The crystal mush paradigm: What we learn from mid-oceanic ridges

Marine Boulanger – Laboratoire Magmas et Volcans, France

From Melt to Mush

The global vision of magma plumbing systems is undergoing a progressive, yet significant revolution. Instead of a large melt-filled magma chamber, natural data indicates that magma is rather stored in a more complex, incrementally grown igneous reservoir which mostly consists of a crystalline mush. Such medium appears essential when accounting for the seismic, geochemical, and textural data collected in many volcanic provinces. The consequences of such a shift are both exciting and critical for the VIPS community. My job for this Focus on VIPS contribution is to guide you through my work and our recently published findings on the magmatic systems below mid-ocean ridges, considering and developing on what we know about mush systems; Cumulate Formation and Melt Extraction from Mush-Dominated Magma Reservoirs: The Melt Flush Process Exemplified at Mid-Ocean Ridges, Journal of Petrology.

Magmas (melt ± crystals ± gases) are a mobilizable medium, meaning that they can be readily transported through the crust to potentially reach the surface and build volcanoes. However, the mobilization of a mush en masse is more complex, if not impossible, due to the rigidity of the connected crystal framework through which melt and gas are distributed [1, 2]. Only the fluid phases (liquid and gas) have the ability to migrate through a mush, and the mush rheology is dictated by the crystalline framework [3]. The presence of mush in igneous reservoirs is not a new concept [4, 5], and the recent paradigm shift towards crystal-rich mushy reservoirs does not contradict either the existence of small melt-rich lenses or transient large melt bodies. Yet, most melt differentiation is now suspected to take place in a porous environment formed by crystals of variable nature and compositions. The presence of this crystalline matrix prevents efficient separation of the minerals from the melt phase during crystallization, hence questioning the prevalence of fractional crystallization during differentiation.

Mid-Ocean Ridge Reservoirs

At mid-oceanic ridges (MOR), the first mush-dominated models date from the early ‘90s (Figure 1) [6, 7]. Despite representing about 75% of the total Earth’s magma budget production [8], these magmatic systems are relatively simple compared to continental ones in terms of the composition of both recharged and extracted melts [9]. Crystal-melt reactions (also known as melt-rock reactions, reactive porous flow processes, or melt-mush reactions) have long been suspected to take place within oceanic reservoirs [10, 11]. These reactions, which are favoured in mushy environments, are now considered to be fundamental processes in oceanic systems – in the same way as fractional crystallization [12]. The unique sampling conducted by dredging or drilling along MOR (by the ODP – IODP consortium) provides an opportunity to characterize fresh gabbroic samples and shed invaluable light on these deep magmatic processes [13–15]

Figure 1: Cross section of representative oceanic crust structure at fast-spreading (left) and slow-spreading ridges (right).

Melt-Mush Reactions & Thermodynamic Models

During my PhD, I started to characterize the architecture and evolution of slow-spreading MOR igneous reservoirs and to study the significance of melt-mush reactions for those systems. My initial focus was on the Atlantis Bank oceanic core complex (Southwest Indian Ridge), where I conducted a series of high-resolution textural and in situ geochemical studies of fresh gabbroic samples (Figure 2). These rocks record the early emplacement and evolution of the oceanic crustal reservoirs and the widespread occurrence of melt-mush reactions at all stages, such as during magma intrusion within the mush [16]. By combining my observations at the crystal scale (Figure 3) with the structure of the magmatic systems at depth, I developed a conceptual model for the formation and evolution of a slow-spreading ridge igneous reservoir. This model is the first to consider the impact of melt-mush reactions at the scale of the entire system [17], and confirms their significance for melt differentiation.

$Left - Bathymetric map of the Atlantis Bank oceanic core complex located along the Atlantis II fracture zone on the Southwest Indian Ridge (purple to red scale represents an elevation difference of 5000 m). Middle – IODP Expedition 360 logo (2015-2016). Right – Thin section scan of olivine gabbro sample ODP 735B 79R6, 96-100(B) cm.$

Figure 2: Left – Bathymetric map of the Atlantis Bank oceanic core complex located along the Atlantis II fracture zone on the Southwest Indian Ridge (purple to red scale represents an elevation difference of 5000 m). Middle – IODP Expedition 360 logo (2015-2016). Right – Thin section scan of olivine gabbro sample ODP 735B 79R6, 96-100(B) cm.

Figure 3: Photomicrographs in plane polarized light of characteristic interaction textures identified in Atlantis Bank gabbros. Plagioclase with irregular contacts with surrounding grains (a-c, e, h) and complex zoning patterns (b). Symplectitic or intergrown clinopyroxene grains (a, d) and clinopyroxene oikocrysts engulfing resorbed plagioclase or olivine chadacrysts (a,e,g,i). Olivine grain with typical heavily rounded and irregular shapes and irregular grain boundaries suggesting resorption (b, f, i). Scale bar: 500 µm.

