Chemical maps as a memory of magma solidification: from crystallization onset to trapped melt

In igneous systems, mineral compositions are controlled by the chemistry of the parental melt, thermodynamic conditions, and the kinetics of magma solidification, all of which may vary along the crystallization path and generate chemical zonation. Additional processes such as magma recharge, mixing, and reactive melt percolation in crystal-rich mushes can further induce partial dissolution and overgrowth of minerals, while diffusion-driven re-equilibration can modify initial compositions. Although point analyses provide precise compositional data, they often fail to capture the full spatial complexity of chemical variability. High-resolution chemical mapping partly overcomes this limitation and, when properly interpreted, offers powerful constraints on magma solidification histories.

The use of selected chemical elements with contrasting diffusivities and crystal–melt affinities provides key insights into successive crystallization stages, from early nucleation to the solidification of late interstitial melts. In olivine, phosphorus is a particularly robust tracer: it is preferentially incorporated during rapid growth, and diffuses extremely slowly in the crystal. It also displays slightly incompatible behaviour, leading to its enrichment in evolved melts and late-stage olivine growth. Coupling P with faster-diffusing incompatible elements such as Al allows relative differences in crystal residence histories and storage conditions to be revealed.

Here we present striking examples from volcanic and plutonic settings where high-resolution P and Al maps in olivine reveal magma solidification dynamics. Chemical maps uncover hidden skeletal to dendritic growth morphologies. They record early disequilibrium crystallization followed by morphological ripening toward near-equilibrium conditions and repeated cycles of partial resorption and overgrowth. A clear dichotomy emerges between volcanic autocrysts, characterized by coupled P–Al skeletal patterns, and mush-derived crystals, in which P preserves early growth features while Al is homogenized during prolonged storage. Finally, in crystal-rich domains, the trapping of highly differentiated melt upon porosity closure is commonly quantified using chemical mass balance. Here we show for the first time that specific chemical tracers can identify olivine crystallization from such trapped, highly evolved melts. Mapping these tracers in plutonic rocks thus provides unique constraints on late-stage porosity distribution.