LAUSR

Continental collision zones have been proposed as primary sites of net continental crustal growth. Therefore, studies on syn-collisional granitoids with mafic magmatic enclaves (MMEs) are essential for testing this hypothesis. The Baojishan (BJS) and Qumushan (QMS) syn-collisional plutons in the North Qilian Orogen (NQO) on the northern margin of the Tibetan Plateau have abundant MMEs in sharp contact with host granitoids, sharing similar constituent minerals but with higher modal abundances of mafic minerals in MMEs. The QMS host granitoids have high Sr/Y and La/Yb ratios showing adakitic compositions, different from the BJS granitoids. Based on bulk-rock compositions and zircon U-Pb age dating, recent studies on these two plutons proposed that MMEs represent cumulates crystallized early from the same magmatic system as their host granitoids, and their parental melts are best understood as andesitic magmas produced by partial melting of the underthrusting upper ocean crust upon collision with some terrigenous sediments under amphibolite facies. Here, we focus on trace element geochemistry of the constituent mineral phases of both MMEs and their host granitoids of the QMS and BJS plutons. We show that different mineral phases preferentially host different trace elements, e.g., most rare earth elements (REEs and Y) reside in titanite (only found in the QMS pluton), amphibole, apatite, epidote and zircon (mostly heavy-REEs), and high field strength elements (HFSEs) reside in biotite, titanite, amphibole and zircon. Based on the mineral chemical data, we testify that for these two plutons, MMEs are of similar cumulate origin, crystallized from primitive andesitic melts in the early stage of granitoid magmatism. The primitive andesitic melts for these syn-collisional granitoids are most likely produced by partial melting of the oceanic crust, supporting the hypothesis of continental crustal growth considering the syn-collisional granitoids represent juvenile continental crust. As evidenced by distinct mineral compositions, the two plutons have different parental magma compositions, e.g., higher TiO2 content, higher Sr/Y and La/Yb ratios in the QMS parental magmas, a signature best understood as being inherited from the source. The higher TiO2 content of the parental magma for the QMS pluton leads to the common presence of titanite in the QMS pluton (absent in the BJS pluton), crystallization of which in turn controls the trace element (REE, Y, Nb, Ta and others) systematics in the residual melts towards an adakitic signature. Therefore, parental magmas with high TiO2 content and high Sr/Y and La/Yb ratios, as well as their further fractionation of titanite, are important factors in the development of adakitic compositions, as represented by the QMS host granitoids. This model offers a new perspective on the petrogenesis of adakitic rocks. The present study further demonstrates that in general, mineral chemistry holds essential information for revealing the petrogenesis of granitoid rocks.