The contributions of (meta-)sedimentary or granitic orthogneissic sources to the Cenozoic Himalayan granites

The Cenozoic Himalayan leucogranites are generally regarded as the representative S-type granites, derived from partial melting of (meta-)sedimentary rocks in the Higher Himalayan Sequences. This interpretation is challenged by the increasing finding of abundant relict zircons with Cambrian–Ordovician ages in the granites. These ages are lacking in the assumed (meta-)sedimentary sources but are close to the formation timing of early Paleozoic granites (presently as orthogneisses) in the Himalaya. Therefore, it is unclear how the early Paleozoic relict zircons were incorporated into the granites and what the real sources of the Cenozoic Himalayan leucogranites are. This study presents U–Pb ages, trace-element, and Lu–Hf isotope compositions of zircon from both the Cenozoic granites and their country rocks (Paleozoic orthogneisses) from Kuday in the Sakya dome, central Himalaya. Our results indicate that the autocryst zircons from the gneiss samples formed at 494–499 Ma and display obvious negative correlations in the Hf–Ti, Hf–Th/U, and Hf–Eu/Eu* binary plots. Their ε Hf (420 Ma) values have relatively restricted variations of −7.2 to −0.5. In contrast, the early Paleozoic relict zircons from granites form scattered and contrasting fields in these binary plots with largely varied U–Pb ages (ca. 410–520 Ma) and ε Hf (420 Ma) values of −35.4 to + 6.7. These observations suggest that the autocryst zircons from gneisses are a cogenetic population and their compositional variations are controlled by crystallization differentiation, whereas the relict zircons from granites were probably originated from sedimentary rocks in which detrital grains were weathered from a variety of protoliths. We argue that the early Paleozoic relict zircon population was not incorporated into the Cenozoic granites by assimilation of the orthogneissic country rocks. This is supported by whole-rock Nd–Hf isotope analyses, which yield ε Nd (t) and ε Hf (t) values of −12.6 ~ −12.5 and of −10.4, respectively, for granites and of −8.8 ~ −8.1 and of −5.9, respectively, for orthogneisses (all calculated at 20 Ma). Furthermore, our compiled whole-rock Sr–Nd isotope data indicate that the Cenozoic Himalayan granites differ from the Paleozoic orthogneisses but overlap with the Himalayan (meta-)sedimentary rocks. Phase equilibrium modeling demonstrates that the Paleozoic orthogneisses are relatively infertile due to their low fractions of hydrous minerals and thus low bulk water contents as compared with the metapelites assuming no free fluid is present. Therefore, although the contribution of the Paleozoic orthogneisses cannot be totally precluded, the observed evidence suggests that they are not the main source components of the Cenozoic Himalayan granites. It is speculated that Ordovician and later sediments which can receive the weathered clasts of the early Paleozoic granites can act as the unrecognized sources of the Cenozoic granites. However, more work is required to characterize these sediments before drawing firm conclusions on the sources of these granites.