Introduction

With a total production of 2.2 Moz Au until 2000, the early to Mesoarchean terrane of the Pilbara Craton in Western Australia (Fig. 1) has historically shown lower Au production in comparison to the Neoarchaean Eastern Goldfields Province of the Yilgarn Craton in Western Australia (130 Moz Au until 2000; Goldfarb et al. 2001; Huston et al. 2001). As a result, most of the lode gold mineralization studies have been conducted within Neoarchaean, Yilgarn Craton leaving Mesoarchean terrane relatively unexplored. Archaean lode gold mineralizations represent a coherent class of gold deposit called ‘orogenic gold deposit’ that developed over a crustal depth range from granulite to sub-greenschist facies environments (crustal continuum model; Groves 1993). According to this model, gold mineralization resulted from a syn-tectonic epigenetic emplacement process associated with fluids channelled along crustal-scale shear zones (Goldfarb et al. 2001; Groves 1993; Kerrich and Wyman 1990). Fluid inclusions and stable isotopes studies supporting this interpretation remain uncertain concerning the source and composition of the ore-related fluid, suggesting either a surface-derived, or a metamorphic, or a magmatic fluid (Groves and Phillips 1987; McCuaig and Kerrich 1998). Multi-techniques including μ-SR-XRF are essential to further constrain the source and composition of ore fluids. This paper documents the composition and the origin of the fluids associated with the lode-gold within the Warrawoona Syncline in the Eastern Pilbara (EPGGT, Fig. 1). Our multi-techniques analytical approach shows that quartz veins preserve the fingerprints of three main stages of fluid infiltration involving a primary aqueous-carbonic fluid of deep origin (magmatic and/or metamorphic) and two aqueous fluids of surficial origin. Based on our data set, we show that the fluid chemistry and quartz isotopic composition associated with gold deposit from the Mesoarchean Warrawoona Syncline are similar to that of far more richly endowed Neoarchaean orogenic gold deposits.

Fig. 1
figure 1

Simplified geological map of the North Pilbara terrane modified after Van Kranendonk et al. (2002). EPGGT East Pilbara Granite–Greenstone Terrane, CPTZ Central Pilbara Tectonic Zone, WPGGT West Pilbara Granite–Greenstone Terrane, MB Mallina Basin, MCB Mosquito Creek Basin, KT Kuranna Terrane, MGB Marble Bar Greenstone Belt

Full size image

General setting

Geological and structural setting

The Pilbara Craton (Western Australia) is a well-exposed 3.6–2.8 Ga granite–greenstone terrane overlain by volcanic and sedimentary rocks of the Hamersley Basin (2.77–2.40 Ga, Fig. 1). The East Pilbara Granite–Greenstone Terrane (EPGGT, Figure 1) is characterized by circular to elliptic granitic domes, 40–120 km in diameter, encircled by variably deformed greenstone belts. Granitic domes include 3.46 to 3.43-Ga-old TTG gneisses and younger syn-kinematic and post-kinematic suites of high-K granitic rocks emplaced between 3.32 and 3.30 Ga (Fig. 2) and derived through partial melting of the early TTG (Hickman and Van Kranendonk 2004; Smithies et al. 2003; Van Kranendonk et al. 2002).

Fig. 2
figure 2

The study area in the Warrawoona Syncline. Geological map of the Warrawoona Syncline

Full size image

Gold-bearing quartz veins from the Warrawoona Syncline are the focus of this study. The Warrawoona Syncline is located between the Mount Edgar and Corunna Downs Granitic Complexes (Fig. 2) and consists of ultramafic, mafic and felsic volcanic rocks and minor sedimentary and chert units that are ascribed to the ca. 3.51–3.43 Ga Warrawoona Group and 3.35–3.31 Ga Kelly Group (Hickman 1983, 2001; Van Kranendonk et al. 2004). The volcanic rocks preserve primary igneous features such as pillows, vesicules and spinifex textures. Facing directions on both limbs point away from the granitic complexes towards the core of the syncline (Hickman 1983, 2001). This syncline is asymmetric, with limbs involving rocks of contrasting ages, metamorphic facies and strain intensities. On the northern limb, intensely deformed amphibolite facies mafic and felsic schists of the lower part of the Warrawoona Group are in contact with the southern margin of the Mount Edgar Granitic Complex across a kilometre-scale shear zone. The southern limb consists of the upper part of the Warrawoona Group and the overlying Kelly Group. The southern limb is intruded by ca. 3.30 Ga granitic rocks of the Corunna Downs Granitic Complex.

