Discussion on Aspects of Mineralisation of the Bushveld Granites

Transcription

1 7 Discussion on Aspects of Mineralisation of the Bushveld Granites In the previous chapters various aspects of the mineralisation associated with the granites of the Bushveld were considered, in particular, the alteration of the granites near mineralisation and the petrographic and geochemical features of the alteration. Existing mineralised occurrences and deposits related to the Bushveld granites were reviewed, as were examples of IOCG mineralisation from Olympic Dam, Australia, and Salobo, Carajás mineral province, Brazil. New mineral occurrences were documented. Some of the considerations of the previous chapters will be revisited here in greater detail. Other discussions are also presented here which may have been outside the scope of previous chapters Comparison of Bushveld-type Mineralisation to Olympic Dam & Salobo Similarities exist between the deposits and occurrences of the Bushveld, including Vergenoeg, and those found in recognised IOCG provinces, such as Olympic Dam in South Australia and Salobo in the Carajás mineral province of Brazil (Figure 7.1). These similarities exist on cratonic, regional and locality scales. The characteristics of each of these deposits are compiled in Table 7.1 for comparative review and discussed below. All of the aforementioned deposits are Palaeo- to Mesoproterozoic in age and related to large-scale, A-type granite magmatism. Deposits appear to be located near craton margins, although the Bushveld examples are situated more towards the centre of the present Kaapvaal craton. Deposits are found in a variety of tectonic settings (c.f. Figure 1.2) with high heat flow and magmatism being related to mantle underplating (Hitzman, 2000)

2 Figure 7.1. Location of three major IOCG deposits at Olympic Dam, Carajàs and Vergenoeg. Proterozoic crust shown in dark grey, Phanerozoic rocks in medium grey and light grey (from Groves, 2004). Mineralisation may be both endogranitic and exogranitic, and tends to form subvertical breccia pipes. Mineralisation is generally concentrated in the zone of fluid mixing between hotter, more saline magmatic fluids and colder less saline, more oxidised meteoric fluids. Ores are dominated by iron oxide mineralisation, chiefly hematite and/or magnetite, and are commonly accompanied by copper-bearing sulphides, fluorite and a diverse range of other minerals including gold, uranium- and REE-bearing bearing minerals minerals. Alteration of host rocks and country rocks appears to be a significant attribute of mineralisation with characteristic styles of alteration observed in proximity to ore deposits, notably sodic alteration, K-metasomatism, sericitic and silicic alteration, carbonate alteration, chloritisation and hematisation. The patterns of alteration appear to be consistent between deposits

