PACROFI VI - Electronic Program

Fluid inclusions in Archean high-temperature gold deposits: Can we evaluate post-entrapment modification?

J R Ridley1 and S G Hagemann2

  1. Institute for Isotope Geology and Mineral Resources, ETH, Zürich, CH 8092, Switzerland.
  2. Dept. Geology and Geophysics, University of Wisconsin, 1215 W Dayton St, Madison, WI 53706, USA and Lehrstuhl für Angewandte Mineralogie und Geochemie, Technische Universität München, Lichtenbergstr. 4, 85747 Garching, Germany.

Fluid inclusions studies of syn-metamorphic lode- and vein-style gold deposits in greenschist-facies terrains have shown the importance in these hydrothermal systems of a low-salinity, H2O-CO2 (XCO2 ~ 0.1-0.2) fluid, with in some instances minor CH4 (Ho et al., 1985; Gebre-Mariam et al., 1995). Vein and proximal hydrothermal alteration assemblages equilibrated at the T, XCO2 and fO2 of this fluid, confirming its dominance of the system chemical evolution.

In the Archean Yilgarn Block of Western Australia, a number of syn-metamorphic lode- and vein-type gold deposits occur in amphibolite and lower-granulite facies terrains. Formation conditions for these deposits are estimated to range between 450 and 700oC at 3 to 5 kbar (for details see Groves et al., 1995). That these high-temperature deposits are genetically related to those in greenschist-facies terrains is indicated by their closely similar timing, structural controls on siting, and ore and alteration chemistry. Ore and alteration assemblages poorly constrain XCO2 to within the broad range of 0.05 < XCO2 < 0.5, but importantly indicate a value of fO2 such that CO2 would be the dominant carbonic phase in the fluid.

At each deposit investigated, three or four of the following types of fluid inclusion are present in vein quartz and clinopyroxene (this study; Bloem and Brown, 1991; Ridley et al., 1995):

(i) Mixed low-salinity aqueous-carbonic fluids. Salinities are typically 2 - 6 eq wt % NaCl, and the carbonic phase forms 10 - 90 % of the inclusion by volume and ranges from nearly pure CO2 to pure CH4. Aqueous:carbonic ratios and CO2:CH4 ratios are very variable even within individual deposits.
(ii) Carbonic fluids. As for type (i) inclusions, the compositions range from nearly pure CO2 to pure CH4, and CO2:CH4 ratios are often very variable within an individual deposit. Some of the CH4-rich inclusions of this type are clearly secondary.
(iii) Rare, highly saline (30 - 50 eq wt % NaCl), dominantly aqueous fluids with detectable CO2. Because of their rarity, these inclusions are not considered further here.
(iv) Low to moderately saline (5 - 25 eq wt % NaCl) aqueous fluids in secondary inclusions.
Isochores for type (i) and type (iv) inclusions project generally to slightly too high pressures whereas those for type (ii) inclusions project with temperature to pressures as little as a half those inferred for mineralisation on petrological grounds, with generally lower pressures for CH4-richer inclusions (Fig. 1). The secondary nature and densities of type (iv) inclusions support geological and petrological evidence for near isobaric cooling of the terrain after mineralisation.

At face value, the fluid-inclusion record thus indicates a complex hydrothermal history, with fluids trapped under very variable conditions, with very variable H2O:CO2:CH4 ratios, and also with fO2 generally significantly lower than would be in equilibrium with the alteration and ore assemblages. Large variations in fluid pressure during trapping are considered unlikely given the high temperatures of mineralisation.

Figure 1. Representative isochores for inclusion populations of the Marvel Loch deposit, Western Australia, compared with P-T conditions of mineralisation estimated from mineral equilibria. The vertical bars span the range of isochores for each population.

There is, however, a general trend that fluid inclusion populations have less variable compositions, and their compositions and densities are closer to those predicted from alteration petrology in deposits formed at lower-amphibolite facies conditions than those formed at higher temperatures. For instance, the isochores of type (i) inclusions from the lower-amphibolite facies Three Mile Hill deposit pass through the P-T conditions inferred for mineralisation (Hagemann and Ridley, 1993; Ridley et al. 1995).

The overall increasing divergence between fluid-inclusion properties and wall-rock assemblages with increasing temperature suggests temperature-dependent compositional, and also potentially mechanical, re-equilibration with ambient conditions after entrapment. Three processes of inclusion re-equilibration have been suggested recently on theoretical grounds and from the results of experiments on synthetic fluid inclusions: hydrogen diffusion (e.g. Hall and Bodnar, 1990); wicking of water, most likely through diffusion (Qiu et al., 1992), and mechanical re-equilibration of inclusions to ambient pressure changes (e.g. Vityk and Bodnar, 1995).

In the context of the genesis of high-temperature gold deposits it is important to determine whether all or most early CO2- and CH4-bearing fluid inclusions at these deposits could have had initially a similar composition to the fluid inferred for greenschist-facies deposits, i.e. whether the observed complex record is a result of variable re-equilibration at high temperatures post-entrapment. This question is investigated here by considering whether the observed compositions are consistent with the direction of likely re-equilibration during the P-T-X history subsequent to mineralisation, and whether observed molar volumes are consistent with the inferred compositional changes. Given the environment and post-entrapment P-T history of the hostrocks, the possible driving forces for re-equilibration considered are:

  1. H2 diffusion into the inclusions due a control on ambient fO2 by the sulphide assemblage of the rocks on cooling, hence leading to decreased inclusion CO2:CH4 ratios.
  2. Lowered aqueous:carbonic ratios of the fluid due to diffusional loss of water as a result of buffering of ambient fH2O to lower values by retrogression of alteration assemblages during cooling (Yardley, 1981).
  3. Mechanical re-equilibration to lower inclusion volumes during near-isobaric cooling.

Mernagh and Witt (1995) argued against hydrogen diffusion alone as the cause for inclusion compositional variability at similar gold deposits. However, if combinations of the above three re-equilibration mechanisms operated, post-entrapment re-equilibration of an initially near uniform population is an allowable explanation of the variability of inclusion composition and molar volume. The molar volumes of CO2-rich type (ii) inclusions are consistent with trapping as inclusions with XH2O of 0.5 to 0.8 at the P-T conditions of mineralisation and subsequent diffusional loss of most of the water component. The trend of increasing molar volume with increasing CH4 content can be modelled in terms of conversion of initial CO2 to CH4 + H2O synchronous with, or prior to, diffusional water loss. The slightly lower than expected molar volumes of the relatively water rich type (i) inclusions is interpreted to be due to a combination of some H2 addition, and minor mechanical re-equilibration. The best fit uniform initial composition is a mixed, low-salinity H2O-CO2 fluid similar to that in deposits in greenschist-facies terrains or of slightly higher XCO2.