraction (ϕ), the crystal morphology.
I carried out controlled, high temperature laboratory experiments aimed at crystallizing olivine in a basalt melt. The first question we sought to answer was: how fast do olivine crystals we often find in basalt lavas (~1mm or more in size) grow in magma? Does it take years, as often proposed? Our cooling experiments formed olivine of this size in just a few hours, demonstrating that natural crystals may not have such a protracted subsurface history of growth and residence in magmas prior to eruptions.
Figure (adapted from Mourey and Shea, 2019): 3D rendering of olivine crystallization experiments after (a) 60 degrees of cooling and (b) 40 degrees of cooling. Both experiments were left for 6 hours to crystallize. These experiments demonstrate that there are clear preferential growth directions for olivine, and more importantly, that mm-sized phenocryst 'skeletons' can form in just a few hours. (c) Interpretation of the morphological development of the olivine 'endo-skeleton' and its open melt inclusions.
If interested in learning more, take a look at Mourey and Shea (2019).