3 Table 7.1. Synopsis of characteristics of IOCG deposits considered in this study. Ruigtepoort Vergenoeg Olympic Dam Salobo Deposit Size Not known, minor 178 Mt 28.1 % CaF Mt 42 % Fe Mt 2.5 % Cu Mt 0.86 % Cu Age Craton Position & Tectonic Setting Deposit Host Rock Granite Relationship Palaeoproterozoic Palaeoproterozoic Mesoproterozoic Palaeoproterozoic Ma Ma Ma Ma 3 Central Margin Anorogenic Coarse-grained A-type granite Central Margin Anorogenic Rhyolites Craton Margin Anorogenic Coarse-grained A-type syenogranite Craton Margin Continental Rift Basin Metagreywackes amphibolites Endogranitic Exogranitic Endogranitic Exogranitic Orebody Morphology Model Ore Host Sub-horizontal mantos in breccia pipe Fluid-mixing in a medium- to highlevel environment Hematite-Quartz breccia Steeply-dipping to vertical breccia pipe Fluid-mixing in a high-level environment Hematite-Fluorite breccia Steeply-dipping to vertical breccia pipe High-level environment with exhalative activity and fluid mixing Hematite-Quartz breccia Ore Assemblage F (-Cu-Au) Fe-F (-Cu) Fe-Cu-U-Au-Ag- REE Vein Assemblage Fe minerals Cu minerals Hematite (magnetite) Pyrite, chalcopyrite, arsenopyrite, bornite, covellite Late fluoritesiderite-sulphide Hematite (magnetite) Pyrite, chalcopyrite, arsenopyrite, sphalerite, molybdenite, galena Late barite-fluoritesiderite-sulphide Hematite (magnetite) Chalcopyrite, bornite, chalcocite, Ag, Au U minerals Not determined Uraninite, coffinite, brannerite, pitchblende REE minerals Other Significant Phases Alteration Styles Britholite, bastnaesite Fluorite, ferroactinolite K-metasomatism, sericitic-silicic, chloritisation, hematisation Xenotime, monazite, fluorcerite 4 Fluorite, fayalite, apatite, ferroactinolite Hematisation, K-metasomatism, sideritic Bastnaesite, monazite, xenotime, florencite Fluorite with sulphide in barren core K-metasomatism, sericitic-silicic, chloritisation, hematisation Steeply-dipping shear-bounded duplex structure Amphibolite shear zone Fe-Cu Magnetite (hematite) Chalcopyrite, bornite, chalcocite, cobaltite Co, Au Uraninite Fluorite, fayalite, Na-metasomatism, K-metasomatism, sericitic-silicic, chloritisation Number of Fluids No data 2 fluids 2 fluids 2 fluids Fluid Temperatures and Compositions No data 1) 500 C δ 18 O = 7-8 2) C Salinity No data 1) > 67 % NaCl equiv. 2) 1-35 % NaCl equiv. 1) 400 C δ 18 O = 10 2) C δ 18 O = 10 1) 360 C 2) 195 C δ 34 S = % NaCl equiv. 1) 58 % NaCl equiv. 2) 1-29 % NaCl equiv. 1 Harmer & Armstrong (2000); 2 Creaser & Cooper (1993); 3 Machado et al. (1991); 4 Fourie (2000)

4 Geological Models for Bushveld-type Mineralisation Crocker et al. (1988) first suggested the deposits of Vergenoeg and Ruigtepoort to be genetically equivalent with differences between them attributed to differing levels of formation (Figure 7.2). The breccia pipe of Vergenoeg is hosted in Rooiberg Group rhyolites and had a volcanic/phreatomagmatic expression on the paleosurface consisting of ignimbrites, pyroclastic layers and agglomerates. Olympic Dam is also considered to have been a near-surface volcanic feature (Oreskes & Einaudi, 1990, 1992; Reeve et al., 1990) but was not a significant eruptive centre for coherent lavas or ignimbrites (Reynolds, 2000). The body at Ruigtepoort is modelled as being a sub-horizontal, manto-like feature without a volcanic expression. The abundance of roof-rock lithologies and xenoliths near the deposit suggest that Ruigtepoort was formed towards the top of Figure 7.2. Schematic model of level of formation for some Bushveld-type Fe-F deposits (from Crocker et al., 1988)

5 the Nebo granite sheet and perhaps not as deep as intimated by Crocker et al. (1988). It is not known whether the granite is domed, which may have influenced the distribution of the mineralisation and locally concentrated fluids upwards. The numerous mineralised occurrences extending from the Ruigtepoort mine northwards into the overlying stratigraphy may support this (c.f. Figure 2.10). Mineralisation at Salobo is also modelled to have been developed at some depth (c.f. Figure 1.12). According to Skirrow et al. (2000), a high crustal level is requisite for the formation of Cu-Au ores. Discussion on Mineral Assemblages Bailie & Robb (2004) described the nature of polymetallic mineralisation in the granites of the south-eastern portion of the Bushveld Complex and identified four broad types (Table 2.4), of which the Fe-F mineralisation, such as that of Vergenoeg and Ruigtepoort, formed one (Type IV). The Fe-F association is considered to be epigenetic in nature and occurs as a late-stage overprint of the other types. The small, mineralised occurrences of the study area were classified according to this scheme. This late-stage Fe-F overprint was apparent in each case; however, characteristics of Type II and Type III mineralisation were observed. The occurrences at Slipfontein, Elandslaagte and possibly Blokspruit exhibit mineral associations in line with Type II deposits, whereas Ruigtepoort and Doornfontein occurrences exhibit mineral associations in line with Type III deposits. According to the model of Bailie & Robb (2004), Type II deposits precede Type III deposits, implying that there were (at least) two episodes of mineralisation. However, this may be insufficient on its own to indicate the relative timing of mineralisation in each occurrence because the mineralisation is considered to have been episodic and punctuated in time (Robb et al., 2000). Overprinting of earlier episodes of alteration and mineralisation is expected and indeed observed. Further, considering the pronounced or subdued characteristics of one

