The project aimed at understanding and quantifying the processes that lead to melt-crystals phase separation in crustal magmatic systems. It is well recognized now that crustal magma bodies spend most of their lifespan in a mushy state with roughly equivalent solid (crystals) and melt fractions. Yet, the dynamics of compaction (reduction of pore space by expulsion of melt) is only partially understood for very low (viscous creep in the mantle) and high melt fractions (settling). The proposal here sought to address this gap of knowledge through a combination of laboratory experiments (analog), field work (finding a natural example that can be analyzed) and theoretical-numerical work (dealing with the physics underlying compaction).

The first publication from this project established a new analog experimental framework with silicon oil and glass beads to study a compaction process that we refer to as repacking, whereby grains are not deformed, but find a more efficient packing that reduces the pore space. This is done by rotation and translation of crystals with respect to their neighbors. In Hoyos et al. (2022) we present the results of these experiments, specifically looking into the relative contribution of repacking as a continuous global process interspersed with discrete local force chain disruption events. We find that the crystal shape and sizes controls the frequency and magnitude of these force chain disruption events, this becoming increasingly important for small reservoirs (of a scale comparable to a few crystals in dimensions). For greater spatial domains, as the ones expected in nature, repacking mostly acts as a continuous process and therefore can be modeled from a continuum perspective.

The second important task that was accomplished with this proposal was to use these new experiments in conjunction with published experiments of phase separation to develop a physical model for compaction by repacking, inspired by the literature on hydrogranular dynamics. The few compaction experiments  available provide a picture where the compaction rate decreases by several orders of magnitude and abruptly as the melt fraction decreases below about 0.3. We developed a theoretical model of compaction by repacking using two-phase flow theory and found that repacking controls the compaction rheology at melt fractions above what is expected to be the maximum packing.  Moreover the divergence of the repacking rheology at the maximum packing forces a change of rheology to grain boundary (melt enhanced) diffusion creep. This work will be submitted for publication in January 2024 and reconciles compaction rates established by experiments over a wide range of melt-solid fractions.

The third arch of the proposal is to use a natural laboratory for melt-crystal separation. We combine geochemical analyses on samples from the field site with textural analyses and physics-based compaction models to assess whether significant melt extraction has taken place, quantify it and test what process(es) dominated the melt extraction and how much time it may have taken. We conducted field work on a Miocene silicic intrusion, the Spirit Mtn Batholith (Nevada), an exhumed system that offers a well-documented cross-section through a silicic pluton. Our results indicate that melt loss was quite extensive especially in the bottom half of the system, with up to 30% per mass lost and accumulate up top. This is accompanied with geochemical gradients that support melt separation and accumulation near the top of the system. Textural analyses shows the development of fabric perpendicular to paleo-gravity associated with melt loss, supporting the role of repacking on melt extraction and that gravity was the main force driving it. The minimum trapped melt in the bottom half of the system coincides with the inferred maximum packing similarly to what was observed at the Peninsular Range Batholiths (CA) by Lee et al. (2015). This further supports our inference that repacking dominates compaction in crustal mushes and that the transition to viscous compaction is generally not observed because the compaction rate becomes too slow compared to the thermal lifespan of these intrusions.

Last Modified: 01/02/2024
Modified by: Christian Huber