Invisible gold in arsenian pyrite and arsenopyrite from a multistage Archaean gold deposit: Sunrise Dam, Eastern Goldfields Province, Western Australia

Abstract

The Sunrise Dam gold mine (11.1 Moz Au) is the largest deposit in the Archaean Laverton Greenstone Belt (Eastern Goldfields Province, Yilgarn Craton, Western Australia). The deposit is characterized by multiple events of fluid flow leading to repeated alteration and mineralization next to a major crustal-scale structure. The Au content of arsenian pyrite and arsenopyrite from four mineralizing stages (D₁, D₃, D_4a, and D_4b) and from different structural and lithostratigraphic environments was measured using in situ laser ablation inductively coupled plasma mass spectrometry. Pyrite contains up to 3,067 ppm Au (n = 224), whereas arsenopyrite contains up to 5,767 ppm (n = 19). Gold in arsenopyrite (D_4a stage) was coprecipitated and remained as “invisible gold” (nanoparticles and/or lattice-bound) during subsequent deformation events. In contrast, gold in pyrite is present not only as “invisible gold” but also as micrometer-size inclusions of native gold, electrum, and Au(Ag)–tellurides. Pristine D₁ and D₃ arsenian pyrite contains relatively low Au concentrations (≤26 ppm). The highest Au concentrations occur in D_4a arsenian-rich pyrite that has recrystallized from D₃ pyrite. Textures show that this recrystallization proceeded via a coupled dissolution–reprecipitation process, and this process may have contributed to upgrading Au grades during D_4a. In contrast, Au in D_4b pyrite shows grain-scale redistribution of “invisible” gold resulting in the formation of micrometer-scale inclusions of Au minerals. The speciation of Au at Sunrise Dam and the exceptional size of the deposit at province scale result from multiple fluid flow and multiple Au-precipitating mechanisms within a single plumbing system.

Introduction

Over the past decade, the Laverton Greenstone Belt (LGB) in the Eastern Goldfields Province of the Yilgarn Craton, Western Australia (Fig. 1) has become a major center of gold production with a total Au endowment of over 700 t Au (Salier et al. 2005). The belt contains two world-class deposits (>100 t Au), Sunrise Dam and Wallaby. The Sunrise Dam orogenic gold deposit is located approximately 850 km ENE of Perth and 45 km SSE of Laverton at 290°5′ S, 122°25′ E. With a total production by March 2007 of approximately 5.4 Moz Au and remaining resources + reserves of 5.58 Moz Au, the mine is the single largest deposit in the LGB.

Fig. 1

Simplified geological map of the Laverton Greenstone Belt (modified after Salier et al. 2005), showing the sectors suggested by Cassidy et al. (2002) and the position of Laverton Tectonic Zone (LTZ) that hosts the majority of gold deposits in this area, including Sunrise Dam. Inset: position of the area shown in the map (rectangle) in the Yilgarn Craton and Eastern Goldfield Province (EGP)

From the perspective of mineral processing, gold ores can be broadly classified as free milling (over 90% recovery) and refractory (less than 90% recovery; Vaughan 2004). Refractory gold ores have long been recognized from a range of Au deposits with different geological environments, genetic histories, and ages (e.g., Fleet et al. 1993; Fleet and Mumin 1997; Cabri et al. 1989; Cathelineau et al. 1989; Cook and Chryssoulis 1990; Arehart et al. 1993; Oberthür et al. 1997; Tarnocai et al. 1997; Genkin et al. 1998; Ashley et al. 2000; Simon et al. 1999a, b; Pals et al. 2003; Ciobanu et al. 2009). Submicroscopic or “invisible” (less than 0.1 μm) Au is by far the most common reason for refractory behavior, although low Au recovery may also be caused by other factors, e.g., the presence of Au(Ag)–tellurides (e.g., Kalgoorlie, Western Australia; Shackleton et al. 2003), or of fine-grained inclusions of native gold (less than 10 μm) within sulfides.

The Sunrise Dam deposit has a complex tectonic evolution leading to a protracted history of Au deposition and remobilization (Nugus et al. 2005a, b). Although the deposit is generally regarded as low-sulfide type (Brown et al. 2002), Au occurs as native gold inclusions in sulfides/sulfosalts and also forms Au(Ag)–tellurides (Sung et al. 2007). The deposit contains both free-milling and refractory Au ores, and the average Au recovery is about 80%. The refractory Au ores at Sunrise Dam consist mainly of arsenian pyrite and arsenopyrite.

Consequently, a laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) study has been conducted on pyrite and arsenopyrite to evaluate the concentration and the nature of refractory Au in the deposit. This study provides the basic information about the nature of invisible gold necessary to understand the relatively low gold recovery rate. Combined with petrological observations, the LA-ICP-MS data enable us to discuss aspects of gold deposition and remobilization during the successive deformational stages and via different precipitation mechanisms. Lastly, this should provide an insight into the various mechanisms of Au enrichment in a world-class orogenic Au deposit.

The Sunrise Dam deposit

The Sunrise Dam deposit is located in the Archaean Laverton Tectonic Zone (LTZ), occupying the highly deformed central domain of the LGB (Fig. 1). The majority of the LTZ has undergone metamorphism at lower greenschist facies conditions (Hallberg 1985). The lithologies in the mine area comprise an intermediate volcanoclastic sequence that dips gently to the NW and is overlain by turbidite sedimentary rocks, the latter one of which includes banded iron formation (BIF) sequences (magnetite-rich shales and cherts; Fig. 2). Dikes and sills of dolerite and ultramafic rocks intrude the sedimentary rocks, whereas rhyodacite porphyry intrudes any lithological unit.

Fig. 2

Plan view of major lithologies and structures in the Sunrise Dam gold deposit showing projected locations of LA-ICP-MS samples. 1 GQ 1960 OD, 2 GQ open pit 67.5 RL, 3 Astro UGD 728, 4 Mako 2087 OD, 5 SSZ 2075 OD, 6 SSZ 2077 OD, 7 SSZ 2067 OD, 8 SSZ open pit W-wall, 9 Cosmo 1940 OD, 10 Dolly Volcanic 1931 OD, 11 Dolly Porphyry 1930 OD, 12 Margies open pit S-wall, 13 WSZ 2109 OD, 14 WSZ open pit S-wall, 15 WFZ open pit SE-wall, CUS Cleo Upper Shear, SSZ Sunrise Shear Zone, CS Carey Shear, WSZ Western Shear Zone, WFZ Watu Fault Zone. Mine grid is rotated 45° E of true north

Structures and orebodies

Gold mineralization occurs intermittently along a NE-trending corridor of ca. 4.5 km length, characterized by the presence of strongly magnetic BIF sequences (Newton et al. 2002). The 20 currently defined orebodies are centered onto a series of parallel, gently dipping (~30°), and NE-trending shear zones with a thrust-duplex architecture and high-strain character. These shear zones are, from west to east, Cleo Upper Shear (CUS), Sunrise Shear Zone (SSZ), and Carey Shear (CS; Fig. 2). Such gentle-dipping structures, characteristic of many orogenic gold deposits in the LGB (e.g., Wallaby, Granny Smith, Lancefield, Jubilee), usually carry a significant proportion of the ore. At Sunrise Dam, the SSZ alone hosts in excess of 1.5 of the 11 Moz Au contained in the whole deposit (Nugus et al. 2005a).

A number of orebodies in the deposit are hosted within steep shear zones, e.g., Western Shear Zone (WSZ) and Watu Fault Zone (WFZ) that in turn are located between the gently dipping CUS and the SSZ (Fig. 2). These lodes carry higher and more consistent grades (Au at 15 g/t in WSZ; Au at 13 g/t in WFZ) than those hosted by the SSZ.

All the above structures are NE-trending and contain ~80% of the Au resource whereas the rest is hosted by N–S shears, e.g., Summer Cloud, Predator, Hammer Head, and Cosmo orebodies. These contain an important part of the 1.71-Moz current underground resource (0.55 Moz at 8.96 g/t Au; Nugus et al. 2005a).