At a regional scale, the Warrawoona Syncline shows a main, steeply dipping, foliation (Sm) parallel to the granitic/greenstone boundaries. A pronounced mineral and stretching lineation (Lm) rotates and converges toward a small area where the lineation is vertical (Collins et al. 1998; Teyssier et al. 1990; Zegers et al. 2002; Fig. 2). The finite strain ellipsoid changes from S tectonites near the contact with the granitic complexes to L tectonites where Lm is vertical (Collins et al. 1998; Teyssier et al. 1990; Zegers et al. 2002). Open to isoclinal folds occur at all scales. Folds axial planes are parallel to Sm and fold axes are parallel to Lm (Fig. 2). Locally, later folds refold these regional fabrics.

The S and SW rim of the Mount Edgar is affected by a kilometric-scale shear zone (Fig. 2) with down-dip stretching lineation. Kinematic analysis reveals that the shear zone accommodated the exhumation of the granitic dome (Collins 1989; Collins et al. 1998; Teyssier et al. 1990). Within the greenstones, hectometre to kilometre-scale shear zones involve oblique horizontal dextral and sinistral sense of shear generating pop-up structures (Kloppenburg et al. 2001). As the stretching lineation rotates to become vertical and the finite strain ellipsoid evolves to constriction, the horizontal shear component disappears. Radiometric data (U–Pb zircon on syn- and post granites and Ar–Ar on metamorphic fabrics) constrain the age of the formation of the syncline between 3.32 and 3.30 Ga (Collins et al. 1998; Kloppenburg et al. 2001).

Fluid–rock interactions and lode gold mineralization

The Warrawoona Syncline is one of the largest mafic-ultramafic-hosted goldfields in the EPPGT. This district has produced 745 kg of Au from 25 kt of ore, with an average grade of 29.6 g/t, and an additional 22 kg of alluvial Au (Hickman 1983; Huston et al. 2001). More recently, a geological resource of 9.95 Mt at 1.0 g/t has been identified for the Klondyke deposit (Huston et al. 2001). The ore deposits are mainly composed of lodes hosted within kilometric shear zones (Huston et al. 2001). The age of the mineralization is loosely constrained by three Pb model ages from galenas: 3.38, 3.37 and 3.05 Ga (Huston et al. 2001). In a recent study, oxygen isotopic and geochemical analyses on whole rock and quartz veins were combined with structural observations to constrain the fluid circulation and lode gold mineralization history (Thébaud et al. 2006). This study suggests that the Warrawoona Syncline was the locus of a protracted and polyphased fluid circulation history involving at least two stages of paleo-fluid infiltration of contrasted temperatures and sources. Alteration halo around the main shear zone (Fielding’s Find Shear Zone) is characterized by an intense silicification, HREE depletion and 18O enrichment. These features are interpreted as the result of an alteration process driven by low-temperature hydrothermal convection probably involving seawater. This first stage was followed by a veining episode related to the syn-tectonic infiltration of deep sourced fluids characterized by quartz veins with quartz δ18O values centred around a mean value of 13.2‰, significantly lower than their silicified host rocks. These quartz veins represent the event responsible for the bulk of economic lode-gold formation; however, no definitive constraints exist on the ore-fluid source.

Quartz veins

Quartz veins are ubiquitous in the Warrawoona Syncline. Based on cross-cutting relationships, four main quartz vein populations have been identified (Fig. 3).

Fig. 3
figure 3

Sketches of the different structures and associated vein types (V1–V4) occurring from west to east in the Warrawoona Syncline. S0 and Sm refer to local observations of bedding (S0) and foliation (Sm) relationships (see text)

Full size image

Type 1 veins (V1) can be found throughout the Warrawoona Syncline, but are mainly present along shear zones. Up to a few meters long, they are composed of quartz with minor calcite and ankerite frequently replaced by brown Fe-oxides and sulphides. Locally, they host native gold. These veins are parallel to Sm but can be locally boudinaged or folded. These structural features suggest that V1 were emplaced prior to or during the deformation. Nevertheless, the absence of quartz veins in the less deformed Marble Bar Greenstone Belt (Fig. 1) supports a syn-deformation origin.

