Abstract
The Nova-Bollinger Ni-Cu-platinum group element (PGE) deposit in the Fraser zone of the Albany-Fraser orogen consists of two main orebodies, Nova and Bollinger, hosted by the same tube-shaped intrusion but having distinctly different Ni tenors of around 6.5 and 4.8 wt %, respectively. Nova is also higher in Pd, but Cu and Pt tenors are similar. Both deposits have very low PGE tenors, with average Pd concentrations of 110 ppb in massive sulfide at Bollinger and 136 ppb at Nova. The Nova and Bollinger orebodies show relatively little internal differentiation overall on deposit scale but show strong differentiation into chalcopyrite-rich and chalcopyrite-poor regions at a meter scale. This differentiation is more prevalent at Nova, where massive sulfide-filled vein arrays are more extensively developed, and in massive ores, particularly veins, than in net-textured ores. Net-textured and disseminated ores have on average Ni and Cu grades and tenors similar to those of massive, semimassive, and breccia ores in the same orebody but a smaller range of variation, largely due to a more limited extent of sulfide liquid fractionation and higher average concentrations of Pt and Pd than adjacent massive ores. Unusually for differentiated magmatic sulfides, there is no systematic positive correlation between Pt, Pd, and Cu. A partial explanation for the lack of a Pd-Cu correlation is that Pd was partitioned into peritectic pentlandite in the middle stages of sulfide liquid solidification. This explanation is not applicable to Pt, as Pt characteristically forms its own phases rather than residing in base metal sulfides. PGE tenors are very low in both orebodies, very similar to those observed in other Ni-Cu-Co sulfide ores in orogenic settings, notably the Savannah and Savannah North orebodies. This depletion is attributed to sulfide retention in the mantle source of the parent magmas rather than to previous fractional extraction of sulfide liquid in staging chambers or feeder networks. The higher Ni and Pd tenors at Nova are attributed to reworking and upgrading of precursor sulfide liquid originally deposited upstream at the Bollinger site. Replicate analyses of multiple jaw-crusher splits returned highly variable Pt and Au assays but much smaller relative errors in the other PGEs. The poor Pt and Au reproducibilities are attributed to nugget effects, explicable by much of the Pt and Au in the samples being present in sparse Pt- and Au-rich grains. This is principally true for Pt in massive rather than disseminated ores, accounting for a strong contrast in the distribution of Pt/Pd ratios between the two ore types. Numerical simulation suggests that Pt is predominantly resident in Pt-rich platinum group minerals with grain diameters of 100 μm or more and that at the low (<100 ppb) concentrations in these ores, this results in most assays significantly underreporting Pt. This is likely to be true in other low-PGE ores, such that apparent negative Pt anomalies in massive ores may in such cases be attributable to sampling artifacts.