The role of decarbonization and structure in the Callie gold deposit, Tanami Region of northern Australia

Abstract

The Callie deposit is the largest (6.0 Moz Au) of several gold deposits in the Dead Bullock Soak goldfield of the Northern Territory’s Tanami Region, 550 km northwest of Alice Springs. The Callie ore lies within corridors, up to 180 m wide, of sheeted en echelon quartz veins where they intersect the 500-m-wide hinge of an ESE-plunging F1 anticlinorium. The host rocks are the Blake beds, of the Paleoproterozoic Dead Bullock Formation, which consist of a > 350-m-thick sequence of lower greenschist facies graphitic turbidites and mudstones overlying in excess of 100 m of thickly bedded siltstones and fine sandstones. The rocks are Fe-rich and dominated by assemblages of chlorite and biotite, both of which are of hydrothermal and metamorphic origin. A fundamental characteristic of the hydrothermal alteration is the removal of graphite, a process which is associated with bleaching and the development of bedding-parallel bands of coarse biotite augen. Gold is found only in quartz veins and only where they cut decarbonized chloritic rock with abundant biotite augen and no sulfide minerals. Auriferous quartz veins differ from barren quartz veins by the presence of ilmenite, apatite, xenotime, and gold and the absence of sulfide minerals. The assemblage of gold–ilmenite–apatite–xenotime indicates a linked genesis and mobility of Ti, P, and Y in the mineralizing fluids. Geochemical analysis of samples throughout the deposit shows that gold only occurs in sedimentary rocks with high FeO/(FeO+Fe2O3) and low C/(C+CO2) ratios (> 0.8 and < 0.2, respectively). This association can be explained by reactions that convert C from reduced graphitic host rocks into CO2 and reduce ferric iron in the host rocks to ferrous iron in biotite and chlorite. These reactions would increase the CO2 content of the fluid, facilitating the transport of Ti, P, and Y from the host rocks into the veins. Both CO2 and CH4 produced by reaction of H2O with graphite, effervesced under the lower confining pressures in the veins. This would have partitioned H2S into the vapor phase, destabilizing Au–bisulfide complexes; the loss of CO2 and H2S from the aqueous phase caused precipitation of gold, ilmenite, apatite, and xenotime. It is proposed that this process was the main control on gold precipitation. Oxidization of iron in the very reduced wall rocks, resulting in reduction of the fluid, provided a second mechanism of gold precipitation in previously decarbonized rocks, contributing to the high grades in some samples. Although sulfide minerals, especially arsenopyrite, did form during the hydrothermal event, host rock sulfidation reactions did not play a role in gold precipitation because gold is absent near rocks or veins containing sulfide minerals. Sulfide minerals likely formed by different mechanisms from those associated with gold deposition. Both the fold architecture and subsequent spatially coincident sinistral semibrittle shearing ensured that the ore fluids were strongly focused into the hinges of the anticlines. Within the anticlines, a reactive cap of fine-grained, graphitic, reduced Fe-rich turbidites above more permeable siltstones and fine sandstones impeded fluid flow ensuring efficient removal of graphite, and the associated effervescence of CO2 from the fluid caused the precipitation of gold. Exploration for similar deposits should focus on the intersection of east–west shear zones with folds and Fe-rich graphitic host rocks.