Experimental study of high-Ti and low-Ti basalts: liquid lines of descent and silicate liquid immiscibility in large igneous provinces

Dichotomous high-Ti and low-Ti magmas are ubiquitous in large igneous provinces (LIPs). These magmas often form economically critical Fe–Ti oxide ores in layered intrusions via mechanisms that remain debated. To constrain the evolution of high-Ti and low-Ti basalts during fractionation, we performed stepwise equilibrium crystallization experiments at atmospheric pressure. We specifically aimed to quantify the influences of starting composition and oxygen fugacity (fO2) on phase stability, phase compositions, and the onset of silicate liquid immiscibility during cooling. Both types of magma crystallize similar phase assemblages at QFM (quartz–fayalite–magnetite thermodynamic equilibrium) and QFM + 2: olivine, clinopyroxene, plagioclase, Fe–Ti oxides, and whitlockite. Tridymite crystallizes late in experiments at QFM + 2. The starting composition exerts a strong influence on phase and melt compositions. High CaO and Al2O3 contents in the melt favor the early crystallization of plagioclase and enhance FeO enrichment before Fe–Ti oxide saturation. fO2 affects the composition and stability of Fe–Ti oxides, and high fO2 conditions may promote melt differentiation into the calc-alkaline field. Silicate liquid immiscibility occurs in both compositional trajectories, producing Fe-rich melt globules in equilibrium with Si-rich melts. Strong iron enrichment is not necessary for immiscibility to develop; unmixing also occurs in Fe depleted compositions. We propose a new parameterization to map the binodal surface in temperature-composition space that successfully fits the two-liquid field in experiments and natural immiscible compositions. Our results indicate that Fe–Ti oxide ores in layered intrusions associated with LIPs form by the segregation of Fe-rich melts and/or the accumulation of early crystallized Fe–Ti oxides during fractionation.