P-T-T-DEFORMATION-FLUID CHARACTERISTICS OF LODE GOLD DEPOSITS: EVIDENCE FROM ALTERATION SYSTEMATICS

Abstract

Structurally hosted lode gold-bearing quartz vein systems in metamorphic terranes possess many characteristics in common, spatially and through time; they constitute a single class of epigenetic precious metal deposit, formed during accretionary tectonics or delamination. The ore and alteration paragenesis encode numerous intensive and extensive variables that constrain the pressure—temperature—time—deformation—fluid (P—T—t—d—f) evolution of the host terrane and hence the origin of the deposits. The majority of lode gold deposits formed proximal to regional translithospheric terrane—boundary structures that acted as vertically extensive hydrothermal plumbing systems; the structures record variably thrust, and transpressional—transtensional displacements. Major mining camps are sited near deflections, strike slip or thrust duplexes, or dilational jogs on the major structures. In detail most deposits are sited in second or third order splays, or fault intersections, that define domains of low mean stress and correspondingly high fluid fluxes. Accordingly, the mineralization and associated alteration is most intense in these flanking domains. The largest lode gold mining camps are in terranes at greenschist facies; they possess greenschist facies hydrothermal alteration assemblages developed in cyclic ductile to brittle deformation that reflects interseismic—coseismic cycles. Interseismic episodes involve the development of ductile S—C shear zone fabrics that lead to strain softening. Pressure solution and dislocation glide microstructures signify low differential stress, slow strain rates of ≤ 10−13 s−1, relatively high confining stress, and suprahydrostatic fluid pressures. Seismic episodes are induced by buildup of fluid pressures to supralithostatic levels that induce hydraulic fracturing with enhanced hydraulic conductivity, accompanied by massive fluid flow that in turn generates mineralized quartz veins. Hydrothermal cementing of ductile fabrics creates ‘hardening’, lowers hydraulic conductivity, and hence promotes fault valve behaviour. Repeated interseismic (fault valve closed), coseismic (valve open) cycles results in banded and/or progressively deformed veins. Alteration during both interseismic and coseismic episodes typically involves the hydrolysis of metamorphic feldspars and Fe, Mg, Ca-silicates to a muscovite/paragonite—chlorite ± albite/K-feldspars assemblage; carbonization of the metamorphic minerals to Ca, Fe, Mg-carbonates; and sulphidation of Fe-silicates and oxides to sulphides. Geochemically this is expressed as additions of K, Rb, Ba, Cs, and the volatiles H2O, CO2, CH4, H2S in envelopes of meter to kilometer scale. K/Rb and K/Ba ratios are close to average crustal values, potentially ruling out late stage magmatic fluids where K/Rb and K/Ba are respectively lower and higher than crustal values. Smaller deposits are present in subgreenschist, and amphibolite to granulite facies terranes. The former are characterized by subgreenschist facies alteration assemblages, vein stockworks, brittle fracturing and cataclastic microstructures, whereas the latter feature amphibolite to granulite facies alteration assemblages, ductile shear zones, ductilely deformed veins, and microstructures indicative of dislocation climb during interseismic episodes. Hence the lode gold deposits constitute a crustal continuum of deposits from subgreenschist to granulite facies, that all formed synkinematically in broad thermal and rheological equilibrium with their host terranes. These characteristics, combined with the low variance of alteration assemblages in the higher temperature deposits, rules out those being metamorphosed counterparts of greenschist facies deposits. Deposits at all grades have a comparable metal inventory with high concentrations of Au and Ag, where Au/Ag averages 5, with enrichments of a suite of rare metals and semi-metals (As, Sb, ± Se, Te, Bi, W, Mo and B), but low enrichments of the base (Cu, Pb, Zn, Cd) and other transition (Cr, Ni, Co, V, PGE, Sc) metals relative to average crust. The hydrothermal ore-forming fluids were dilute, aqueous, carbonic fluids, with salinities generally ≤ 3 wt.% NaCl equivalent, and X(CO2 ± CH4) 10–24 wt.%. They possess low Cl but relatively high S, possibly reflecting the fact that metamorphic fluids are generated in crust with ∼ 200 ppm Cl, but ∼ 1 wt.%S. Primary fluid inclusions are: (1) H2OCO2, (2) CO2-rich with variable CH4 and small amounts of H2O, and (3) 2-phase H2O (liquid-vapor) inclusions. Inclusion types 2 and 3 represent immiscibility of the type 1 original ore fluid. Immiscibility was triggered by fluid pressure drop during the coseismic events and possibly by shock nucleation, leading to highly variably homogenization temperatures in an isothermal system. A thermodynamic evaluation of alteration assemblages constrains the ore fluid pH to 5–6; redox controlled by the HSO4/H2S and CO2/CH4 buffers; and XCO2 that varies. The higher temperature deposits formed under marginally more oxidizing conditions. Stable isotope systematics of the ore and gangue minerals yields temperatures of 200–420°C, consistent with the crustal spectrum of the deposits, very high fluid rock ratios, and disequilibrium of the externally derived ore fluids with wall rocks. The ore fluid δD and δ18O overlap the metamorphic and magmatic ranges, but the total dataset for all deposits is consistent only with dominantly metamorphic fluids. Carbon isotope compositions of carbonates span −11 to +2% and show provinciality: this is consistent with variable proportions of reduced C (low δ13C) and oxidized C (higher δ13C) in the source regions contributing CO2 and CH4 to the ore fluids. In some instances, C appears to have been derived dominantly from proximal to the deposits, as in the Meguma terrane (δ13C ∼ − 22%). Sulphur isotope compositions range from 0 to +9‰, and are consistent with magmatic S, dissolution or desulphidation of magmatic sulphides, or average crustal sulphides. 34S-depleted sulphides occur in ore bodies such as Hemlo where fluid immiscibility led to loss of H2S and consequent fluid oxidation. Gold is probably transported as an Au(HS)−2 complex. Relatively high S but low Cl in the hydrothermal fluid may explain the high Au slow base metal characteristic of the deposits. Gold precipitated in ore bodies due to loss of S from the ore fluid by sulphidation of wall rock, or immiscibility of H2S; and by oxidation or reduction of the fluid, or by chemisorption, or some combination of these processes. Most lode gold deposits have been brittly reactivated during uplift of host terranes, with secondary brines or meteoric water advecting through the structures. These secondary fluids may remobilize gold, generate retrograde stable isotope shifts, reset mineral geochronometers, and leave trails of secondary fluid inclusions. Data on disturbed minerals has led to invalid models for lode gold deposits. The sum of alteration data leads to a model for lode gold deposits involving a clockwise P—T—t evolution and synkinematic and synmetamorphic mineralization of the ‘deep later’ type. During terrane accretion oceanic crust and sediments are subcreted beneath the terrane boundary. Thermal equilibration generates metamorphic fluids that advect up the terrane structure, at lithostatic fluid pressure, into the seismogenic zone where the majority of deposits form. Thus many lode gold deposits are on intrinsic part of the development of subduction—accretion complexes of the high-T, low-P type.