The key question becoming gradually more important for the community is whether these melt-mush reactions (based on the description of natural samples) are thermodynamically viable. Most reactions described in the literature involve the main mush-forming minerals and more primitive melts compared to the liquids from which these minerals crystallize [18]. In our recent contribution, we provided the first thermodynamic data of this kind. I used the Assimilation – Fractional Crystallization (ACF) mode of the Magma Chamber Simulator (MCS) [19, 20] to test the feasibility of typical oceanic melt-mush reactions (Figure 4). The MCS is an open-source numerical model which relies on the rhyolite-melt engine, and allows tracking the enthalpy and both compositions and temperatures of up to three components in a magmatic system (the main magma subsystem, a subsystem assimilated by the magma and a recharge subsystem). The results show that not only are the reactions feasible, but they can also significantly impact mush minerals, as up to 20% of these can be assimilated during the modelled reactions. These results are only the beginning of what remains to be discovered about these processes [21].

Simplified view of a mush at the crystal scale where melt flush occurs. This figure shows the progressive removal (+/- mixing) of interstitial evolved melt (Ln+1) from the crystal matrix replaced by the primitive melt (Ln) that partially assimilates the minerals before/together with crystallization of a new generation recording both the average more primitive character of the melt and chemical evidence for RPF. — Figure 4: Simplified view of a mush at the crystal scale where melt flush occurs. This figure shows the progressive removal (+/- mixing) of interstitial evolved melt (Ln+1) from the crystal matrix replaced by the primitive melt (Ln, *Magma* in the MCS) that partially assimilates the minerals before/together with crystallization of a new generation recording both the average more primitive character of the melt and chemical evidence for RPF.

Melt-mush reactions at the reservoir scale: the melt flush as a new cumulate forming process

Cumulates are magmatic rocks that form at depth within the reservoirs and provide valuable insight into mushy systems. They are characterized by a refractory composition that is interpreted as reflecting the extraction of a melt phase from the minerals that crystallize during differentiation. Various processes can account for the formation of cumulates such as melt buoyancy, crystal settling, filter pressing, viscous compaction, crystal repacking, and compositional convection. However, these processes are limited by the size of the systems and the duration of the magmatic activity in oceanic mushy reservoirs, especially at slow-spreading ridges [22, 23]. Textural evidence is also lacking for the occurrence of these processes, e.g. widespread crystal-plastic deformation would be expected if compaction played a prominent role in cumulate formation [24]. To overcome these discrepancies, whilst considering what we’ve learned from melt-mush reactions so far, we propose a complementary cumulate forming process called the melt flush [18]. The melt flush is the progressive replacement of the interstitial melt within a mush column by a less evolved melt (with potential partial mixing), leading to reactions between the mush minerals and the crystallization of more refractory ones. Hence in this model, not only melt extraction from the mush but also melt recharges account for the progressive acquisition of a cumulative geochemical signature (e.g. depleted in incompatible elements, Figure 5). The melt flush is therefore a new first order process that can explain cumulate rocks formation.

Figure 5: Schematic diagram of the melt flush process and summary of the processes occurring during the formation of igneous rocks from a petrologic perspective. Continuous and cyclic replenishment refers to oscillatory melt replenishment cycles by melt intrusion that vary through time.

The melt flush process relies on the three main characteristics of (oceanic) magma plumbing systems:

Mush-dominated during melt differentiation
Repetitive recharges of primitive mantle melts
Widespread melt-mush reactions occurring.

One can wonder how relevant and applicable this process is to magmatic systems beyond MOR igneous reservoirs. Since for example continental reservoirs are both commonly composed of mush and likely grow incrementally by primitive melt recharges, we expect that the melt flush process may be viable in numerous igneous systems. A lot of work remains to be done to detail our understanding of melt-mush reactions, from their quantification and kinetics in oceanic systems to their very consideration in continental reservoirs.

Marine Boulanger (marine.boulanger@uca.fr) started her work on mid-oceanic ridge magmatic systems during her PhD at the CRPG lab of Université de Lorraine (France) and at Institut für Mineralogie of Hannover (Germany) with Lydéric France & Juergen Koepke. Before arriving at Laboratoire Magmas et Volcans (Clermont-Ferrand, France) in 2022, Marine was an IODP-France postdoctoral research fellow at Université de Montpellier. Her research combines high-resolution textural and (in situ) geochemical study of natural samples and experimental petrology.

Overall, her research is interested in crystalline mush and the impact of melt migration on differentiation, especially in understanding the significance of crystal-melt interactions on the magmatic systems.

Get in contact with Marine via email and Twitter @marinebou_FR, or through Research Gate and Scholar.

References:

Would you like to be featured on our blog?

We’re on the hunt for fellow VIPS enthusiasts to share their stories, whether you’re a student, an early career researcher (ECR) or more experienced researcher – we want to hear from you!

Do you have a new paper to tell the world about? Or have you been involved in some exciting fieldwork recently? Get in touch at info@vipscommission.org to tell us what’s going on in your field of VIPS.

Share this post:

X (Twitter) Facebook Pinterest LinkedIn E-mail