We report new results of a combined focused ion beam and high-resolution transmission electron microscopy (FIB/HRTEM) investigation of platinum group element-rich (PGE) samples from the Merensky Reef of the Bushveld Complex in South Africa. Chromitites and pegmatitic melanorite from the Northam and Rustenburg platinum mines contain Ru-Rh-Pt arsenide nanocrystals (<50 nm in size) as inclusions in pyrrhotite and pyrite, as well as Ru-Rh-Os, Ru-Rh-Pt-Ir, and Ru-Rh-Pt sulfide nanocrystals in pyrrhotite and pentlandite. Chalcopyrite did not incorporate nanometer-sized PGE-bearing minerals (PGM). These nano-scale PGM inclusions form a new population that was hitherto unknown from the Merensky Reef. The PGM inclusions are found as idiomorphic, plate-like nanocrystals, which are differently oriented with respect to the host base metal sulfide (BMS). In one TEM foil we observed Ru-Rh-Pt-sulfide nanocrystals exsolved from a pentlandite matrix with a common crystallographic orientation that is visible in TEM dark-field images. However, this phenomenon is only patchily developed and ahead of the exsolution front the pentlandite still contains Ru-Rh in solid solution.
Occasionally the PGM inclusions form clusters, develop a flow texture, or are aligned like pearls on a string. These observations are inconsistent with sub-solidus exsolution and suggest that the nano-scale PGM crystallized before solidification of the sulfide melt and the formation of their BMS hosts.
There is evidence from recently performed laboratory experiments on silicate melts with additives of Ru and Pd that discrete Fe-rich Fe-Ru-Pd alloys (nanometer-sized) can indeed crystallize early. These experiments raise the possibility that the PGE-bearing nanocrystal inclusions found in the BMS of the Merensky Reef might represent preserved relicts of an early phase of magmatic PGM that precipitated from the silicate melt and afterwards were collected by the sulfide melt. Combined FIB/HRTEM techniques open up a window through which the nano-scale mineral population can be viewed. PGE previously thought to be present in BMS solid solution might actually be an artefact of inadequate spatial resolution.