The Witwatersrand Supergroup is a 2,9 to 2,7 Ga sedimentary basin located on the Kaapvaal Craton (Fig. 1) and is world-renowned for its super-giant gold endowment. The gold in the basin is largely associated with a series of quartz-pebble conglomerate beds (reefs) that formed during continental sedimentation in alluvial fans and braid-plain depositional environments (Frimmel, 2019). Historical gold production from the Witwatersrand sedimentary basin accounts for ~ 32% of the total gold that has ever been mined globally (Frimmel, 2014). Despite the quantitative importance of the Witwatersrand deposits, the ultimate provenance of the gold is still a topic of scientific debate, in part because any obvious hinterland gold source has been removed from the geological record through erosional processes. Genetic models for the formation of the mineralisation include solution transport models with algal mat ‘traps’ (Heinrich, 2015), placer models (Kirk et al., 2001), hydrothermal models (Barnicoat et al., 1997; Phillips & Law, 2000), and the ‘modified-placer’ model (Frimmel et al., 2005; Hayward et al., 2005). Perhaps the most broadly-accepted theory is this modified-placer model since it combines the microscale textural observations of the hydrothermal model with the observation that mined gold grades follow spatial distributions that are better explained by the placer model. Briefly, this model suggests that native gold grains, likely sourced from continental-scale chemical and mechanical weathering and erosion of granite-greenstone dominated hinterlands (Frimmel, 2014; Koglin et al., 2010), were initially concentrated in conglomerates along with other dense detrital minerals. Subsequent hydrothermal fluid overprints resulted in local-scale remobilisation of gold within these conglomeratic units, thus accounting for the observed crystalline gold textures ( e.g., Minter et al., 1993) and sulphide morphologies (e.g., da Costa et al., 2020).
Sedimentation of the Witwatersrand Supergroup occurred under a reducing Archean atmosphere, as evidenced by the presence of detrital pyrite (FeS2) and detrital uraninite (UO2) within the auriferous quartz pebble conglomerates (England et al., 2002). Given the possibility that Witwatersrand gold was derived from orogenic-type primary gold deposits in an Archean granite-greenstone hinterland, it is reasonable to infer that a proportion of the detrital pyrite may similarly have been sourced from such deposits. Remnant vestiges of Archean greenstone belts located on the Kaapvaal craton commonly show elevated gold endowments (e.g., Barberton Greenstone belt = 376 t gold; Amalia-Kraaipan Greenstone belt = 32 t gold; Giyani Greenstone belt = 12 t gold; Pietersburg Greenstone belt = 4 t gold (Pearton and Viljoen, 2017)). Much of this gold can be classified as ‘invisible gold’, a loosely-defined term reserved for gold that behaves in a refractory manner to direct cyanidation (Coetzee et al., 2011) and which encompasses both gold dissolved in the sulphide mineral structures (generally arsenopyrite and arsenian pyrites) as a cation substituent (Merkulova et al., 2019; Reich et al., 2005), and gold occurring as submicron and nano-scale inclusions within sulphide mineral hosts (Paleniket al., 2004). For example, arsenopyrite and arsenian pyrite mined as part of the run-of-mine from orogenic gold shoots in the Barberton greenstone belt commonly host high concentrations of invisible gold (up to 1000 ppm (Agangi et al., 2014)), and the prevalence of this ore type has warranted the use of biological oxidation (BIOX®) beneficiation strategies (Van Aswegen & Marais, 1999). Similarly, fine and invisible gold mineralization is reported in the Kraaipan Greenstone Belt, strongly associated with pyrite and pyrrhotite (Hammond & Moore, 2006; Morishita et al., 2019).
The mining of Witwatersrand conglomerates and their contained gold and associated detrital sulphides dates back to 1885 (Frimmel & Nwaila, 2021). This extended mining legacy has resulted in a massive accumulation of tailings material (6 billion tons; Wymer, 2001) which, largely on account of historical processing inefficiencies, is currently being re-mined as a secondary gold resource (Bosch, 1987; Sibanye-Stillwater Limited, 2021). Most of the dedicated plants that reprocess Witwatersrand tailings material apply re-milling followed by carbon-in-leach extraction as the beneficiation strategy. This strategy targets native gold and results in a gold recovery that ranges between 30–50% (Bosch, 1987; Harmony Gold, 2021; Sibanye-Stillwater Limited, 2021). The present contribution seeks to better understand the mineralogical deportment of the remaining 50–70% of the gold that is deemed refractory to this beneficiation strategy, with particular emphasis on the role of Archean detrital pyrites as viable mineral hosts for the ‘missing’ recovery.