6 mineralisation type compared with another may not fully appreciate the continuum of assemblages that might occur. Differences between the occurrences may simply reflect the degree of overprinting by subsequent mineralising episodes. The primary assemblage at Vergenoeg is considered to have been composed of fayalite, ilmenite, titanium magnetite, fluorite and REE minerals (Borrok et al., 1998; Fourie, 2000). A secondary assemblage of magnetite, titanium magnetite, siderite, quartz, apatite, sulphides and REE minerals overprinted this. A late-stage overprint of Fe-F completed the paragenesis, equivalent to the Type IV assemblage proposed by Bailie & Robb (2004). The assemblage observed at Blokspruit may most resemble the primary assemblage at Vergenoeg. The coarse ferroactinolite that occurs as the vein fill in the granite mega-breccia, may represent an alteration of a primary fayalite-dominated assemblage, such as that found in the lowest parts of the Vergenoeg deposit. The subsequent alteration of ferroactinolite to Fe-chlorite and quartz-hematite pseudomorphs is indicative of the subsequent stages in the evolving paragenetic sequence. The similarities in primary mineralogy between Blokspruit (and Ruigtepoort) and Vergenoeg lend strong support to idea that all of these bodies formed under similar geological and geochemical conditions. Discussion on Alteration Patterns and Sources The presence of fayalite in a granite system is indicative of the Fe-rich nature of the mineralising hydrothermal fluids, far more so than one would expect to be developed from a typical granite. A pertinent question is that of where and how such Fe-rich fluids were derived. This problem may be resolved by considering the alteration patterns and fluid compositions. The styles of alteration associated with the Bushveld-type mineralisation indicate a loosely defined zonation around mineralised veins and bodies from

7 chloritisation closest to the mineralisation (< 2 m), to sericitisation accompanied by silicic or epidote alteration beyond this (< 5 m), and K-metasomatism on a much broader scale (<100 m), with albitisation occurring on a more regional scale, although not easily identifiable in the field (Figure 7.3). This broad alteration pattern has been described around most IOCG deposits, sometimes with local variations. Hitzman et al. (1992) described the relationship between the styles of alteration and included the relationship as a function of depth and constructed the model reproduced in Figure 7.4. Sodic alteration is shown to occur at the greatest depth, with potassic alteration, then sericitic alteration becoming dominant towards the surface. Massive magnetite and magnetite stockworks are shown to occur at intermediate levels but largely corresponding to the distribution of potassic alteration. Massive hematite bodies and mantos are shown to correspond to the distribution of sericitic alteration. The implication is that the redox-controlled boundary between magnetite and hematite formation and the pressure-temperature-controlled boundary between potassic and sericitic alteration occur at similar levels. Sericitic-silicic alteration appears to be confined to late-stage quartz-hematite breccias. According to Barton & Johnson (2000), the alteration observed around IOCG deposits can be explained by the fluid chemistry. They identified two broad endmember types of hydrothermal IOCG deposits with a continuum of hybrid deposits between (Table 7.2). One end-member is typified by relatively hightemperature mineralisation, and relatively high K/Na and Si/Fe ratios in the alteration; that is to say that potassic and silicic alteration is dominant. The fluids are distinctive of magmatic fluid sources. These deposits represent the highest potential for the development of economic IOCG deposits. The other end-member is typified by more oxide-rich, sulphide-poor mineralisation, low Si/Fe ratios, and voluminous alkali-rich alteration with sodic types exceeding potassic types. The fluids of this group reflect the influence of