Gold deposition and deformation history

Four major deformational stages are currently recognized at Sunrise Dam. The earliest deformation (D₁) formed F₁ folds and thrusts (i.e., SSZ, CUS, and CS) in response to N–S compression and shortening (S₁; Nugus et al. 2005a). The second event (D₂) was characterized by regional E–W shortening associated with the formation of the regional Spartan Anticline (F₂ fold; see Fig. 1) and penetrative N–S trending cleavage (S₂). These two deformational stages were associated with hydrothermal alteration of country rocks (carbonate–chlorite–sericite) and only minor low-grade Au mineralization (generally <1 g/t). The two following stages, D₃ and D₄, are responsible for the bulk of the mineralization. D₃ produced major shear and/or fault zones such as WSZ and WFZ as a result of deformation caused by NNW–SSE to NW–SE oriented stress. This was also responsible for reactivation of S₁ and S₂ with sinistral strike-slip movement, as well as development of mineralized S₃ fractures and NE-trending extensional vein arrays. The last deformational event (D₄) resulted from NW–SE shortening with local dextral reactivation on S₃ surfaces and produced extensional Au lodes. The D₄ veins can be subdivided into steeper-dipping, early (D_4a), and gentle-dipping late veins (D_4b); D_4b veins crosscut the D_4a veins.

The D₃ and D₄ veins differ in terms of mineralogy, texture, and geochemical signatures. The vein filling consists mainly of quartz and carbonates in various proportions; quartz dominates in the D_4b veins. Pyrite is the main ore mineral present in all veins, arsenopyrite is particularly rich in D_4a, whereas base-metal sulfides and Pb–Sb sulfosalts are mainly found in the D_4b veins. Tellurides are present in both types of D₄ veins, but particularly abundant and diverse in D_4b veins (Sung et al. 2007). Coarse native gold is common in D_4a veins. Replacement of BIF by pyrite during faulting in the D₃ stage is another common aspect (e.g., Fig. 3a). Evidences of overlap between the D₄ and D₃ stages are widespread. For example, two types of veins are recognizable in Fig. 3b: (a) a thin, white, isoclinically folded D₃ vein and (b) a thicker D_4a vein consisting of laminated gray quartz; the gray color is due to fine-grained arsenopyrite. Characteristic of the D₄ veins are dark internal laminae (Fig. 3c). In detail, such laminae may have a stylolitic aspect and complex mineralogy, e.g., muscovite, tourmaline, apatite, rutile, as well as pyrite and minor native gold and tellurides. Post-D₄ deformation is also recognizable as oblique laminae crosscutting the internal fabric of the D₄ veins, e.g., Fig. 3d. It is on such internal structures that local-scale remobilization can lead to formation of the most diverse telluride assemblages (Sung et al. 2007).

Fig. 3

Mineralization styles in the Sunrise Dam mine. a Sulfidation of BIF (outlined by dotted lines) along a fault plane; 2067 ore drive, SSZ. b Laminated D_4a vein crosscutting folded D₃ vein, WSZ open pit. Note more highly deformed and bleached appearance of D₃ veins. c Typical high-grade D_4b vein showing stylolitic laminae. The numerous quartz veins in the host rock belong to the D₃ and D_4a stages. Also note the collocation of the D_4b vein with a subparallel breccia zone on the top. 2066 ore drive, SSZ. d Post-D₄ laminae affecting a D_4b vein

Country-rock alteration is similar in both D₃ and D₄ stages. Importantly, this is always associated with dissemination of fine-grained pyrite and arsenopyrite. The most intense alteration is sericite + ankerite/dolomite adjacent to veins, whereas distal alteration includes mainly chlorite + calcite. Chromium-bearing mica (“fuchsite”) is present preferentially in the central part of the deposit below the SSZ (e.g., SSZ, GQ, and Cosmo orebodies) and is usually associated with Cr-rich mafic dikes and sills and mafic volcanogenic lithofacies.

The geochronological framework established by Brown et al. (2002) indicates that the D₃ and D₄ mineralizations occurred in the interval ~2,670–2,650 Ma. Intrusion of rhyodacite porphyry at 2,674 ± 3 Ma (SHRIMP II U–Pb on zircons) predates the formation of D₄ veins (corresponding to D₃ in the nomenclature of Brown et al. 2002). This age overlaps with the 2,663 ± 11 Ma Re–Os age obtained on D₁ molybenite in quartz veins crosscutting the porphyry. Brown et al. (2002) consider that the main stage mineralization is constrained by a 2,667 ± 3 Ma age from ⁴⁰Ar/³⁹Ar dating of Cr-mica from the SSZ and a 2,654 ± 8 Ma SHRIMP II U–Pb age on hydrothermal xenotime and monazite from the WSZ.

Fluid sources and gold deposition mechanisms

Fluid inclusion data on samples from gently (SSZ) and steeply dipping (WSZ; “Western Lodes” in Brown et al. 2003) orebodies indicate that the chemistry of Au-bearing fluids was identical (Brown et al. 2003). The mineralization is considered to be formed from a CO₂–H₂O dominant, moderate-salinity fluid (6–9 wt.% NaCl equivalent) with homogenization temperatures ranging from 280°C to 320°C. Assuming no posttrapping inclusion modification, Brown et al. (2003) estimated trapping pressures of 80–130 MPa from isochores at 280°C, corresponding to a depth of formation of ~4 km under transient hydrostatic pressure. Based on variations in H₂O/CO₂ ratios among secondary inclusions with homogenization temperatures similar to the main fluid, Brown et al. (2003) suggest that Au deposition in the WSZ lodes may be related to intermittent phase separation, whereas for SSZ, precipitation of Au was mainly triggered by wall rock sulfidation of BIF units.

Samples selection, preparation, and analytical techniques

Representative ore samples were collected from ten orebodies (Table 1; Fig. 2), and a total of 138 polished blocks were prepared and examined under a Leitz Laborlux-12-Pol dual reflected–transmitted light polarizing microscope and Philips XL30 scanning electron microscopy (SEM) with energy dispersive X-ray spectrometry and back-scattered electron (BSE) imaging capabilities. Native gold (including electrum), pyrite, and arsenopyrite were analyzed with a CAMECA SX-51 electron microprobe (EMP) with wavelength dispersion spectrometers at Adelaide Microscopy Centre, University of Adelaide, Australia. The measurements were performed at 20 kV with a beam current of 19.5 nA, using the following X-ray lines and standards: FeKα (CuFeS₂), SKα (PbS), AsKα (CoAsS), SbLα (Sb₂S₃), and AuLα (Au).

Table 1 Sample locations and descriptions for the LA-ICP-MS analysis

A total of 260 LA-ICP-MS spot analyses were performed on 38 polished blocks representing different paragenetic stages and orebodies (Table 1). The analyses were carried out using an Agilent HP 4500 ICP-MS with a New Wave UP-213 Nd:YAG laser ablation system at CODES, University of Tasmania, Australia. Single-spot analyses were performed with a laser beam diameter between 20 and 40 μm, depending on the size of ablated mineral. The CODES standard STDGL2b-2 was used as primary standard in this study. It consists of powdered sulfides doped with certified element solutions and fused with lithium borate into a glass disc (Danyushevsky et al. 2003). Fe concentrations were used as internal standard. Data were collected during 100 s, including an initial 30 s of the carrier gas background before the laser was switched on. A total of 11 elements were monitored: ⁵³Cr, ⁵⁹Co, ⁶⁰Ni, ⁶⁵Cu, ⁷⁵As, ¹⁰⁷Ag, ¹²¹Sb, ¹²⁵Te, ¹⁹⁷Au, ²⁰⁸Pb, and ²⁰⁹Bi. The lowest values of minimum detection limit are (in ppm): Cr 0.393, Co 0.025, Ni 0.066, Cu 0.331, As 0.369, Ag 0.038, Sb 0.022, Te 0.302, Au 0.014, and Pb 0.009. The integration ranges used for calculating concentrations were selected to avoid inhomogeneities and spikes likely related to buried inclusions; examples of integration intervals are shown in Fig. 4. A more detailed interpretation of time-resolved LA-ICP-MS spectra is given below in section “LA-ICP-MS spectra”.

Fig. 4

Different types of LA-ICP-MS profiles obtained on Sunrise Dam pyrites (a–e) and arsenopyrite (f). The selected integration intervals for concentration calculations are also shown. a Pyrite crystal from the D₃ stage, showing smooth pattern for As and other trace elements, but a spikier pattern for Au, indicating the presence of suboptical gold particles. b, c Pyrite crystal from the D_4a stage (WSZ orebody) showing a relatively smooth and steady ablation profile for Au, attributed to invisible gold (i.e., Au in solid solution or as inclusion <0.1 μm). Note the positive correlation between the Au and As signals during the first 30 s of ablation in c. d D_4a stage pyrite crystal (WSZ) displaying the highest Au concentration measured in this study, due at least in part to the presence of submicrometer gold inclusions. e Pyrite from the SSZ orebody (2077 ore drive) closely associated with nagyágite in a D_4b stage quartz vein. LA-ICP-MS traces reveal sharp variations in Au signal that are mimicked by the Te and Pb signals. This is characteristic for the presence of buried inclusions of nagyágite. f Arsenopyrite from the D_4a stage, showing inhomogeneous Ni distribution, but homogeneous Au distribution. Variations in Fe and As signals indicate the presence of quartz inclusion among the fine-grained arsenopyrite

Composition of pyrite and arsenopyrite: results of LA-ICP-MS analyses

In this section, we summarize the concentrations and distribution of trace and minor elements in pyrite and arsenopyrite from the Sunrise Dam deposit, based mainly on LA-ICP-MS analyses. Sample location, mineralogy, stages, and host rocks are listed in Table 1. A summary of gold concentrations in pyrite from different stages is given in Table 2, all the data obtained for arsenopyrite are listed in Table 3, and the full dataset for pyrite is included in Appendix.

