Mixing, fluid Infiltration, Leaching, and Deformation (MILD) processes on the slab-mantle wedge interface at high T and P conditions: Records from the Dalrymple Amphibolite, Philippines

Abstract

The whole rock compositions of the blocks and the surrounding matrix of the Dalrymple Amphibolite are investigated in this study to determine the protolith of the blocks and the effect of mechanical mixing and fluid infiltration in the matrix of this fossil slab-mantle wedge interface. The major and trace element contents of the metamafic blocks indicate their mid-oceanic ridge basalt origin similar to the mafic lavas of the crustal section of central Palawan Ophiolite. Similarities in their rare earth and trace element patterns indicate the genetic relationship between the mafic lavas of the Palawan Ophiolite and the metamafic blocks of the Dalrymple Amphibolite. This confirms that the metamafic blocks represent the basalt to gabbro section of the oceanic lithosphere of the subducting slab. The matrix surrounding the blocks exhibit highly variable phase assemblages. In order to determine its petrogenesis, we distinguished groups of components/elements which behave similarly (Group 1 TiO₂, Al₂O₃, Zr, Th and the light rare earth elements; Group 2 Cr, Ni and MgO) based on statistical (correlation coefficient) analyses. These groups indicate mixing of metasedimentary (Group 1) and metamafic (Group 2) end-members to form the matrix. The mixing proportions of the end-members were estimated by employing regression analysis wherein the measured concentration of fluid immobile elements (Cr, Ni, Zr, TiO₂ and Al₂O₃) in the matrix samples were fitted against a modelled concentration by changing the end-member and their relative proportions. The end-members and mixing ratio with the highest regression value (r²) was selected to obtain the modelled composition of the matrix. The modelled and the measured matrix compositions were then used as the original (unmetasomatized) rock and the altered rock respectively in the isocon analysis, assuming that TiO₂, Al₂O₃, Cr, Nd, Zr, and Hf are immobile. This assumption is supported by the prevalence of kyanite, ilmenite and zircon in the matrix mineral assemblage. This procedural workflow helped distinguish end-member components, estimate their mixing ratios, and determine the effects of infiltrating fluids. In particular, the whole rock composition of the matrix was controlled by mixing of a subordinate amount of metamafic blocks in a metasedimentary-dominated shear zone. This is supported by the Cr-Nb content of rutile grains included in the matrix samples which indicate mixed metamafic and metapelitic signatures. The metamafic-metasedimentary dominated matrix in the Dalrymple Amphibolite contrasts with other high-pressure/temperature (P/T) type metamorphic terranes which are dominated by low T minerals (serpentine, Mg-chlorite, and talc) derived from an ultramafic end-member, and could be reflective of conditions in warmer subduction zones. Mass balance calculations further revealed that an early fluid infiltration event likely occurred following the mixing process. This preferentially leached out elements which are either fluid-mobile (e.g. CaO and SiO₂) or are not incorporated into the growing minerals in the matrix. The strong control of mineral assemblage of the matrix in its chemistry is exhibited by a number of samples which showed variable degrees of losses and gains in elements traditionally interpreted to be fluid immobile (e.g. heavy rare earth elements and Y). A later hydration event linked to retrograde metamorphic stage imprinted gains of K₂O, Rb, and Ba in the matrix samples with the growth of replacement minerals (e.g. muscovite on kyanite). This later fluid infiltration event possibly masked the original loss of these fluid-mobile elements in the matrix samples during the earlier fluid-rock interaction.