The origin of shoshonites: new insights from the Tertiary high-potassium intrusions of eastern Tibet

Abstract

The shoshonitic intrusions of eastern Tibet, which range in age from 33 to 41 Ma and in composition from ultramafic (SiO₂ = 42 %) to felsic (SiO₂ = 74 %), were produced during the collision of India with Eurasia. The mafic and ultramafic members of the suite are characterized by phenocrysts of phlogopite, olivine and clinopyroxene, low SiO₂, high MgO and Mg/Fe ratios, and olivine forsterite contents of Fo₈₇ to Fo₉₃, indicative of equilibrium with mantle olivine and orthopyroxene. Direct melting of the mantle, on the other hand, could not have produced the felsic members. They have a phenocryst assemblage of plagioclase, amphibole and quartz, high SiO₂ and low MgO, with Mg/Fe ratios well below the values expected for a melt in equilibrium with the mantle. Furthermore, the lack of decrease in Cr with increasing SiO₂ and decreasing MgO from ultramafic to felsic rocks precludes the possibility that the felsic members were derived by fractional crystallization from the mafic members. Similarly, magma mixing, crustal contamination and crystal accumulation can be excluded as important processes. Yet all members of the suite share similar incompatible element and radiogenic isotope ratios, which suggests a common origin and source. We propose that melting for all members of the shoshonite suite was initiated in continental crust that was thrust into the upper mantle at various points along the transpressional Red River-Ailao Shan-Batang-Lijiang fault system. The melt formed by high-degree, fluid-absent melting reactions at high-T and high-P and at the expense of biotite and phengite. The melts acquired their high concentrations of incompatible elements as a consequence of the complete dissolution of pre-existing accessory minerals. The melts produced were quartz-saturated and reacted with the overlying mantle to produce garnet and pyroxene during their ascent. The felsic magmas reacted little with the adjacent mantle and preserved the essential features of their original chemistry, including their high SiO₂, low Ni, Cr and MgO contents, and low Mg/Fe ratio, whereas the mafic and ultramafic magmas are the result of extensive reaction with the mantle. Although the mafic magmas preserved the incompatible element and radiogenic isotope ratios of their crustal source, buffering by olivine and orthopyroxene extensively modified their MgO, Ni, Cr, SiO₂ contents and Mg/Fe ratio to values dictated by equilibrium with the mantle.

Appendix: Analytical methods of elements and isotopes

Major element XRF analyses of the whole rocks samples were analysed by Prof. Bruce Chappell at Macquarie University using the method of Norrish and Hutton (1969), except for samples marked by star symbols (*), which were analysed by wet chemical method at the Guangzhou Institute of Geochemistry, Chinese Academy of Science (GIG–CAS). The rare earth and trace elements concentrations were measured by laser ICP–MS at the ANU, using the method described by Campbell (2003). The glasses for analyses were prepared by mixing finely powered rock with lithium borate flux in the ratio 2:1, heating for 15 min at 1,200 °C, then quenched in water. The absolute concentration of elements was determined by ratioing to Ca using the NIST glass 610 as a standard. Analytical uncertainties are ±1–2 % for major elements, and between ±2–10 % (2σ) for the trace elements, depending on the element. Zircon U–Th–Pb dating was performed at the Research School of Earth Sciences, the Australian National University following the procedure of Harris et al. (2004).