Table 2 LA-ICP-MS analyses of pyrite in the Sunrise Dam gold deposit (data in parts per million)

Gold and other trace elements in pyrite

The Au content of pyrite ranges from below detection limit (<0.014 ppm) up to 3,067 ppm, with an average of 1.61 ppm (based on 224 spot analyses; Table 2; Appendix). Unless noted otherwise, reported averages for trace elements in this paper were calculated on logarithmic concentrations, and analyses below detection limits were included at half the value of the detection limit. Histograms show that pyrite from each stage of mineralization has wide ranges in Au concentrations (Fig. 5). The few analyses of pyrite from D₁ (n = 6; average 1.12 ppm, range <0.3 to 6.56 ppm) show similar contents as D₃ (n = 53, average 0.56 ppm, range <0.04 to 25.8 ppm) and D_4b (n = 61, average 1.17 ppm, range <0.1 to 18.2 ppm). The D₁ dataset is limited because it is difficult to identify pristine D₁ pyrite. The average Au content for the D_4b stage reduces to 0.89 ppm if three points (circled in Fig. 5) that contain an order of magnitude more Au (140–290 ppm) and Te (499–6,426 ppm) than the other 59 points (sample Mk2-2, Mako orebody) are excluded. Such high Au and Te contents are attributed to calaverite inclusions, an interpretation that is consistent with the abundance of telluride inclusions in this sample (Table 1). The D_4a stage shows a significantly wider range in Au concentrations than the other stages, resulting in higher average concentration (n = 104, average 3.65 ppm, range <0.1 to 3,067 ppm). The highest value is an outlier and the mean becomes 3.42 ppm when this is excluded; this leaves the maximum value at 540 ppm.

Fig. 5

Distribution of some trace elements (LA-ICP-MS data) in Sunrise Dam pyrite as a function of ore stage and arsenopyrite from the D_4a stage. The circles in Py D_4b (Au, Te) emphasize three analyses in sample Mk2-2 with anomalously high Au and Te concentrations

Arsenic concentrations vary from 8 ppm to 6.7 wt.%, spanning nearly four orders of magnitude. The dataset does not reveal any significant difference in As distribution from one stage to another (Fig. 5). The wedge-shaped plot obtained for Au vs. As concentration in pyrite (Fig. 6) is similar to that known for pyrite in Carlin and epithermal Au deposits (Reich et al. 2005). The pyrite from D_4b plots at the upper part of the wedge and includes the highest As and Au values.

Fig. 6

Plot of Au vs. As concentrations for arsenian pyrite (as a function of ore stage) in the Sunrise Dam gold deposit analyzed by LA-ICP-MS. Solubility limit for gold is based on Reich et al. (2005). The five analyses plotting above the solubility limit were obtained on pyrite grains from the WSZ (D_4a) and Mako (D_4b) orebodies

Tellurium concentrations vary from <0.3 to 6,426 ppm, also spanning over four orders of magnitude (Figs. 5 and 7). As for Au, the highest Te contents are found in the D_4a and D_4b stages (Fig. 5), with D_4a showing the highest average Te concentration (8.17 ppm); the other stages have average Te concentrations at 4.41, 2.22, and 2.20 ppm for D_4b, D₃, and D₁, respectively. A broad positive correlation between Au and Te is revealed in Fig. 7b.

Fig. 7

Plots of Au concentrations vs. Te concentrations in arsenopyrite (a) and pyrite (b) from the Sunrise Dam deposit, as a function of ore stage and ore deposit. Two different Au-enrichment trends are clearly visible in the arsenopyrite dataset

In addition, the following trace elements were also measured in pyrite: Ni (11.8 ppm to 2.0 wt.%), Co (2.1 ppm to 1.3 wt.%), Cr (<0.7 to 3,900 ppm), Cu (1.7 ppm to 4.7 wt.%), Ag (0.15 to 3,310 ppm), Sb (0.4 to 5,260 ppm), Pb (0.9 to 5,750 ppm), and Bi (0.05 to 2,180 ppm; Table 2). These elements also have wide ranges in concentration that span several orders of magnitude but in this case, there are no obvious differences among stages (Fig. 5; Appendix). Prominent correlations are observed between Ni and Co (Fig. 8a) and also among the elements Pb, Ag, Sb, and Bi (e.g., Pb vs. Bi in Fig. 8b). In contrast, there is no correlation between Ni and Cr (Fig. 8c).

Fig. 8

Plots of the concentrations of some trace elements in Sunrise Dam pyrite and arsenopyrite. The symbols for pyrite distinguish among the orebodies and ore stages. a Ni vs. Co; b Bi vs. Pb; (c) Ni vs. Cr

Gold and other trace elements in arsenopyrite

Arsenopyrite is a major ore mineral in the D_4a stage and it occurs mainly in the central and deeper part of the deposit, e.g., WSZ, WFZ, Astro, Cosmo, Dolly Volcanic, Dolly Porphyry, and GQ orebodies. Of these, the WSZ and Dolly Volcanic orebodies are particularly As-rich, containing approximately equal amounts of arsenopyrite and pyrite.

The LA-ICP-MS data show significantly higher values of Au in arsenopyrite (Table 3) compared to pyrite; concentrations range from 0.81 to 5,767 ppm Au with an average of 562 ppm Au. In contrast to pyrite, the histogram of Au concentration in arsenopyrite is skewed toward the higher concentrations (Fig. 5).

Table 3 LA-ICP-MS analyses of arsenopyrite from different orebodies in the Sunrise Dam gold deposit (data in parts per million)

The Te concentrations vary from 6 ppm to 1.7 wt.% (Fig. 5); contrary to pyrite, there is no correlation between Te and Au, but instead, two distinct populations emerge (Fig. 7a). It appears that arsenopyrite from samples collected near porphyry intrusions forms a distinct population, higher in Te relatively to samples from the other orebodies. Arsenopyrite can also have high concentrations of Sb (to 1.7 wt.%), Ni (to 1.0 wt.%), and Pb (to 1.5 wt.%). The highest concentrations of Cr (288–1,853 ppm) were measured in arsenopyrite from Cosmo (other arsenopyrites have ≤399 ppm), which also contains relatively high Co (307–473 ppm) and Ni (6,200–9,974 ppm; Table 3).

LA-ICP-MS spectra

Different types of time-resolved LA-ICP-MS profiles were obtained from pyrites (Fig. 4). Smooth and steady ablation profiles for Au, with a positive correlation between the Au and As signals, were obtained from most D_4a pyrites (Fig. 4b, c). This type of pattern is attributed to invisible gold, i.e., gold in solid solution or as inclusions <0.1 μm. We observe, however, that the correlation between the As and Au signals can change across the ablation interval (Fig. 4c). The Au signal in the richest auriferous pyrite (D_4a) is irregular, although it remains at relatively high level, whereas the As signal is smooth in comparison; parallel signals are obtained for Pb and Te in the same spectra (Fig. 4d). This type of pattern is interpreted as reflecting the presence of submicrometer gold inclusions, a fact confirmed by BSE images showing the presence of such native gold inclusions (<0.5 μm in size). Some of the D₃ pyrites have similar LA-ICP-MS profiles, although the Au concentrations are much lower (Fig. 4a).

Another type of spectra, characteristic of D_4b pyrite, but found in a few D_4a pyrite crystals as well, shows sharp variations in Au signal that are mimicked by the Te and Pb signals (Fig. 4e). This is interpreted as characteristic for the presence of buried inclusions of Au-bearing minerals, in the case of Fig. 4e nagyágite (Pb₅Au(Te,Sb)₄S_5–8).

In contrast to pyrite, the spectra obtained for arsenopyrite are relatively similar one to another. Figure 4f shows the LA-ICP-MS profile corresponding to the highest Au concentration measured in this study. The Au signal is more erratic than that of any other element, and there is no correlation between the Au signal and that of any other element.