In contrast to V1, type 2 veins (V2) are oriented at high angle to Lm and therefore perpendicular to Sm. V2 veins are < 1 m thick extensional fractures occasionally forming ‘en échelon’ vein sets and are composed of quartz with minor sulphides. V2 veins are progressively reoriented following the attitude of the stretching lineation (Lm) (Fig. 4b–d). Locally, quartz fibres are sub-parallel to the stretching lineation suggesting that V2 formed during the development of Lm and thus, as V1, can be considered as syn-deformation quartz veins formed in the brittle–ductile domain.

Fig. 4
figure 4

Outcrop scale photographs of hydrothermal veins encountered in the Warrawoona Syncline. a Deformed V1 quartz vein cross-cut by V2. b View north of steeply west-dipping V2 extensional quartz veins, plunging ∼ 30° toward the ESE at high angle to Lm (toward the person). c Sub-horizontal V2 in the core of the syncline, at high angle to vertical Lm. d Stereographic projections of V2 orientation (poles to planes) from west (top) to east (right) in the Warrawoona Syncline highlighting the progressive change in orientation from steeply dipping to sub-horizontal toward the core of the Warrawoona Syncline (see text). e V3 breccia vein along Fielding’s Find Shear Zone at the outcrop scale; black arrow points to a person for scale. f Detail of a V3 breccia vein. g V4 “massive” type quartz veins

Full size image

Type 3 veins (V3) occur as irregular-shaped breccias forming the core of the shear zones in particular within the Fielding’s Find Shear Zone running parallel to the syncline axial surface (Fig. 4e, 4f). V3 veins consist of pervasive and randomly oriented vein sets within a highly silicified and fuchsite-enriched wall rock. The breccias vary in size but can be followed almost continuously for kilometres along strike, as in the Fielding’s Find Shear Zone. The localization of V3 in the core of the shear zones is best interpreted in terms of a late emplacement possibly associated with hydraulic fracturing during brittle faulting.

Type 4 veins (V4) form massive quartz pods tens of meters in length (Fig. 4g). Mainly composed of quartz with minor sulphides, they show little to no deformation and occur commonly in mafic material parallel to S1. V4 is considered to be the result of a late deformation feature postdating the formation of the main shear zones. As fluid inclusions in V4 were too small to be analysed, they are not integrated in this study.

Analytic approach and methodologies

Our analytical approach relies on the combination of routine geochemical techniques and more recently developed in situ analyses. The combination of microthermometry and Raman spectroscopy provides robust constraints on the inclusion fluid chemistry, including molar proportions of anhydrous gas species (CO2, CH4, N2) and the salinity, as well as constraints on pressure and temperature associated with the (minimal) trapping temperature. μ-SR-XRF techniques were used to constrain the nature and content of solutes (Cl, Br and K, Ca, Fe, Cu, Zn, As, Rb, Sr). Finally, oxygen isotopic analyses were used in conjunction with fluid inclusion analyses to discriminate the various fluid sources.

Microthermometry and Raman spectroscopy

Typology and petrography of fluid inclusions were documented from representative samples of the various quartz vein populations. Microthermometric analyses were performed on fluid inclusions observed in quartz in 16 doubly polished thin sections. Microthermometry was carried out on a Fluid Inc. Modified USGS heating–freezing stage attached to a Leitz Wetzlar Laborlux S binocular microscope. Fluid inclusion types were distinguished using the nomenclature of Boiron et al. (1992). Microthermometric results are presented using the following abbreviations: TmCO2 = melting temperature of solid CO2; ThCO2 = homogenization temperature of CO2 vapour to the liquid phase; Tmf = final melting temperature of ice or hydrate; Tmclath = melting temperature of clathrate; Th = temperature of final homogenization; Td = temperature of decrepitation. The rate of heating was kept under 0.2°C/s during freezing and under 1°C/s during heating. Salinity, expressed as equivalent wt% NaCl, was calculated from microthermometric data using the equation of Bodnar (1993). In vapour-bearing fluid inclusions, the presence of CO2 amongst other anhydrous phases (CH4, H2S) was detected on the basis of the melting temperature of the solid anhydrous phase during freezing and melting cycles. Salinity was calculated using the clathrate melting curves of Collins (1979). The volumetric fraction of the aqueous liquid in carbonic fluid inclusions was estimated using the volumetric chart of Roedder (1972).