8 Figure 7.3. Generalised distribution of styles of alteration about a mineralised vein or body, with intense chloritisation, hematisation and silicification (Fe Si) occurring closest to the vein (< 2 m), sericitic alteration (H) in close proximity (< 5 m) which may possess an epidote and/or silicic component, K-metasomatism (K) in a more broad pattern around the vein (100s of m), and albitisation (Ca Na) occurring on a regional scale. Figure 7.4. Schematic cross section representing proposed relationships of alteration zoning in iron oxide (-Cu-U-REE-Au) deposits, drawn to represent examples in volcanic and plutonic host rocks (from Hitzman et al., 1992)

9 non-magmatic brines, and due to the sulphide-poor characteristic of this type, metals are less likely to be retained to form economic ore deposits. The intermediate or hybrid examples of these two end-members are influenced by both magmatic and non-magmatic fluids. All of the examples discussed in this study were formed by the effect of two distinct fluids. It is likely that the case of either end-member alone would be the exception rather than the rule. An important derivation from the Barton & Johnson (2000) model is the understanding of the source and distribution of metals within the systems and the controls on ore deposition. The characteristics of each type is summarised in Table 7.2. Table 7.2. Characteristics of deposits from systems dominated by magmatic fluid and nonmagmatic fluids (reproduced from Barton & Johnson, 2000). For magmatically sourced systems: Histories will tend to be clearly related to one or a few magmatic events, Mineralisation is correlated with igneous compositions and perhaps texturally distinct units, Moderate volumes (a few km 3 ) of alkali-rich alteration (K>Na in almost all cases), most of which is proximal; distal acid alteration carries metal but generally less overall, Metals commonly correlated with high-t, magnetite-rich, quartz-rich centre of system, Cooling is the fundamental trap, locally aided by wall-rock reaction and mixing, Favourable areas will thus: a) Need proper magma types, b) Be near the tops of (former) magma chambers, c) Be localised with intense Si and K + Fe (as magnetite) metasomatism. For externally sourced systems: Histories can be prolonged and not clearly related to individual magmatic events, Large volumes (many km 3 ) of alkali-rich (Na- Ca, + K) alteration, much of which may be distal (and metal depleted), Most metals commonly external to the high-t, magnetite-rich centre of the system, Nature of the metal traps is varied (mixing, wallrock reaction, boiling, cooling); good traps may not be present, thus many well-developed hydrothermal systems can be barren (or have metals trapped distal to the iron-oxides), Favourable areas will thus: a) Need an external brine source (surface, connate, remobilised), b) Be structurally or stratigraphically controlled or near a heat source, c) Be localised in upwelling zones with Na, K or H + Fe (magnetite or hematite) + Si metasomatism

10 The fluids of the Bushveld-type deposits have been shown to be dominated by a high-temperature magmatic component (Borrok et al., 1998; Freeman, 1998). Mineralisation is generally situated near the top of the granite sheet, with potassium metasomatism alteration assemblages dominating over Na-Ca metasomatism. Sericitic alteration is most proximal to mineralisation (although not necessarily proximal to the source of heat or fluid), and may possess an intense silicic component. This sericitisation is commonly obscured or overprinted by chloritic or hematitic alteration which accompanies mineralisation. None of the occurrences in the study area, however, demonstrate significant sulphide contents and probably did not present suitable metal trap sites. The sulphide contents of Vergenoeg, on the other hand, may be as much as 60 % of the volume of some rock assemblages (Fourie, 2000). The Bushveld-type mineralisation also possesses some characteristics of the externally derived fluid end-member. Mineralisation appears to be structurally controlled, distal albitisation has been noted by other authors (Kleeman & Twist, 1989; Freeman, 1989), and some appreciable metal deposits are distal to the granites and Fe-oxides (e.g. Rooiberg Sn Mine, Phalaborwa, etc.). Evaporitic halite casts are observed in sandstones of the Transvaal country rocks (Eriksson et al., 2004), although there exists no evidence that the intrusion of the Bushveld Complex caused dissolution of evaporate layers or that they had any influence on subsequent mineralisation. The system that produced the Bushveld-type deposits may have derived fluids from different sources but appears to have been dominated by magmatic fluids. An important feature of this model is the fact that similar geological features may be generated under different conditions, some of which may produce ore deposits and others only barren ironstones. Oliver et al. (2004) emphasised the importance of sodic alteration in the generation of IOCG deposits and postulated that regional albitisation was