Textures, mineral associations, and sulfide chemistry

We investigate the trace-element contents as a function of the textures and paragenesis of the sulfide minerals. A variety of distinct pyrite and arsenopyrite textures and mineral associations occur within the Sunrise Dam orebodies, which broadly correlate with the deformation stage. Pyrites in D_4a and D_4b structures often exhibit evidence of episodic growth, recrystallization, and remobilization.

Pyrite from D₁ stage

D₁ pyrite occurs as fine-grained (50 ~ 120 μm), anhedral to subhedral grains associated with minor amounts of tennantite, chalcopyrite, and molybdenite; the sulfides are observed both in the vein and host rock. Gold content in the pyrite varies from less than 0.3 to 6.56 ppm Au with an average value of 1.12 ppm Au (n = 6; Table 2).

Pyrite from D₃ stage

D₃ pyrite in general occurs as subhedral to anhedral grains, or as distinct pentagonal dodecahedral (pyritohedral) or cubic crystals, and commonly with sponge-like surfaces and numerous cavities. Pressure shadows often observed around coarser-grained pyrite are mainly composed of elongated and curved quartz crystals. Euhedral pyrites in the BIF unit often contain a linear array of magnetite inclusions that are parallel to the orientation of magnetite banding, indicating porphyroblastic growth. The following description focuses on pyrites from three locations in the SSZ orebody, illustrating the differences between vein pyrites and those resulting from BIF sulfidation: SU2 is a small D₃ quartz vein and SU11 and SU17 illustrate increasing degrees of BIF sulfidation (Table 1).

The coarser pyrite grains in SU2 contain locally abundant minute inclusions of gangue minerals (Fig. 9a). Although such grains appear otherwise homogeneous under the SEM, a qualitative LA-ICP-MS profile across a grain shows significant variations in the Au and As signals, with an As- and Au-poor “zone” followed by an As- and Au-rich portion (Fig. 9b). The Au concentrations also appear to be positively correlated with Bi concentrations. Three spot analyses (Appendix) confirm this internal inhomogeneity: Au concentrations vary from <0.15 (in the inclusion-poor areas) to 0.48 ppm (inclusions-rich area). There are also some marked differences in the time-resolved depth profiles in LA-ICP-MS analyses (Fig. 9c, d). Co and Ni in spot 3 are detected as relatively steady lines over the time-integrated ablation (Fig. 9c), but significant variations with depth are observed from the spots 2 and 4 (e.g., Fig. 9d) suggesting compositional zoning with depth within the pyrite crystal. A spiky pattern in the Cu signal clearly indicates the presence of buried chalcopyrite inclusions; a larger (50 μm) chalcopyrite inclusion is visible in Fig. 9a.

Fig. 9

Pyrite in D₃ vein. a Back-scattered electron image showing the location of a LA-ICP-MS line raster single and three spot analyses. Note the distribution of areas with plenty of inclusions and those with few inclusions. The arrow shows the direction of line raster. b LA-ICP-MS line raster across the grain shown in a. c, d Typical ablation signals for selected elements from spots 3 and 4 in the BSE image. Abbreviations: see Table 1

Sample SU11 is a near-pristine BIF containing only minor amounts of pyrite, occurring as subhedral to anhedral fine (<100 μm) grains scattered in magnetite–quartz–dolomite–ankerite bands. Ankerite or ferroan dolomite is observed as infill between grain boundaries of subhedral magnetite. In addition, scheelite and chalcopyrite are found in trace amounts. One spot analysis shows 0.38 ppm Au. In contrast, in the highly sulfidized BIF sample SU17, pyrite is the dominant ore mineral, having replaced most of the original magnetite. This pyrite is coarser-grained and more porous than that in less sulfidized BIF. Fine-grained magnetite is disseminated sparsely and euhedral arsenopyrite crystals are found along selvages of the ankerite veins. Larger pyrite crystals in SU17 display very characteristic microstructures, which are interpreted as overgrowth of D₃ pyrite during D_4a (see below; Fig. 10e).

Fig. 10

Back-scattered electron images of textures of pyrite from the D4 stage at the Sunrise Dam deposit. a D_4a stage pyrite crystal enclosed by galena, tennantite, and chalcopyrite from WFZ, showing patchy As zoning. b Pyrite with three distinct growth zones marked as I, II, and III, from a quartz vein in the Cosmo orebody. c Pyrite with the same three distinct growth zones as in b, but from a highly sulfidized BIF. The white square indicates the location of a detail shown in d. d Detail of zone II of the pyrite crystal shown in c, showing crystallographically oriented arsenopyrite (white) and gangue minerals (dark spots). e D₃ (?) idiomorphic pyrite crystal showing a reaction-front rich in As; pyrite behind the reaction front displays porosity and many small gangue inclusions. f Small (40-μm) pyrite crystal recovered from the tailings dam, showing an As-rich core partially replaced by As-poor pyrite with numerous inclusions of arsenopyrite and increased porosity (filled with quartz). g, h D_4b stage pyrite, showing inclusions of native gold, chalcopyrite, sphalerite, and tennantite (g) and association with telluride (h). Abbreviations: see Table 1

Pyrite and arsenopyrite from D_4a stage

D_4a pyrite crystals were identified in all the studied orebodies (see Table 1). Pyrite from D_4a veins commonly occurs as subhedral to anhedral grains with numerous pores or inclusions of silicates, carbonates, oxides, sulfides, and minor telluride minerals.

In the WSZ, most pyrite consists of agglomerates of euhedral to subhedral crystals that are overgrown or completely rimmed by euhedral arsenopyrite. EMP results show that these pyrites have distinctively higher arsenic content (average 1.27 wt.%) than those from other stages or orebodies in the deposit, though the As distribution in these pyrites is somewhat patchy (evidence of episodic growth or recrystallization). Pyrite in the WSZ is disseminated within altered intermediate volcanic rocks and infrequently replaced by arsenopyrite near ankerite veins.

Some pyrite from the GQ orebody contains inclusions of chromite (abundant), melonite (NiTe₂), and gersdorffite. Other associated minor inclusions include altaite (PbTe), quartz, K-feldspar, rutile, titanite, sericite, monazite, dolomite, ankerite, and tourmaline. The Cr concentrations obtained from these pyrites can exceed 100 ppm (Fig. 8c); similar high (and higher) concentrations are found in D_4a pyrite from the nearby Cosmo orebody. These high Cr contents generally also are coupled with high Ni and Co contents (Fig. 8a). Overgrowth textures are very common in the D_4a pyrite from GQ and native gold is found as inclusions in pyrite or fills the fractures of pyrite either as separate grains or intergrown with gersdorffite. Au and Te concentrations in GQ pyrite are variable, from below detection limits to 22.4 ppm Au and 21.3 ppm Te. The overgrown pyrite contains lower Au concentration (Table 4 in Appendix) and there is generally a positive correlation between As and Au concentrations.

In the WFZ orebody, the arsenian pyrite (1.76 wt.% As) is sometimes enclosed successively by galena, tennantite, and chalcopyrite. BSE images reveal that the pyrite is composed of numerous irregular domains within an As-richer groundmass (Fig. 10a). Tetradymite (Bi₂Te₂S) was found as rare inclusions in pyrite from the WFZ orebody, either enclosed in pyrite, or sealing fractures in pyrite together with native gold. A single analysis of a pyrite with inclusions of tetradymite and native gold showed 0.85 ppm Au, 25.3 ppm Te, and 11.4 ppm Bi.

Many D_4a pyrites display a characteristic zoning consisting of three distinct zones (Fig. 10b, c): (I) a porous core of lowest As content surrounded by a narrow As-richer zone, (II) a porous inclusion-rich (mainly arsenopyrite with minor galena, quartz, feldspar, Mg–Fe carbonates; e.g., Fig. 10d) zone displaying corona-like texture, and (III) an outermost overgrowth zone consisting of bands or layers with different As contents. A change in the crystal morphology during the growth of these crystals is also noticeable; the cores through to zone II show cube or octahedral forms, subsequently changing to pyritohedron whose orientations are at a slight angle to the outline of the cores, indicating deformation during crystal growth. Coarse-grained pyrites often have corroded margins indicating dissolution. Such zoning is common among coarse crystals in sulfidized BIF from WSZ (Fig. 10c) and in As-rich quartz veins from the Cosmo orebody (Fig. 10b). There are wide variations in Au contents within the three different zones. The highest Au concentrations of up to 148.7 ppm were found in zone II (spot SU17-2), whereas the core (zone I) in the same grain has the lowest Au content (0.38 ppm, spot SU17-1). The LA-ICP-MS Au-profile indicates the presence of native gold inclusions near the surface, and this interpretation is supported by the occurrence of tiny native gold grains (<400 nm) in zone II of multiply overgrown pyrites from the Astro orebody. Similar evidence exists for the presence of micro-inclusions of a Pb–Sb sulfide or sulfosalt in zone II of these pyrites. The core (zone I) regions of these zoned pyrites are quite similar to those from less sulfidized BIF and may represent D₃ pyrite.

