The newly discovered Zaorendao gold deposit is in the Tongren-Xiahe-Hezuo polymetallic district in the westernmost West Qinling orogenic belt. The estimated pre-mining resource is approximately 13.6 t of Au at an average grade of 3.02 g/t. Mineralization is predominantly controlled by NW-trending and EW-trending faults within diorite intrusions and surrounding sedimentary rocks. In the present study, in situ zircon U–Pb geochronology and Lu–Hf isotopic analyses of the ore-hosting diorite at Zaorendao were measured using LA-ICP-MS. The data suggest that the diorite was emplaced at ca. 246.5 ± 1.9 Ma. The large variation of zircon Hf isotopic composition (ɛ_Hf(t) values ranging from −12.0 to −1.8) indicates a two-stage model age (T_DM2) that ranges from 1.4 Ga to 2.0 Ga. Such Lu–Hf isotopic compositions indicate that the diorite was dominantly derived from a Paleo- to Meso-Proterozoic continental crust. The wide range of ε_Hf(t) and the presence of inherited zircon can be interpreted to suggest the mixing of Paleo- to Meso-Proterozoic continental crust with a mantle component. Combining such characteristics with the geochemistry of coeval rocks that are associated with the diorite, we therefore proposed that the gold-hosting Triassic diorite in the Zaorendao gold deposit formed in an active continental margin that was associated with the northward subduction of the paleo-Tethyan ocean.

Magmatic and metasomatic zircon occurs in many alkaline igneous rocks and both are potential economic reservoirs of Zr, and in some places, rare-earth elements. The Baerzhe deposit in China is an example of a system where both types of zircon occur. Previous studies recognized deuteric and variably altered magmatic zircon in a transsolvus miaskitic granite, as well as four types of metasomatic zircon in a transsolvus agpaitic granite. In this study, the relationships among, and origins of, zircon and how these relate to models for rare-metal mineralization are assessed. In situ backscattered electron (BSE) and cathodoluminescence (CL) imaging, Raman spectroscopy (including mapping), and chemistry of zircon from the agpaitic granite were conducted, combined with evaluation of published data on zircon from Baerzhe. Their textural, spectroscopic, and chemical characteristics suggest that the four types of metasomatic zircon in the agpaitic granite were not subjected to metamictization or intense alteration, with trace-element accommodation largely following a xenotime substitution mechanism. The most abundant type of metasomatic zircon in the agpaitic granite occurs in zircon-quartz pseudomorphs and exhibits comparable CL, Raman spectral, and chemical features to rare zircon that has partially replaced elpidite. This confirms that the pseudomorphs formed by complete replacement of elpidite. The pseudomorph zircon occurs in association with snowball quartz that contains inclusions of zircon, aegirine, and albite, and with secondary quartz containing aegirine. This is consistent with their coeval formation during Na metasomatism. The restriction of Na metasomatism to the agpaitic granite indicates that this event and the associated zircon formation resulted from early autometasomatism of the agpaitic phase. REE- and Be-rich zircon that replaced magmatic amphibole crystallized as a result of reaction with a REE- and Be-rich fluid that most likely was responsible for the later REE-Nb-Be mineralization that affected both the miaskitic and agpaitic granites. The miaskitic granite contains deuteric and altered magmatic zircon with different chemical characteristics to the four types of metasomatic zircon in the agpaitic granite. This suggests that secondary Zr mineralization in the miaskitic granite formed from different fluids to those that metasomatized the agpaitic granite and may also have resulted from autometasomatism. This study reveals a complex picture for the formation of zircon at Baerzhe, the character of which can vary significantly, both temporally and spatially. Such variable chemistry of the various types of zircon resulted not only from their different origins (magmatic vs. metasomatic), but also from localized water-rock interaction that involved multiple stages of fluids. Zircon in both the miaskitic and agpaitic phases was mainly the product of autometasomatism that was constrained to their parental granites.

The Baerzhe deposit represents world-class mineralization of rare metals, in particular rare earth elements (REEs) and Nb, and shares mineralogical and geochemical similarities with the Strange Lake deposit in Canada. However, the mechanisms responsible for the rare-metal enrichment at Baerzhe are still under debate. Mineralization at Baerzhe is characterized by multiple stages of dissolution, re-precipitation and enrichment of rare-metal elements, as with the Strange Lake deposit. Zirconium was concentrated during magmatic fractionation of elpidite in a transsolvus agpaitic granite (Stage I), and was redistributed into zircon–quartz pseudomorphs during Na-metasomatism (Stage II). Rare earth elements, Nb and Be occur as a variety of mineral assemblages, mainly as pseudomorphs after complete replacement of unknown precursor minerals that are disseminated through the hematized transsolvus granites and were precipitated in three sequential stages (Stages III–V): Stage III is characterized by the mineralization of hingganite, aeschynite-(Y), columbite-(Fe) and euxenite-group minerals, accompanied by hematite alteration; the following Stage IV is defined mainly by monazite-(Ce) crystallization, and Stage V by the preceding REE minerals being replaced by light REE minerals (i.e. bastnäsite-(Ce) and fluocerite-(Ce)). The formation of various pseudomorphs as replacement for precursor minerals is also a common feature documented at Strange Lake. Our results call for a re-evaluation of the timing and role of meteoric waters at Baerzhe, which were suggested by previous researchers to have entered the ore-forming system prior to the mineralization of zircon. New in situ O isotopic analyses on pseudomorph-hosted mineral pairs and investigation of silicate melt–fluoride melt immiscibility through observations and analyses of melt inclusions could provide further insights into the nature of the Baerzhe system.