Hot and juicy - Melt-assisted deformation controls the rheology of the oceanic crust beneath the detachment fault footwall at (ultra)slow-spreading ridges

At (ultra)slow-spreading ridges, oceanic core complexes are commonly interpreted to form through dislocation creep in the lower crust, followed by strain localisation into evolved, oxide-rich domains along detachment faults due to fluid-assisted weakening. New observations from the Atlantis Bank (Southwest Indian Ridge) suggest that this framework underestimates the role of melt-present deformation in controlling lower-crustal rheology.

We investigate gabbroic samples from IODP Hole U1473A (Expedition 360), which penetrated ~800 m of the footwall of the Atlantis Bank detachment system. We developed a semi-quantitative microscale strain intensity classification (grades 0–IV, from undeformed to ultramylonite) and compared it with the IODP shipboard macroscopic fabric classification. We furthermore integrate electron backscatter diffraction, scanning electron microscopy, full thin-section chemical mapping, and in situ major and trace element analyses by laser ablation, allowing deformation, melt–rock interaction, and chemical evolution to be assessed from the macro- to the microscale.

Intracrystalline deformation is pervasive across all samples, including those classified as undeformed at the shipboard macroscopic scale, with no systematic relationship between strain intensity and bulk compositional evolution. Whole thin-section chemical maps reveal strong asymmetric zoning in porphyroclasts, with rims and neoblasts consistently enriched in more evolved compositions and preserving microstructural evidence of melt–rock interaction across all lithologies and strain classifications. In addition to neoblasts and overgrowths, evidence for the former melt presence is manifested by locally elevated modal proportions of secondary phases (pargasite, oxides, and enstatite), low apparent dihedral angles (<60°) between mineral phases, films or thin elongate grains interpreted as pseudomorphs after melt along grain boundaries, and cuspate grain boundaries that affect all phases. These microfabrics occur across the full range of microstructural gradients and rock types, but are most pronounced in higher-strain samples. In situ trace element profiles further confirm that rims and neoblasts are more evolved than their host minerals, marked by enrichment in light rare earth elements.

Although all mineral phases display well-defined crystallographic preferred orientations, these fabrics are not consistently related to known slip systems. Instead, we suggest that deformation is accommodated by stress-controlled precipitation and anisotropic growth, consistent with the observation of asymmetric zoning and neoblasts. Coupled rare earth elements and plagioclase–amphibole geothermometry indicates progressive cooling from ~1150 to ~850 °C, constraining the thermal conditions of melt-present deformation prior to brittle localization. This deformation is pervasive throughout the entire sampled hole interval, with protomylonites and mylonites comprising most of the deformed rocks and occurring in shear zones up to 10 m thick.

These results support a model in which the lower oceanic crust deforms and evolves predominantly by melt-assisted dissolution–precipitation creep, largely independent of bulk composition. Dislocation creep is interpreted as a secondary, local response to high strain rates imposed by melt-enhanced reactions. This process produces substantial rheological weakening at the base of the crust, promoting the initiation and long-term activity of detachment faults in oceanic core complexes.