Another style of complex zoning in arsenian pyrite is characterized by textures suggestive of recrystallization along a reaction front (Fig. 10e, f). The reaction front in the crystal shown in Fig. 10e is underlined by As-rich composition; the pyrite behind the front has higher porosity than the “pristine” pyrite. The crystal shown in Fig. 10f has been retrieved from the tailing dams, but was left untouched by the processing. It has an As-rich core with weak oscillatory zoning and an overgrowth of As-poor pyrite. A large part of the core has been replaced by As-poor pyrite, with a composition similar to the overgrowth. This replaced pyrite is characterized by many inclusions of arsenopyrite as well as pores and shows a sharp and convolute contact with the core. The arrow on Fig. 10f points to an As-rich “ghost” that still underlines the original limit between the original crystal and an overgrowth with the same composition as the replacement.

Arsenopyrite occurs generally as idiomorphic, lath-, or rhomb-shaped crystals ranging in size between 50 and 200 μm in diameter, but occasionally reaching several mm in the Dolly Volcanic. Single crystals are dispersed in the host rocks or are concentrated in aggregates of numerous intergrown individuals occurring in thin interlayers and selvages of the ankerite veins and wall rock fragments in the veins. The arsenopyrite crystals generally show preferred orientation parallel to the schistosity of host lithologies. In some cases, rosettes of euhedral arsenopyrite crystals rim pyrite (e.g., Fig. 10b) and fine-grained euhedral arsenopyrite occurs as inclusions within pyrite crystals. When observed by BSE imaging, most arsenopyrite crystals display compositional zoning related to the As contents. Electron microprobe analyses reveal that arsenopyrite is systematically As deficient, ranging from 25.7 to 30.4 at.% As (33 at.% is the stoichiometric value). Such As deficiency is common in Archaean orogenic Au deposits, and these values place Sunrise Dam among the “Au-bearing arsenical type” deposits in the classification of Vaughan and Kyin (2004). Arsenopyrite from the Astro orebody has the lowest average As (26.22 at.%), whereas the highest average As content (30.23 at.%) is found in coarse-grained arsenopyrite in the Dolly Volcanic orebody. The As zoning in arsenopyrite seems to be more complex in the ore veins compared to that in the wall rock. Rims of arsenopyrite crystals are usually richer in As than the cores (Fig. 11a), most likely from a change in fluid composition or temperature during crystallization. The crystal in Fig. 11b displays a different type of As zoning: The complex patchy zoning is related to crystallographic control of composition, most probably through polysynthetic twinning.

Fig. 11

SEM back-scattered electron images of arsenopyrite textures from the WSZ at the Sunrise Dam deposit; brighter areas contain more arsenic. a Zoning in fine-grained arsenopyrite. Sample WSZ333. b Twinned arsenopyrite crystal with inhomogeneous As content. Sample SU7. Abbreviations as in Table 1

In most cases, arsenopyrite in veins or vein salvages shows Au concentrations that are one to two orders of magnitude higher than in wall rock arsenopyrite (Table 3). A euhedral arsenopyrite crystal from highly sulfidized BIF also contains 23.4 ppm Au and elevated Te concentrations up to 435 ppm (spot SU17-8). The highest Au value was measured in a sample from the Astro orebody (Table 3). Although this value is anomalously high, the ablated Te signals exhibit smooth profiles and a good correlation with As and Fe signals. Thus, Te appears to be present in solid solution in the crystal structure of arsenopyrite.

D_4b stage

D_4b vein pyrites occur mostly as subhedral to euhedral grains (20 μm to ~2 mm in size; type 1_eh) and aggregates (type 1_ag) possessing a pyritohedral or cubic form. Locally, smaller (0.1–10 μm) euhedral to subhedral disseminated grains are also present (type 2). In some parts of D_4b quartz veins that show swelling, subhedral to anhedral pyrite (type 1_sw) is accompanied by other sulfides (sphalerite, galena), sulfosalts (tetrahedrite, bournonite), and nagyágite (Fig. 10h). Rare skeletal pyrite grains (type 3) are found in quartz veins that also contain pyrite grains (up to 1 mm in length) with inclusions of native gold. In contrast to other pyrite from Sunrise Dam, many D_4b pyrite crystals contain micrometer-sized native gold inclusions (e.g., Fig. 10g). Other minerals found as inclusions in D_4b pyrite include hessite, chalcopyrite, tennantite, galena, cassiterite, quartz, K-feldspar, fluorapatite, monazite, xenotime, ankerite, and dolomite. Among these, the sulfides and sulfosalts are generally aligned parallel to the growth planes, indicating pre- or syn-tectonic growth of D_4b pyrite, under relatively stable conditions allowing the nucleation of ore minerals on the pyrite surface. In contrast, silicate and carbonate inclusions generally occur as mineral aggregates with semi-oriented texture at the margin, providing evidence for fluid pressure fluctuations during pyrite growth (Ciobanu et al. 2004).

Concentrations of 0.52 to 18.2 ppm Au, 2.22 to 53.3 ppm Te, and 922 to 5,053 ppm As were obtained from type 1_sw pyrite grains (spot SU4-1, SU4-2, and SU10-2; Table 4 in Appendix). Laser ablation profiles show that gold has positive correlations with Te, Pb, Sb, and Ag rather than with As (e.g., Fig. 4e), indicating that Au is present mainly in the form of nagyágite inclusions.

Silver content of native gold

Coarse native gold is found within the different mineralizing stages and alteration zones throughout the deposit, but D_4b veins are characterized by a large number of particulate gold grains, mainly occurring as inclusions within pyrite or along microfractures and grain boundaries in pyrite (e.g., Fig. 10g). They also occur in tetrahedrite–tennantite, galena, sphalerite, nagyágite, altaite, and Au–(Ag)–As–telluride and are intergrown with hessite and tetradymite.

EMP analyses revealed that gold is alloyed with silver at Sunrise Dam, with Ag concentrations ranging from 2.78 to 42.80 wt.%. The purest gold (97.2 wt.% Au; n = 6) occurs as coarse free gold in ankerite veins from the D_4a stage. Native gold in intermediate volcanic rocks from WSZ (D₃ or D_4a stage) is also relatively pure (average 96.4 wt.%, n = 6), as is native gold from the decomposition of nagyágite (average 94.6 wt.%, n = 7). Native gold in D_4b structures shows a wide variability: gold inclusions in D_4b pyrite have an average Au content of 91.8 wt.% (n = 12), free gold in D_4b quartz vein 84.7 wt.% (n = 5), and the lowest purity (56.9 wt.% Au) was found in electrum filling fractures in D_4b pyrite.

Discussion

Gold incorporation in sulfides

In natural arsenian pyrite, the highest reported Au concentration exceeds 1.1 wt.% (Pals and Spry 2003; Pals et al. 2003), but a maximum Au content of only 0.64 wt.% was obtained from synthetic arsenian pyrite (Fleet and Mumin 1997). In contrast, gold values up to 1.52 wt.% have been reported in natural arsenopyrite (Johan et al. 1989), and Fleet and Mumin (1997) prepared synthetic arsenopyrite containing up to 3.0 wt.% Au. Vaughan and Kyin (2004) used secondary ion mass spectrometry measurements on arsenian pyrite and arsenopyrite to characterize refractory Au ores from nine mesothermal Au deposits in the Yilgarn Craton. They reported maximum Au contents of 327 ppm (average 43 ppm) for arsenian pyrite from Wiluna (Norseman-Wiluna Belt, ~345 km NW of Sunrise Dam) and 1,069 ppm (average 276 ppm) for arsenopyrite from Lancefield (Laverton Greenstone Belt, ~65 km N of Sunrise Dam). Highly refractory arsenical ores are predominant in both of these deposits.