Molar fractions of CO2, CH4, H2S and N2 were determined in individual inclusions by micro-Raman analyses performed with the Labram Raman spectrometer at CREGU at Nancy, France. Bulk compositions and molar volumes were computed from P–V–T–X properties of individual fluid inclusions in the C–O–H–S system using the equations of Bakker (2003), Dubessy (1984) and Dubessy et al. (1989).

Synchrotron radiation X-ray fluorescence

Synchrotron radiation X-ray fluorescence is a non-destructive and in situ technique that allows, thanks to the high synchrotron resolution, to achieve reliable quantitative single fluid inclusion analyses. μ-SR-XRF measurements were carried out at the μ-fluorescence, μ-imaging and μ-diffraction beamline (ID22) of the European Synchrotron Radiation Facility (ESRF). A double-crystal fixed-exit monochromator was used to create a monochromatic beam of 17 keV. The beam was focused with a KB mirror to a 1.5 × 3 μm spot. Focused beam intensity was 3 × 1011 photons/s. X-ray fluorescence spectra were recorded with a Gresham Si(Li) detector of 140 eV resolution, placed at 90° angle to the incident beam in the polarization plane to reduce X-ray scattering. Detection of S, Cl, K and Ca was improved by placing the sample and the semi-conductor detector under a helium atmosphere, which greatly reduces the absorption of low-energy fluorescence emitted by light elements.

Detailed information on the experimental protocol, absorption correction, calibration procedures and error calculations can be found in Cauzid et al. (2004, 2006), Ménez et al. (2002) and Philippot et al. (1998). As the quartz matrix absorbs elements fluorescence, elemental ratios are strongly dependent on the length of the X-ray path through the quartz (i.e. inclusion depth). To minimize this absorption, inclusions (8–20 μm in diameter) were brought to within 1–10 μm of the polished quartz surface. Calibration of the X-ray spectrum was achieved using the Cl concentration measured by microthermometry in each inclusion. When these data were not available, solutions of known composition filled in quartz capillaries were used as external standards (Cauzid et al. 2004; Mavrogenes et al. 1995). Error associated with these calibration procedures were at most ± 20%. The thickness of each inclusion and their depth relative to the polished surface were optically estimated with an error of ± 1 μm. Error on the inclusions depth propagates into a content error of 9–22% depending on the mass of the elements (Cauzid et al. 2006). Total errors were calculated using the error propagation formula detailed in Cauzid et al. (2006). These are ± 20–50% for K, Ca, Fe, Cu, Zn, Br, As, Rb and Sr (Table 2), ± 20% for Cl (from microthermometry) and 43% for Br/Cl ratios.

Oxygen isotope data

δ18O data for the quartz minerals were obtained on a LaserPrep system on line to a VG Isotech (now GV instruments) Optima dual inlet (Mattey 1997) at the Royal Holloway University of London (Egham, UK). Samples (1.7 g) were combusted using a CO2 laser in the presence of excess BrF5. The laser spot is approximately 250 μm in diameter. The liberated O2 passed through cryogenic traps for clean up before being directly analysed in the IRMS. Samples below 90% of the expected yield were rejected. Three mineral standards have been analysed during the runs to calibrate the data. Two are internal, GMG II (a garnet) and QBLC (a quartz) and the international NBS-30 (a biotite). All δ18O values are reported relative to V-SMOW. The overall precision on standards and sample replicates is better than ± 0.1‰.

Results on fluids characteristics

Fluid inclusion textural settings and compositions

Based on textural observations and fluid compositions, three types of fluid inclusions Lw-c, Lw-1 and Lw-2 have been recognized, irrespective of the host rock lithologies. In this section we present their microthermometric, Raman spectroscopy (Table 1) and μ-SXRF analyses (Table 2).