11 responsible for the generation of abundant Fe and K in the fluid (Figure 7.5). In their model, fluids responsible for regional albitisation are coeval with the intrusion of the granite batholiths and are derived, at least in part, from the crystallising plutons. In order to produce intense albitisation, a highly saline fluid is required. These fluids may have been derived by dissolution of evaporitic layers in country rocks, or from the unmixing of a NaCl-H 2 O-CO 2 + CaCl 2 fluid derived from the intrusive granitoids into hypersaline brine and CO 2 -rich gas (Pollard, 2001). Figure 7.5. Schematic cross-section of the Eastern Mt. Isa Block Succession explaining the distribution and generation of IOCG deposits and the likely chemical reaction paths between source rocks, albitisation and ore deposits. Black arrows are inferred pathways of brines; white arrows are speculated sulphur-bearing fluids. Fluid modification from albitisation indicated by the variable grey shading (reproduced from Oliver et al., 2004)

12 According to Oliver et al. (2004), the optimal conditions for ore genesis require: a precursory modification of the fluid from albitisation to release K and Fe, a structural trap, a reactive host rock, and a sulphur source. Barren ironstones will be produced along the modelled albitisation path where sulphur is lacking. The availability of Cu is not necessarily a limiting factor, as all pyrite-magnetite ironstones appear to contain chalcopyrite. The source of sulphur in the system is postulated to have come from crystallising or leached gabbros (or possibly mantle-derived fluids), which is constrained by δ 34 S values of sulphides from the deposits near 0 (Mark et al., 2000; Baker et al., 2001). Also, although the data of Oliver et al. (2004) strongly support a major magmatic component to the alteration and mineralisation (including Cu), the possibility of exotic NaCl replenishment cannot be excluded. Their model supports the formation of ore by the mixing between magmatic S-bearing fluids and saline brines

13 7.2. General Discussions on Aspects of the Mineralisation Hematite Stage vs. Magnetite Stage The stability of Magnetite Stage assemblages suggest high temperature reducing conditions of mineralisation, whereas Hematite Stage assemblages represent the lower temperature, oxidising equivalent. Magnetite, therefore, develops first in a magmatic-hydrothermal system, and as in the case of Vergenoeg, occurs with associated fluorite, fayalite and sulphides. In order to produce Iron Oxide-Copper- Gold deposits, two important factors are required at the time of the hematite stage. Firstly, the existence and preservation of a plumbing stockwork and breccia, secondly perhaps even more importantly, the presence of sufficient sulphur to precipitate and concentrate metals. The sulphur may be sourced from the magnetite stage sulphides or from surrounding/adjacent sulphur-rich country rocks (as discussed in the previous section; c.f. Figure 7.3). Unless sulphur is present in the mineralising system, sulphides cannot form even if Cu and Au are present. This has been shown from fluid inclusions containing 5000 ppm Cu, however, the host rock was barren as no sulphides were present to scavenge and precipitate the metals. The metals were essentially lost from the system. It is well known that within large IOCG provinces a great majority of Magnetite-Stage Iron Oxide bodies are barren of Cu-Au mineralisation. Previously proposed ideas that magnetite-stage and hematite-stage IOCG deposits may be mutually exclusive appear now to be unfounded. Even in the classic example of Olympic Dam, which until now has been purported to be entirely hematite-stage, without the presence of the magnetite stage, appears to be untrue. Remnant magnetite cores to hematite masses have been observed in Olympic Dam ores, suggesting that the hematite is a later stage overprint of an earlier magnetite-bearing ore. This feature is likewise observed in Ruigtepoort ores. Within the Salobo Deposit of the Carajás Mineral Province, Brazil, the mineralization consists mainly of iron oxides where the proportion of magnetite is greater than that of hematite. These occur with dissemination of chalcopyrite,

