The hydrothermal evolution and genesis of the porgera gold deposit, Papua New Guinea

Abstract

High grade gold mineralization at Porgera is the result of two spatially and temporally distinct mineralization events referred to as stages I and II. Stage I mineralization resulted in the formation of pyrite, Fe-rich sphalerite and galena rich vein sets, spatially associated with intrusions, and hosted in intrusions, the surrounding altered sediment and calcareous black shale. Early sulphide rich veins (A/B veins) contain several grams of Au/tonne but due to their scattered nature do not constitute high grade ore. The fluid responsible for this mineralization contained about 1 molal NaCl (corrected for C02 after Richards and Kerrich, 1993), a minimum of 2 molal C02, had an average Th of about 280°C (Richards et al., 1993) and was relatively reduced. Stage II mineralization which constitutes the bulk of the high grade ore is hosted in banded quartz and quartz-roscoelite­ pyrite-barite breccia veins (D veins), which are spatially associated with the Roamane Fault. Altered sediment and intrusive rock are the main host to stage II mineralization. Fluid inclusion data from stage II quartz yields a bimodal salinity at 4.2 and 7.8 wt % NaCl equiv,and an average Th of 145oC (Richards et al., 1993). The C02 content of these fluids is extremely variable ranging up to at least 2 molal. High grade ore shoots occur at the intersection of stage I and II veins sets. Munroe(1996) concluded that the Roamane Fault was not active during stage I, but was the main spatial control upon development of stage II mineralization. Stage I mineralization is hosted in and disseminated about A and B veins. In these veins quartz and carbonate are intimately intergrown with stage I sulphide but are rarely observed in equilibrium with sulphide. Textural evidence shows that pyrite, sphalerite, galena and arsenopyrite are aggressively replaced by quartz and carbonate. No evidence was found to support the existence of early A vein quartz and all gangue in these veins is interpreted to postdate pyrite, sphalerite and galena. It is inferred that sulphides formed early,in stage I, and were dissolved during stage II quartz and carbonate deposition. Evidence from this study suggests that all quartz and carbonate in A veins is temporally equivalent to material in the more spatially restricted stage II D veins. Alteration of the sediments was one of the earliest hydrothermal events and predates stage I. It is associated with the bleaching of black shales along thin grey pyrite veinlets termed G stringers, and may be equivalent to Richards and Kerrich's propylitic alteration of the intrusive phases. This alteration is associated with gains in K20, Rb, Ba, and As; and losses in Na20, CaO, C02, Pb, Zn, and Sr. Local geometry, relative paragenetic position and geochemistry suggest that the fluids responsible for this early alteration of the sediments emanated from the currently exposed intrusions. Intrusive rocks are altered next to both AlB (stage I) and D (stage II) veins. Altered selvedges next to both vein types have undergone gains in K20, C02, H20, Pb, As and S and losses in Na20, Sr and Ba. Electron microprobe analyses of stage II D vein pyrite indicate highly anomalous gold and arsenic values of up to 0.9 wt% Au and 7.7 wt% As. The highest values (both Au and As) are found in the high grade M126/Roamane Fault veins as well as A-D intersections. In A vein pyrite, gold is below detection and arsenic ranges from 300 ppm to 2.4 wt %. X-ray images of D vein pyrite for Au, As and Cu show well developed growth banding - with Au and As showing a moderate correlation. These images suggest that Au, As and Cu occur in solid solution within the pyrite structure. Bulk ()34S analysis of Stage I and Stage II pyrite by Richards and Kerrich (1993) gave values of +3 to +5 per mil and -14 to -10 per mil respectively. SHRIMP sulphur isotope analysis (this study) of Stage II pyrite yielded ()34S values ranging from -18 to -14 per mil. Pyrite from the early G stringers returned SHRIMP ()34S values of -4 per mil whereas bulk sulphur analyses returned values ranging from -1.2 to 3.7 per mil. While G stringers are not strongly negative they are more negative than the bulk rock values for stage I sulphides. These data indicate reducing fluids during stage I, oxidizing fluids during Stage II and an intermediate redox state for the early pre-stage I fluids. Since there are no thermodynamic data available for roscoelite and the only data available for other vanadium species are at 25°C, it was necessary to construct a thermodynamic data set to estimate the relative stability of roscoelite and other vanadium