Table 1 Microthermometric data and chemical compositions obtained by Raman microprobe spectrometry for Lw-c fluid inclusions
Full size table
Table 2 μ-SR-XRF results (ppm)
Full size table

Carbonic Lw-c fluids are common in V1 and V2 (Fig. 5a, b). In contrast to V1 and V2, V3 contains rare and isolated Lw-c inclusions. Lw-c appear either as aligned inclusion along healed micro-fractures terminating at the edge of a single quartz grain or as irregular clusters in quartz grains. These features support a primary or a pseudo-secondary origin. Lw-c shows CO2 melting temperatures in the range of − 60.0 to − 56.6°C indicating the presence of anhydrous species such as CH4, H2S or N2 in addition to CO2 (Fig. 6a). Raman spectrometry revealed that CH4 and in lesser extent N2 are the two anhydrous species contributing to lowering the CO2 melting temperature. Clathrate melting temperatures between 6 and 10°C (Fig. 6b) yielded salinities between 0 and 7 wt% NaCl equivalent (Fig. 8) (Collins 1979). Final homogenization in the liquid phase occurs between 234 and 272°C (Figs. 6c, 7). Decrepitation occurred between 195 and 340°C, in most cases prior to homogenization. These temperatures can be seen as minimum trapping temperatures. The wide range of decrepitation/homogenization temperature and the homogenous salinity suggest that Lw-c may have been trapped during successive stages of micro-fracturing (Fig. 7). Lw-c inclusions contain various amounts of Ca, K and transition metals (Cu, Zn, Fe) (Figs. 11, 12) and are enriched in As (up to 1,700 ppm) which is often associated with lode gold deposits. K/Rb ratio range from 2 to 400, with most data comprised between 100 and 400.

Fig. 5
figure 5

Transmitted light photomicrographs and sketches of fluid inclusions from quartz veins sampled in the Warrawoona Syncline. a Pseudo-secondary carbonic fluid inclusions (Lw-c) trails in V1. b Early secondary carbonic fluid inclusion trails (Lw-c) and aqueous secondary Lw-2 fluid inclusions in V2. c Decrepitated primary fluid inclusions (Ld) located along lining overgrowth zones, and Lw-1 located along healed microfractures. d Secondary aqueous Lw-2 fluid inclusion trails in V3

Full size image
Fig. 6
figure 6

Microthermometric data of Lw-c fluid inclusions from V1 and V2. a Frequency histogram of TmCO2. b Frequency histogram of TmClath. c Frequency histogram of ThCO2

Full size image
Fig. 7
figure 7

Salinity versus Th and Td plot of aqueous and carbonic fluid inclusions

Full size image

Lw-1 aqueous inclusions have been observed within a few V1 veins and in one V3 vein. Lw-1 occurs as inclusion trails along micro-fractures cutting across several quartz grains and primary inclusion trails (Fig. 5). Therefore Lw-1 can be seen as secondary inclusions. Ice melting temperature between − 16.1 and − 1.5°C yielded moderate salinities between 2.5 and 19.5 wt% NaCl equivalent. Homogenization temperatures of Lw-1 are within the range of 126–157°C. μ-SXRF data indicate that Lw-1 inclusions are depleted in all elements except Cl and K, and in lesser extent Ca and Fe.

Lw-2 occurs in all veins as secondary fluid inclusion trails along healed microfractures (Fig. 5). Ice melting temperature between − 6.2 and 0°C yielded salinities between 0 and 8.9 wt% NaCl equivalent with most of the analyses at around 2 wt% NaCl equivalent. Similar to Lw-1, Lw-2 homogenized in the liquid phase between 126 and 157°C. It is worth noting that in a salinity versus Td and Th diagram, Lw-1 and Lw-2 present a wide range of salinity that overlap each other (Fig. 7). This distribution may reflect an isothermal mixing process between two different fluid sources. This is further supported by the composition of Lw-2 showing a marked enrichment in K, Rb, Sr, Zn and Br when compared to Lw-1 (Fig. 11, Table 2).

In summary, three main fluid inclusion populations have been found within the various quartz vein families (Fig. 8). (a) Lw-c is a H2O − CO2 ± CH4 ± N2 primary fluid with a relatively high-density volatile phase, a moderate to low salinity and a significant enrichment in K and base metals. (b) Lw-1 is a moderate to high salinity aqueous fluid depleted in most trace elements. (c) Lw-2 is a low-salinity aqueous fluid showing variable enrichment in trace and major elements. On the basis of microthermometric and Raman spectroscopy, past studies of quartz carbonate veins from the Klondyke mine documented similar primary carbonic and secondary aqueous fluid inclusion populations (Huston et al. 2001).

