New constraints on the open magma chamber processes in the formation of giant Hongge Fe-Ti-V oxide deposit

Abstract The Hongge layered intrusion in Emeishan Large Igneous Province (ELIP), SW China, contains the largest Fe-Ti-V oxide deposit in China (4500Mt ore reserve). The mechanism by which million tons of Fe-Ti-V oxide concentrated is debated. In this study, new petrography, mineral chemistry (~3600 microprobe analyses for olivine, clinopyroxene and plagioclase) and bulk rock major element chemistry (108 samples) are presented from samples collected over a 940 m drill core (ZK11003) in the southern part of the Hongge layered intrusion. In contrast to the Bushveld and Skaergaard layered intrusions, a broad normal fractionation trend upwards recorded by cumulus clinopyroxene and plagioclase is absent, whereas frequent reversals to more primitive compositions are observed in the Hongge intrusion. According to the mineral reversals, we divided the entire intrusion into eleven stratigraphic units. Notably, in the Middle Zone (MZ) and Lower Zone (LZ), the appearance of Fe-Ti-V oxide ore layers are closely related to these reversals. These distinct units are regarded as indicators for major replenishments of primitive high Ti ferrobasaltic/picritic magmas. In these samples, various clinopyroxene populations with both low and high Mg# are also observed and phases with high Mg# are interpreted to be inherited from the replenishing magmas. Previous studies have proposed that the replenishing magma might be anhydrous. However, the presence of a considerable amount of hydrous cumulus minerals including phlogopite and hornblende indicates that there is a high water content in melts at some stages of the formation of the Hongge layered intrusion, especially in the LZ and MZ. Contamination of crustal materials (e.g., introduction of aqueous fluids) or replenishing magmas with different water contents could lead to high water activities in the shallow crystallizing magma chamber. As a consequence, mixing of hydrous residual melts with replenishing magmas would trigger large scale precipitation of Fe-Ti-V oxide ores. Relatively high water activity, which could suppress the crystallization of cumulus plagioclase, would also explain the absence of this phase in the LZ and MZ. Thus, based on all these observations, we propose that the Hongge layered intrusion was formed by multiple emplacements of crystal-laden magmas that carried olivine and clinopyroxene already crystallized in deep-seated chambers.