The LA-ICP-MS data in this study indicate that arsenian pyrite and arsenopyrite in the Sunrise Dam deposit incorporate significant amounts of refractory Au. The average Au content in arsenian pyrite is 1.61 ppm (n = 224) with a maximum value of 3,067 ppm from the WSZ orebody. In arsenopyrite, the average Au content is 562 ppm (n = 35), with a maximum value of 5,767 ppm. The highest Au values, both average and maximum, are obtained from D_4a stage pyrite and arsenopyrite, which is consistent with D₄ being the major mineralizing event in the deposit. The average Au concentration (calculated using raw parts per million values) in arsenian pyrite is almost 20 times higher than that from Lancefield (21.0 vs. 1.2 ppm Au), but is lower than that from the Wiluna deposit (43 ppm Au). The Au concentrations in the Sunrise Dam arsenopyrite are the highest yet reported in the Yilgarn Craton and are even higher than those reported from the arsenopyrite-rich Ashanti mesothermal deposit in the Ashanti Belt of Ghana (1 to 2,500 ppm, average 190 ppm Au; Oberthür et al. 1997). Our data indicate a preferred enrichment of Au in arsenopyrite relative to coexisting pyrite (e.g., Fig. 5); such preferential enrichment has not only been documented in the Yilgarn (e.g., Vaughan and Kyin 2004; Morey et al. 2008) but also from many gold deposits worldwide (e.g., Mumin et al. 1994; Fleet and Mumin 1997).

Element correlations and the nature of LA-ICP-MS traces indicate that Au occurs in different forms in sulfides from Sunrise Dam, including “invisible gold” (either solid solution in the crystal lattice or nanoparticles <100 nm in size), particulate native gold, and Au compounds, notably tellurides. The complexity of the paragenesis and speciation of Au found at Sunrise Dam is the result of a number of different processes, active at different scales. The most likely contributors are:

• 1. Different mechanisms for mineral precipitation. Brown et al. (2003) suggests that two main processes resulted in Au deposition at Sunrise Dam: desulfidation of the hydrothermal fluid caused by fluid–rock interaction, resulting in the destabilization of the Au(I)–bisulfide complexes in solution (D₃–D_4a), and phase separation during D_4b. At Sunrise Dam, two types of lithologies are particularly rich in Fe(II) and hence highly reactive in the presence of a bisulfide-bearing hydrothermal fluid: BIF and (ultra-) mafic rocks. Both lithologies appear to have reacted with D₃/D_4a hydrothermal fluids, causing significant Au mineralization. Intermediate volcanic and volcanoclastic rocks also can participate in scavenging sulfur from the hydrothermal fluid.

• 2. Different mechanisms for Au incorporation in sulfides. The role of As-rich sulfides, foremost arsenian pyrite and arsenopyrite, in scavenging Au has been long recognized, but the scavenging mechanism remains controversial. Several mechanisms have been proposed that may all be active under specific conditions. These mechanisms range from surface electrochemical reactions (e.g., Maddox et al. 1998), to chemisorption (e.g., Fleet and Mumin 1997), and to highly reducing conditions localized at the surface of arsenic-rich minerals during dissolution (Pokrovski et al. 2002).

• 3. Local redistribution of Au—and possibly upgrading of Au grades—during remobilization of preexisting Au mineralization. This aspect may have been underestimated in many polyphase Au deposits, and we will demonstrate how dissolution–reprecipitation reactions may account for a significant upgrading of Au grades at Sunrise Dam.

In the next section, we attempt to identify some of the factors that result in the complex Au speciation observed at Sunrise Dam.

Sulfidation reactions: BIF vs. (ultra-) mafic vs. intermediate volcanic and volcanoclastic rocks

Pyrite from partially sulfidized BIF and core zones of pyrite (zone I) in highly sulfidized BIF contain low Au concentrations (<0.4 ppm) and relatively low concentrations of trace elements (Table 4 in Appendix). These pyrite crystals likely grew during D₃. The low Au concentrations may reflect sulfidation from a Au-poor fluid, or the inefficiency of the sulfidation reaction in affecting Au solubility under the conditions of formation of this pyrite. The higher Au concentrations in zones II and III could indicate that Au solubility was controlled by the activity of bisulfide in the fluid, itself controlled by the extent of fluid–rock interaction (amount of pyrite formed); however, the texture suggests that zone II resulted from the recrystallization of a preexisting pyrite during D_4a (see below) and hence that the higher Au concentrations are more likely related to a different Au-scavenging process or to different physical and chemical conditions.

The GQ orebody at the 1960 ore drive is mainly composed of arsenian pyrite with minor amounts of Cr–Ni–Ti minerals and tellurides. The As and Au contents are higher in the core of these pyrites than in the overgrowth, implying that deposition of pyrite was initiated from Au–As–(Te)-rich fluids followed by progressive overgrowth of pyrite from more dilute fluid, possibly during the waning stage of D_4a mineralization. The nature of the host rock protolith is hidden by a very strong carbonate–sericite alteration, but the Cr contents of the sericite (“fuchsite”) and the chemistry of the GQ pyrite provide very strong evidence that the altered host rock had some (ultra-) mafic affiliation. It displays elevated contents of Cr (up to 92.9 ppm), Co (up to 537 ppm), and Ni (up to 7,474 ppm). Moreover, GQ pyrite preserves inclusions of relict chromite, which most likely represents the source of Cr found in the secondary “fuchsite”. The presence of Ni-bearing minerals (e.g., gersdorffite), which is restricted to Cr-rich assemblages, also is consistent with the (ultra-) mafic nature of the protolith.

A similar situation exists in the Cosmo orebody at the 1940 ore drive, where both pyrite (D₃ and D_4a) and arsenopyrite (D_4a) are enriched in Cr, Co, and Ni (Tables 2 and 3). Coarser pyrite grains in D₃ stage contain inclusions of pyrrhotite, indicating a reducing character and/or higher temperature environment (Mernagh et al. 2004). In contrast to the GQ orebody, Cr- or Ni-bearing minerals were not observed at Cosmo; however, elevated contents of Cr, Ni, and Co in sulfide minerals may reflect an affiliation with (ultra-) mafic host rocks.

In a location where the WFZ orebody is hosted in intermediate volcanic rock, high Au (45.2 ppm) and As (1.76 wt.%) concentrations occur in pyrite successively encapsulated by other sulfide and sulfosalt minerals (Fig. 10b). Textural evidence and high As contents suggest that the pyrite was precipitated from D_4a stage ore fluids enriched in As and Au and that it retained its original chemical composition during later mineralizing events. Judging from the smooth and flat profile of the laser ablation signal (Fig. 9b), Au is likely to occur in a chemically bonded form. Pyrite containing inclusions of tetradymite and native gold shows elevated contents of Bi and strong correlation with the Au signal. This association is interpreted as a geochemical affinity between Au and Bi and their simultaneous incorporation into pyrite structure from the D_4a mineralizing fluid that formed tetradymite and native gold. Moreover, differences in chemistry and inclusion mineralogy in D_4a pyrites from GQ and WFZ are also interpreted to reflect the influence of local fluid–rock interaction; it is possible that the Bi was derived locally from the intermediate volcanic rocks.

The above examples confirm that fluid–rock interaction was a major precipitation mechanism for sulfides in D₃/D_4a stages. In many cases, the nature of the host rock affected the chemistry of pyrite. It is also noticeable that the early BIF sulfidation reactions did not produce high Au grades at Sunrise Dam, most probably reflecting strong undersaturation of the mineralizing fluids with respect to gold during the sulfidation.

Association of Au–As and Au–Te–(Sb)

Refractory Au is commonly incorporated within the structures of arsenian pyrite and arsenopyrite, and it has a strong affinity to As content of arsenian pyrite and inversely correlates with the Fe content of arsenopyrite (e.g., Wu and Deldove 1989; Cook and Chryssoulis 1990; Fleet et al. 1993; Yang et al. 1998; Pals et al. 2003). LA-ICP-MS analyses of arsenian pyrite from the Sunrise Dam deposit show a positive correlation between Au and As (Fig. 6), which is common for invisible gold-bearing arsenian pyrites elsewhere (e.g., Arehart et al. 1993; Cook and Chryssoulis 1990; Ashley et al. 2000; Pals et al. 2003). Furthermore, Au contents in pyrite also have a positive correlation with Te (Fig. 7b). In contrast, a plot of Au and Te values from arsenopyrite (Fig. 7a) distinguishes two arsenopyrite populations. Arsenopyrite from the Astro, Cosmo, and Dolly Porphyry orebodies are relatively Te rich and show a weak positive correlation between the Au and Te concentrations. Arsenopyrite from the other orebodies has lower Te concentrations and no, or a slightly negative, correlation between Te and Au. The Astro, Cosmo, and Dolly Porphyry orebodies are within and adjacent to porphyry intrusions, and the As–Te–Au signature may reflect this lithological association.