Fig. 8
figure 8

Synthetic diagram showing the relationship between vein type, fluid inclusion occurrence and microthermometric results

Full size image

Vein quartz oxygen isotopic composition

Oxygen isotopic compositions of vein quartz from the Warrawoona Syncline are given in Table 3. Figure 9 shows that the different types of veins display a narrow range of δ18O values between + 11.5 and + 16.6‰ with a mean value at + 13.2 ± 2‰ (Fig. 9a). This suggests that quartz precipitated from a homogeneous fluid under near-isothermal conditions. The latter interpretation is reinforced by the clustered homogenization temperature distribution recorded by the primary fluid inclusions. The slightly higher δ18O value of the quartz veins hosted in felsic, compared to that hosted in mafic rocks, may indicate a slight buffering by the host rock (Fig. 9b).

Table 3 Vein quartz oxygen isotopic values
Full size table
Fig. 9
figure 9

Quartz δ18O values of a the different vein types and b veins by different lithology sampled in the Warrawoona Syncline

Full size image

Discussion

Potential fluid sources

There is an ongoing debate on the source of metals and ore fluids in Archaean lode gold deposits. Three alternative sources have been suggested: (1) a meteoritic and/or seawater source (e.g. Gebre-Mariam et al. 1993; Hagemann and Ridley 1993; Nesbitt and Muehlenbachs 1989); (2) a magmatic source (e.g. Burrows et al. 1986; McNaughton et al. 1990; Meinert 1993) and (3) a metamorphic source (e.g. Fryer et al. 1979; Groves and Phillips 1987; Phillips et al. 1996). Traditionally deep sourced fluid model have been mostly favoured. However, it remains often difficult to discriminate between fluids of magmatic or metamorphic origin.

Commonly, the minimum temperature of entrapment measured for fluid inclusion is used in conjunction with the mineral oxygen isotopic composition to constrain the δ18O value of the fluid from which it precipitated. In the Warrawoona Syncline, quartz veins precipitated from a near-isothermal fluid between 250 and 400°C as shown by primary fluid inclusion analyses from Lw-c. Using Zheng (1993) fractionation equation, the δ18O of the fluid ranges from + 2.4 to + 12.1‰ (Fig. 10). This δ18O range corresponds to a fluid in equilibrium with a crystalline basement at temperature in excess of 400–500°C. Alternatively, the vein quartz could have obtained its δ18O value from the infiltration of secondary aqueous fluids. Using a fluid temperature between 120 and 160°C as shown by Lw-1 and Lw-2 measurements, the fluid δ18O calculated ranges from − 7.3 to 2‰ (Fig. 10). Values around 0‰ are indicative of a meteoric or seawater-derived fluid. On a δ18O versus δD diagram, the range of the quartz forming fluid δ18O thus overlaps with the surfacic, magmatic and metamorphic fluid fields. It is thus essential to explore the chemistry of fluid inclusions using μ-SR-XRF.

Fig. 10
figure 10

δD and δ18O diagram showing the fluid composition calculated from Zheng (1993) quartz fractionation equation (see text)

Full size image

Fluid compositions and sources

The major premise arising from individual fluid inclusion analysis using μ-SR-XRF technique is to provide chemical information on the different aqueous fluids or the aqueous portion of carbonic fluid inclusions. With such information, it is then possible to discuss the composition of potential fluids reservoirs. Furthermore, some fluid reservoirs have characteristic elemental fingerprints. In addition, as the concentration of Cl and Br is higher in seawater and sedimentary aqueous brines than crustal rocks, the Cl/Br ratio is little affected by fluid–rock interaction processes and therefore can be used as a source tracer (Banks et al. 2000; Yardley and Graham 2002).