species in f02- pH space at temperatures up to 300°C. Results indicate that at neutral to even moderately acidic pH, vanadium is only mobile under oxidizing conditions and that roscoelite is stable under reducing conditions. This suggests that vanadium would be effectively transported by an oxidizing fluid and may be fixed as roscoelite upon reduction. It is proposed that Porgera evolved initially during stage I, within the Pb-Zn rich moderately distal reaches of a porphyry hydrothermal system, beneath a regionally extensive tectono-stratigraphic boundary. Base metals and gold were complexed as chloride and sulphide respectively, and were deposited through passive boiling of a C02 rich moderately saline fluid at about 280°C and under pressures approaching lithostatic. Rupture of the Roamane Fault resulted in pressure drop and associated throttling and adiabatic decompression of the fluid, which prematurely ended the stage I hydrothermal system. Fluid that originally deposited base metal sulphide and gold in the early veins, immediately began dissolving both, and simultaneously deposited quartz and carbonate as the fluid flowed into low pressures zones along the Roamane Fault. Here, the same fluid further decompressed and continued to deposit barren quartz and lesser carbonate. Decompression also permitted deeper sourced, fresh, oxidized magmatic fluid to quickly achieve similar levels within the Roamane Fault and mix with the reduced pregnant ore bearing fluid. Mixing of the two fluids at the intersection of stage I and II structures resulted in oxidation of the reduced gold bearing fluid and precipitation of the high grade gold rich quartz-roscoelite layers. That gold was carried in the reduced rather than the oxidized fluid is evidence by the replacement of pyrite by gold within the assemblage quartz-roscoelite-pyrite-barite. Key elements of the model are the carbonate rich host rocks, the C02 rich (high fluid pressure) composition of the early fluid and the presence of a regionally extensive pressure seal. Rupture of the Roamane Fault breached the seal causing rapid decompression of the fluid and loss of both C02 and H2S to the vapour phase which resulted in the dissolution of both sulphide and gold. Most previous studies of boiling hydrothermal systems emphasize the potential of boiling to deposit metals (Drummond, 1981; Drummond and Ohmoto, 1986; Cole and Drummond, 1986; Reed and Spycher, 1985). The current study emphasizes the potential of boiling to dissolve and remobilize both base metals and gold. Whether boiling results in sulphide deposition or dissolution is related to the relative balance between C02 loss (favouring pH increase and deposition) and H2S loss (favouring dissolution). Small amounts of boiling (- 5%) favour sulphide and gold deposition whereas large amounts of boiling/throttling (> 10%) favour dissolution. Reaction path modelling of various boiling scenarios has indicated that both sulphide and gold dissolution may be expected from a boiling fluid whose pH is buffered by the surrounding host rock. At Porgera it is suggested that boiling under similar conditions led to the remobilization of a significant proportion of the widely distributed stage I gold and its subsequent redeposition, through mixing induced oxidation, in a more concentrated form within a much smaller area. In the New Guinea Highlands most of the major Au/Cu deposits show a close spatial relationship to the contact between Tertiary Limestone and underlying Mesozoic clastic sediments. It is suggested that: (1) this contact/decollement surface acted as a major barrier to fluid flow causing a major discontinuity in fluid pressure of evolving hydrothermal systems; and (2) rupture of this contact lead to large scale decompression of large overpressured hydrothermal fluid reservoirs which lead to focussing of large amounts of fluid which simultaneously dissolved pre-existing sulphide and Au. The physical parameters necessary for decompression induced remobilization of gold are commonly achieved in hydrothermal systems. Many of the Carlin type deposits have similar characteristics (to those at Porgera) which include: (1) calcareous shale host rocks beneath regionally extensive tectono-stratigraphic boundary; (2) early widely dispersed disseminated/veinlet hosted mineralization deposited by early C02 rich fluids; and (3) later overprinting spatially restricted gold rich mineralization associated with arsenical pyrite within a relatively oxidized assemblage. In the Carlin type deposits it is possible that remobilization/leaching affected distal sediments as well as early scattered base metal sulphide mineralization (this is also a possibility at Porgera).