Gold concentrations in arsenopyrite are inversely correlated with grain size: Arsenopyrite crystals in the Dolly Volcanic are coarse-grained (up to several millimeters in length) with lower average Au values (15.7 ppm Au, N = 5), whereas finer grained (<100 μm) arsenopyrite crystals in the WSZ have much higher average Au concentrations of 1,611 ppm (N = 10); this relationship is consistent with the findings of Vaughan and Kyin (2004) in that finer-grained arsenopyrite from gold deposits in the Yilgarn Craton generally contains higher concentrations of refractory gold. It is noteworthy that pyrite and arsenopyrite in porphyry host rock, even within a single orebody such as Dolly, have higher average Au concentrations (4.79 ppm in pyrite and 1,821 ppm in arsenopyrite) than those from volcanic host rock (0.78 ppm in pyrite and 15.7 ppm in arsenopyrite). Because volcanic rocks are less competent than porphyry (Brown et al. 2002), they are inferred to have hosted coarser-grained (but comparatively Au-depleted) sulfide minerals by providing a suitable porosity compared with adjacent rhyodacite porphyry.

According to Reich et al. (2005), the solubility of gold in arsenian pyrite from Carlin-type and epithermal Au deposits increases with the amount of As. They argued that high Au values above the solubility limit correspond to particulate native gold (Au⁰), whereas those below the limit represent structurally bound Au occurring as solid solution (Au⁺). At Sunrise Dam, five data points plot above or on the Au solubility line (Fig. 6), three of them are from WSZ (D_4a) and two from Mako (D_4b). The presence of metallic gold in the sampled volume is suggested by the spiky pattern of ablated signals, and, in some cases, SEM imaging revealed the presence of submicrometer-sized native gold particles. All the other analyses fall below the Au solubility limit. However, metallic Au inclusions still exist below the Au solubility limit in some pyrites grains from sulfidized BIF, D_4b pyrite, and D₃ (Fig. 4a). This discrepancy may result from different formation mechanisms and/or environment between Carlin-type and Archean orogenic Au deposits. The former type is lower temperature (150 ~ 250°C) and contains Au that is incorporated during the growth of As-rich bands in arsenian pyrite (Fleet et al. 1993). The presence of particulate gold below the Au solubility limit in the Sunrise Dam pyrite is interpreted as an evidence of the migration (remobilization) of submicroscopic gold from the arsenian pyrite lattice into inclusions during dissolution–reprecipitation (D_4a) and recrystallization (post-D_4b) process (see below).

As reported by Genkin et al. (1998) and Vaughan and Kyin (2004), Au-bearing arsenopyrite in mesothermal deposits is usually As deficient (27–32 at.%), with the most As-deficient varieties usually carrying the highest Au contents. Vaughan and Kyin (2005) found that arsenopyrite in highly refractory ores in the Yilgarn Craton contains high average concentrations (based on raw values) of invisible Au ranging from 60 to 270 ppm without any native gold inclusions (>100 nm). For example, arsenopyrite from the Lancefield deposit is As deficient (28.2–32.0 at.%), fine grained (average 20 μm), and occurs within a shear zone that crosscuts earlier formed pyrite and pyrrhotite (Vaughan and Kyin 2004). For Sunrise Dam, the occurrence of arsenopyrite is similar to Lancefield, in that arsenopyrite replaces pyrite near the vein salvages, but the arsenic content is much lower (25.7–30.4 at.%) and the concentration of invisible Au is much higher, ranging from 0.8 to 5,767 ppm with an average of 562 ppm (log based) compared to 0.1 to 1,069 ppm with an average 276 ppm at Lancefield.

D_4b stage mineralization is characterized by the relative abundance of Sb–Te–Pb–Zn minerals and by a varied assemblage of telluride minerals. D₄ stage mineralization began with steeper veins (D_4a) and concluded in the formation of gently dipping veins (D_4b) by pressure failure or change of stress field. Based on such an assumption, it is suggested that nagyágite was formed from residual remobilizates, which were enriched in base metals, Sb, and Au, whereas the later replacement of these minerals resulted from destabilization by fluids generated during a post-D_4b event (Sung et al. 2007). The pyrite associated with nagyágite shows Au concentrations as high as 18.2 ppm and has a positive correlation with Te, Pb, Sb, and Ag (Figs. 6 and 8); this is related to the presence of small inclusions of nagyágite and probably reflects the geochemical signature of the D_4b stage. This suggests coeval precipitation of pyrite and nagyágite.

Hence, invisible gold in Sunrise Dam pyrite and arsenopyrite is correlated to the As and/or Te contents in these minerals. Whereas arsenopyrite appears to contain mainly “invisible” gold, pyrite from different stages and different localities within a single stage show either invisible gold, native gold inclusions, or telluride inclusions. On the one side, this not only reflects the greater susceptibility of arsenian pyrite to recrystallization and gold remobilization but also a wider range of conditions for pyrite crystallization. In particular, the relatively As-poor and Te-rich D_4b stage is characterized by coprecipitation of gold with pyrite as Au–telluride inclusions, as well as remobilization of invisible gold to native gold inclusions.

Remobilization of invisible Au

Sung et al. (2007) described the breakdown of nagyágite to native gold, altaite, galena, aurostibite (AuSb₂), and semseyite (Pb₉Sb₈S₂₁), as a result of tectonic activity that took place late to post-D_4b. This illustrates how postdepositional processes can affect the speciation of gold. It is likely that similar processes affected D_4b stage pyrite.

In contrast to those formed during the D₁, D₃, and D_4a stages, the coarser pyrite grains in the D_4b stage generally contain numerous native gold inclusions, and Au concentrations are not correlated with the As contents, e.g., As-rich rims do not contain any detectable Au (spot SU5-3 in Appendix), whereas the relatively As-poor core tends to have higher Au contents (10.1 ppm, spot SU5-2). In the latter case, the Au content appears to coincide with elevated Te, Ag, and Pb values (Fig. 4e), probably due to small inclusions of telluride minerals incorporated during the early stage of pyrite crystallization. These features can be explained by assuming that recrystallization during a late to post-D_4b event has caused an increase in grain size, resulting in the liberation of structurally bound Au from the pyrite lattice (Larocque et al. 1995; McClenaghan et al. 2004; Vaughan and Kyin 2004). Remobilized Au may (1) be precipitated as native gold inclusions, which are oriented parallel to the crystal face; (2) infill fractures as native gold, together with other ore minerals including chalcopyrite, galena, tennantite, and hessite (e.g., Fig. 10g); or (3) be redeposited as invisible Au, as illustrated by Au concentrations up to 15 ppm in small pyrite grains, which crystallized later than adjacent coarser-grained pyrite (spots SU15-3 and SU15-4; Appendix). The majority of native gold grains (ca. 60%) in the D_4b stage are found within, or in contact with, sulfides, sulfosalts, and tellurides rather than in quartz vein (12%). This implies that Au remobilized from D_4b pyrite reprecipitated more or less in situ, within the same sulfur- and telluride-rich environment, probably restricted to centimeter-scale migrations. The appearance of visible gold as a result of arsenian pyrite recrystallization is a common feature in Au deposits (e.g., Coleman 1957; Mumin et al. 1994; Genkin et al. 1998) and has been recently described in six deposits from the Eastern Goldfields of the Yilgarn Craton (Morey et al. 2008).

Preferential enrichment of Au in arsenopyrite seems to reflect primary coprecipitation and not the result of postdeposition processes. The lack of remobilization of the arsenopyrite-hosted Au is supported by the fact that there is no native gold occurring either as fracture fill or overgrowths on the auriferous arsenopyrite. Particulate gold grains are rare in arsenopyrite (e.g., Mumin et al. 1994), and according to the BSE images, the rim of arsenopyrite still tends to be As richer (Fig. 11a). Therefore, in sharp contrast with pyrite, post-D_4a Au remobilization and redistribution does not appear to have affected the Au contents and Au speciation in arsenopyrite at the Sunrise Dam deposit.

Arsenopyrite is more refractory than pyrite and commonly retains higher levels of invisible Au (Vaughan and Kyin 2004). However, there is evidence that arsenopyrite easily re-equilibrates during peak metamorphism, resulting in the liberation of Au (Mumin et al. 1994) and As from the outer rim (e.g., Larocque et al. 1995; Lentz 2002). Therefore, preservation of compositional zoning and nonuniform compositions within individual arsenopyrite grains in the Sunrise Dam (Fig. 11a, b) are most consistent with syn- to postpeak metamorphism mineralization (Neumayr et al. 1993; Tarnocai et al. 1997).