Primary fluid inclusions: Lw-c

Lw-c inclusions show a marked enrichment in K relative to Ca and Fe (Fig. 11). Potassium-rich fluids are typical magmatic sources (e.g. Bingham, UT, Butte, Montana and St. Austell, Cornwall; Irwin and Roedder 1995; Sherlock et al. 1999; Solomon et al. 2004; Zaw et al. 2003). In addition, the Cl/Br ratios measured for Lw-c fluid inclusion plots within the field of Au-bearing quartz veins from Alleghany (Böhlke and Irwin 1992) and that of volcanic fumarole condensates compiled by Böhlke and Irwin (1992) (Fig. 13). This suggests that Lw-c may be of magmatic origin. However, an additional metamorphic component is likely involved as some data plot within the field of Ouro Fino retrograde metamorphic quartz veins (Boiron et al. 1999). In addition, calculated K/Rb ratio range between 100 and 400, which overlap with the uniform crustal-like K/Rb ratio (220–400) measured in many other lode gold deposits. This uniform K/Rb ratio has been interpreted as the product of dehydration reactions within a complex lithological rock succession (Kerrich and Fryer 1988; McCuaig and Kerrich 1998), hence strongly suggesting that the Lw-c inclusions comprise an important component of metamorphic origin.

Fig. 11
figure 11

μ-SR-XRF results. Individual fluid inclusion analysis shown as a K–Ca–Fe and b Fe–Cu–Zn diagrams

Full size image

Secondary fluid inclusions: Lw-1 and Lw-2

Lw-1 inclusions are poor in Fe, base metals, metalloids or other trace elements (Fig. 12). Such a depleted fluid signature advocates for a surface origin (meteoric/seawater). On a Cl/Br versus Cl diagram Lw-1 shows a wide range of Cl/Br and Cl concentration values lining along a modern day seawater evaporation trend. In addition, Lw-1 overlaps the Canadian Shield brines field (Frape and Fritz 1987) and that of Archaean saline seawater (Foriel et al. 2004; Fig. 13). This suggests that Lw-1 represents a surface fluid with Archaean seawater affinities. In addition, in a K, Ca, Fe diagram, Lw-1 displays a wide range of composition along the K–Ca line suggesting a significant K enrichment (Fig. 11). The K enrichment as well as the lowering of the Cl/Br ratio could result from fluid interaction with felsic and/or granitic rocks involving the preferential Cl substitution in hydrous silicates. Alternatively it may highlight a fluid mixing episode with a K and Br enriched fluid.

Fig. 12
figure 12

μ-SR-XRF results. Individual fluid inclusion analysis. Element content in (ppm)

Full size image
Fig. 13
figure 13

μ-SR-XRF results. Individual fluid inclusion analysis. Cl/Br versus Cl with (1) Alleghany magmatic Au-quartz veins (Böhlke and Irwin 1992), (2) Ouro Fino retrograde metamorphic quartz veins (Boiron et al. 1999); (3) Canadian shield brines (Frape and Fritz 1987), (4) North Pole seawater (Foriel et al. 2004), (5) Volcanic fumarole condensates ranges compiled by Böhlke and Irwin (1992)

Full size image

Lw-2 inclusions are depleted in Ca but show significant enrichment in K, Zn, Sr, Rb, Br ± Cu and As (Figs. 11, 12). Lw-2 compositions are typical of granite-buffered fluids (Frape and Fritz 1987). Lw-2 shows also a significantly lower Cl/Br ratio and a marked enrichment in Br compared to Lw-c and Lw-1 (Figs. 12, 13). This high Br and low Cl content may not result from the Cl/Br fractionation due to chloride precipitation, as it does not fit the evaporation line. Alternatively, it may result from the preferential Cl substitution in hydrous silicates as proposed for low-temperature brines percolating through crystalline rocks (Frape and Fritz 1987).

With regard to these results, we suggest that Lw-1 and Lw-2 fluids represent two end-members of surficial fluids, possibly of seawater origin, showing different degrees of interaction with the host rocks. According to such interpretation Lw-1 could represent a rather unbuffered pole. On the other end, Lw-2 could represent a second end-member significantly buffered by felsic and/or granitoid lithologies as attested by the low Cl/Br ratio, low salinity and high K, Sr, Rb concentrations. Accordingly, the data dispersion observed in the Cl/Br versus Cl and K, Ca, Fe diagram can either illustrate intermediate stages of fluid–rock interaction or the mixing of both fluid end-members.