The role of coupled dissolution–reprecipitation reactions in upgrading Au grades

As demonstrated above, recrystallization reactions are important in controlling Au speciation. Such reactions can proceed in two different ways: solid-state process or coupled dissolution–reprecipitation reaction (CDRR; e.g. Putnis 2002; Xia et al. 2009). While solid-state reactions are controlled by diffusion (P, T, stress state), CDRR proceeds via the complete dissolution of the parent mineral, coupled at the nanometer scale with precipitation of the product (e.g., Putnis 2002; Xia et al. 2009). Hence, reaction kinetics is controlled by the relative solubilities of the parent and product minerals, which depends on P, T, as well as on solution chemistry (e.g., pH, redox, ligand concentrations). For example, Tenailleau et al. (2006) and Xia et al. (2008, 2009) demonstrate that the transformation of pentlandite ((Fe,Ni)₉S₈) into violarite (FeNi₂S₄), which cannot be achieved at low temperature via a solid-state process, happens in a few days under mild (120°C) hydrothermal conditions following CDRR. The resulting product is characterized by high porosity, even when the product has higher molar volume than the starting material (Putnis 2002). CDRRs can proceed topotactically, preserving both the crystallographic orientation and the external shape of the replaced grain (e.g., Xia et al. 2009).

Here, we argue that the Au-rich zone II pyrite developed via the CDRR mechanism during D_4a and that this process may be responsible for the high Au contents in this pyrite. A good illustration that CDRRs were important in the development of the pyrite textures is provided by the crystal illustrated in Fig. 10f, which shows unambiguously that nearly half the grain was replaced during a subsequent mineralization event; this replacement was also accompanied by an overgrowth on the original crystal. The replaced pyrite is characterized by many inclusions of arsenopyrite as well as pores. These features indicate that the replacement occurred via a CDRR, as a solid-state reaction cannot account for the sharp contact in particular. The crystal from WSZ illustrated in Fig. 10e was affected in a similar fashion by a reaction front moving through the crystal.

The features present in the crystals from Fig. 10e, f correspond to those seen in the zoned crystals in D₃ structures and in BIF (Fig. 10b–d). The Au-rich zone II is characterized by high porosity and by the presence of a myriad of micro-inclusions, which are sometimes preferentially oriented along apparent crystallographic axes of the host pyrite crystal (Fig. 10d). In conjunction with irregular boundaries between zone II and other zones, these features indicate that this zone most probably recrystallized. During this recrystallization, the original arsenian pyrite was transformed into Au-bearing arsenian pyrite with numerous arsenopyrite inclusions; Au and As were probably added to the mineral. This indicates the presence of a fluid enriched in As and Au compared to the “original” D₃ fluid. This fluid most likely corresponds to the D_4a fluid. The features of this recrystallization, i.e., its ability to replace specifically a particular zone in a pyrite crystal (probably characterized by a specific chemistry or microstructure), and the resulting porosity are strongly suggestive of a coupled dissolution–reprecipitation process (Putnis 2002), in which replacement occurs along a reaction front by full dissolution of the mineral followed by tightly coupled reprecipitation.

At Sunrise Dam, such coupled dissolution–reprecipitation reactions may have been important in upgrading the ore grades during D₄. This is because solution chemistry at the reaction front may be different from the conditions in the bulk of the hydrothermal fluid. As first noted by Pokrovski et al. (2002), the dissolution of As-rich pyrite (or arsenopyrite) results in the release of hydrogen:${\text{FeAsS}} + 2{\text{Cl}}^{ - } + {\text{H}}^{ + } + 3{\text{H}}_2 {\text{O}} = {\text{As}}\left( {\text{OH}} \right)_3 \left( {\text{aq}} \right) + {\text{HS}}^{ - } + {\text{FeCl}}_2 \left( {\text{aq}} \right) + 1.5{\text{H}}_2 \left( {\text{aq}} \right)$This leads to reducing conditions at the reaction front, and in most cases, this is expected to decrease Au solubility (e.g., Mikucki 1998) and to favor the coprecipitation of Au with the newly formed pyrite or arsenopyrite.

Diversity of processes and protracted mineralization history

This discussion has emphasized the diversity of processes that are responsible for Au mineralization at Sunrise Dam. Thébaud et al. (2008) demonstrated the role of multiple flow via a single plumbing system in “setting the stage” for major mineralizing events. At Sunrise Dam, the D₁ and D₂ events probably contributed little to the total Au endowment, but put in place a plumbing system that was used by all subsequent mineralization events. Alteration and low-grade mineralization during D₁ and D₂ may also have enabled more efficient transport of Au during subsequent hydrothermal events. A large portion of the gold in many orebodies was deposited during D₄. The mineralogical differences between the D_4a and D_4b stages may reflect different precipitation mechanisms from the same mineralizing fluid: fluid–rock interaction and interaction with preexisting (mainly D₃) pyrite accounting for most of the mineralization during D_4a, whereas D_4b veins probably formed from spent D_4a fluid (As depletion). Pressure variations during D_4b vein formations may have led to mineral precipitation (e.g., Liu et al. 2008), possibly facilitated by phase separation.

Such a complexity is indeed characteristic for many giant orogenic Au deposits. For example, a protracted history of Au flux leading to complex mineralization geometry and mineralogy may be the most striking feature of the giant Kalgoorlie deposit, by far the largest deposit in the Yilgarn Craton (Bateman and Hagemann 2002).

The role of sulfide recrystallization in modifying the speciation of gold (e.g., formation of visible gold from invisible gold) has long been recognized (e.g., Morey et al. 2008 and references in there). However, at Sunrise Dam, we propose a new process of Au deposition/upgrading via recrystallization following the dissolution–reprecipitation process. Given that episodic fluid flow can develop even within single mineralizing events (e.g., Sibson et al. 1988; Brugger et al. 2000, 2008), dissolution–reprecipitation reactions have the potential to play an important role in many gold deposits, and their role in the formation of giant Au deposits needs to be further investigated.

Conclusions

• 1. Arsenopyrite and arsenian pyrite in the Sunrise Dam gold deposit contain significant amounts of “invisible” Au; the average Au concentration is 1.61 ppm in pyrite (n = 220) with maximum value of 3,067 ppm and 562 ppm in arsenopyrite (n = 35) with maximum value of 5,767 ppm, which are the highest concentrations reported so far in Yilgarn Craton.

• 2. The presence of both native and invisible Au and high concentrations of Au in As-deficient arsenopyrite indicate that mineralization at Sunrise Dam is consistent with those of lower temperature deposits in the Yilgarn Craton and with a formation syn- to postgreenschist facies peak metamorphism (Mair et al. 2000).

• 3. Our dataset shows few systematic differences among pyrite crystals from different stages, apart from D_4a showing the highest Au and Te contents (Figs. 6 and 8). The highest Au values (both average and maximum) are obtained from D_4a stage pyrite and arsenopyrite, which is consistent with D_4a being the major mineralizing event in the deposit. The elevated Cr–Ni–Co signature in some pyrite assemblages reflects their crystallization in rock of (ultra-) mafic affinity and pronounced fluid–rock interaction. Similarly, some locally high Bi contents may reflect mobilization of Bi from intermediate volcanics during fluid–rock interaction.

• 4. The sulfidation of BIF originally produced Au-poor pyrite (<0.4 ppm). Upgrading of Au contained in BIF pyrite was partly the result of a coupled dissolution–reprecipitation reactions during the D_4a stage. A similar mechanism also upgraded the Au contents in vein-hosted D₃ pyrites. Textures indicative of coupled dissolution–reprecipitation reactions are common in the deposit and may account for a significant increase in Au grades during D_4a. The importance of coupled dissolution–reprecipitation as a Au-enrichment factor in polystage hydrothermal ore deposits has not been previously recognized, but such reactions have a large potential for efficiently scavenging Au from a hydrothermal fluid, by creating locally highly reducing conditions.

• 5. Recrystallization of pyrite also resulting in altering Au speciation in the deposit. D_4b-stage pyrite contains a significant amount of Au as inclusions of Au–tellurides (e.g., nagyágite). This pyrite also contains numerous inclusions of native gold, which resulted from the expulsion of lattice-bound Au, its micrometer-scale remobilization and redeposition during a late post-D_4b event. Further evidence for postcrystallization remobilization of Au in pyrite is provided by the presence of metallic Au below the gold solubility limit of Reich et al. (2005) in D₃ and D_4b arsenian pyrite, resulting from remobilization during the D_4a and D_4b stages.

• 6. The textures and chemistry of pyrite and arsenopyrite provide insights about mineralization processes. The complex speciation of Au at Sunrise Dam results from the large variety of mineralization processes that were responsible for Au precipitation (fluid–rock interaction with different rocks; possibly phase separation; dissolution–reprecipitation reactions) as well as postcrystallization remobilization. Such a complexity is indeed characteristic for many giant orogenic Au deposits (e.g., Bateman and Hagemann 2002; Mumin et al. 1994) and reflects multiple fluid flow through a single plumbing system.

Appendix

Table 4 LA-ICP-MS analyses of pyrite in the Sunrise Dam gold deposit