The veining episode

Fluid infiltration associated with the veining episode is synthesized in Fig. 14. Fluid inclusions preserved within the quartz veins document at least three fluid infiltration episodes (Fig. 14c, d). The first one is represented by primary carbonic fluid inclusions preserved within all quartz veins types. Our fluid inclusion data indicate that the fluid responsible for the vein forming fluid is of mixed magmatic and metamorphic origin. These fluids have probably been produced during the granitic dome emplacement and associated greenstone devolatilization. In this context, simple calculations show that the heat of crystallization would have released far more water through metamorphic dehydration reactions than released through the crystallizing magmas (McCuaig and Kerrich 1998). Accordingly, Lw-c fluid inclusions may record a small volume of magmatic fluid with a pronounced halogen signature mixed with large volumes of metamorphic fluids with a low halogen signature.

Fig. 14
figure 14

Two-stage model of carbonic versus aqueous infiltration in and around the Warrawoona Syncline shear zones (a and b from Thébaud et al. 2006, see text)

Full size image

The second and third fluid infiltrations are highlighted, respectively, by secondary Lw-1 and Lw-2 aqueous fluid inclusion that reflect a fluid infiltration from shallow seawater reservoir. Lw-2, which is preserved in all quartz veins, is fairly similar to modern shield brines. The characteristics of Lw-2 fluids are best interpreted in terms of the interaction between a surface fluid and felsic rocks (Fig. 14b). In this regard, we suggest that the interaction of a seawater fluid (Lw-1) with crystalline rocks may have led to the formation of the Lw-2 fluid type. The percolation of these surficial fluids has likely been enhanced through a late fracturing event. Flushed within the Warrawoona Syncline, these two fluids may have mixed together thus explaining the composition range.

Lode gold mineralization implications

The fluid inclusions indicate that the fluid responsible for the lode gold mineralization was a carbonic fluid of mixed metamorphic and magmatic origin. This fluid presents similar fluid characteristics to those generally described in typical Neoarchaean orogenic lode gold deposits (e.g. Groves 1993; Groves and Phillips 1987; McCuaig and Kerrich 1998). In addition, vein quartz δ18O values of + 14 ± 1‰ are also typical of many Meso- to Neoarchaean lode gold deposits (de Ronde et al. 1992; Kenich 1987; McCuaig and Kerrich 1998). Taken out of context, this fluid supports the epigenetic crustal continuum model. However, we have seen that this fluid infiltration is the last component of a polyphased and protracted fluid circulation history (Thébaud et al. 2006). This polyphased fluid circulation is not compatible with such a model.

Polyphased and protracted fluid circulation history have been documented elsewhere in the Pilbara (Zegers et al. 2002), the Barbertone Greenstone Belt (Hutchinson 1993; Viljoen 1984) and the Yilgarn (Bateman and Hagemann 2004; Brown and Johnson 2003; Hutchinson 1993; Thébaud et al. 2006). It has been argued that such a polyphased and protracted genetic process is a necessary condition for the formation of economic deposits (Bateman and Hagemann 2004; Hutchinson 1993; Thébaud et al. 2006; Viljoen 1984). We suggest that the epigenetic interpretation favoured in most Archaean cratons only account for the latest fluid infiltration stage preserved in the quartz lodes, therefore missing the possible role of earlier fluid–rock interactions. In other terms, the fluid properties that characterize Archaean lode gold deposits are not discriminatory as they are compatible with both epigenetic and polyphased genetic scenarios.

Conclusion

This paper documents the fluid infiltration recorded within the syn-tectonic quartz veins associated with lode gold mineralization in the Warrawoona Syncline. Oxygen isotopic and fluid inclusion analyses including microthermometry, Raman spectroscopy and Synchrotron radiation X-ray fluorescence were used to decipher fluids composition and source. Our results suggest that three main stages of fluid infiltration were recorded in the quartz veins: a primary ore-related carbonic fluid infiltration is interpreted as of dominantly magmatic origin and two secondary aqueous fluid infiltrations of surface origin (seawater) exhibiting various levels of fluid–rock interaction. The composition of the primary fluid presents similar chemical and oxygen isotopic properties than that preserved in other Archaean lode gold deposits. However, in regard to the protracted history documented in Meso- to Neoarchean cratons, we question the robustness of the epigenetic crustal continuum model. We suggest ore-fluid characteristics equally support an epigenetic or a polyphased mineralization process