14 bornite and chalcocite (Requia & Fontboté, 1999). Accessory ore minerals are hematite, molybdenite, ilmenite, uraninite, graphite, digenite, covellite and sulphosalts. The suggestion of this is that whilst the hematite stage is an important component of the formation of IOCG deposits, it may not be necessary for this late-stage oxidation to be absolutely dominant in order to produce economic mineralisation. The Role of Fluorine The role of fluorine in IOCG mineralising systems remains poorly understood. Few workers analyse fluorine as a matter of course and as a result, a paucity of data exists with regard to the significance of this element to the genesis of IOCG deposits, if any. Fluorine is a large, weakly charged and easily polarised negative ion, and is one of the most strongly electronegative of all the ions. It has a powerful affinity for large electropositive, weakly polarised cations, such as calcium. Ionic substitution is largely temperature dependent, but also dependent on the ionic radius and oxidation state of the substituted ion, in particular dominated by those most resembling calcium, such as the lanthanides and the elements yttrium, strontium and sodium (Crocker et al, 1988). A lot is known of the activity and role of other complexing agents, such as Cl -, which is well documented in Au mineralising systems. Fluorine on the other hand is difficult to analyse for and is further digested by normal wet chemistry analytical techniques. The aggressive nature of fluorine compounds such as HF is known, in particular the ability of this acid to digest silica in great volumes. It would therefore follow that the presence of this acid could promote the development episyenitic granites; the myrialitic cavities of which subsequently exploited for mineral/metal precipitation

15 Fluorine is recognised to be an important component of specialised two-feldspar granites responsible for tin-tungsten mineralisation, where fluorine may be present in amounts as much as 4 wt% F, although more usually < 1 wt% F. An important characteristic of fluorine-rich granites is their position on a Qz-Ab-Or ternary diagram (c.f. Figure 5.3) which is displaced towards the albite apex suggesting that they may be produced by albitisation of a pre-existing fluorine-poor granite. The presence of fluorine in a melt has been shown to have the effect of replacing alkali-feldspar with quartz as the first phase to crystallise from the melt (Manning, 1982). It has been determined experimentally that the solidus temperature of a granite may be greatly depressed by the presence of fluorine to as low as 630 C at 1 kbar with 4 wt% F (Smith & Parsons, 1975). The low temperatures that fluorine-rich granites exist at are below the solidus for normal granites. This suggests that they may crystallise late in the history of a compound batholith and may be available for late-stage intrusion. Therefore during crystallisation, increasingly volatile-rich magma will enhance equilibrium between crystallising feldspars and produce granite compositions which plot nearer the albite apex, but of primary magmatic origin. The implication is that one may obtain compositions that are indicative of albitisation but primary in origin. The anomalous number of small fluorite occurrences in the Bushveld Complex is considered to reflect the great volume of the felsic component of the Complex, with the highly-incompatible fluorine being super concentrated in the late-fluid derivatives of the crystallising felsic magmas. More work needs to be done to determine the effect of fluorine in IOCG systems, in particular, in characterising the regional albitisation described in many accounts to determine whether it is a primary feature due to the increasing volatile content, or indeed a secondary alteration effect

16 Carbonatite Association In some IOCG districts a spatial and temporal relationship between carbonatite and other alkaline intrusions has been recognised (Harmer, 2000 b; Vielreicher et al., 2000). The significance of this association, however, may not yet be fully understood. It is likely that magmatism associated with mantle pluming involves crustal melting of deep continental crust. This lower crust may be composed of igneous rocks and carbonate sediments, resulting in peculiar magmatic assemblages. The distribution of alkaline intrusives indicates that it is possible to generate alkaline magmas in both intra-plate, anorogenic magmatic environments and in extensional rift environments. The Bushveld Complex is intruded by a conspicuous number of small carbonatites and other alkaline intrusives. Incomplete or unreliable geochronology exist for these intrusions, however two principal sets are defined one Bushveld-aged (~2050 Ma) and the other Pilanesberg-aged (~1300 Ma). The distribution of the known alkaline intrusives is shown on Figure 7.6. The largest of these intrusions is the Phalaborwa Complex located to the east of the Bushveld Complex. Although it does not lie within the limits of the Bushveld Complex it has been shown to be coeval with Bushveld magmatism. The geological and mineralogical characteristics of the intrusion, which appear consistent with those of IOCG-type deposits, have drawn some authors to regard it as such (Harmer, 2000 b; Vielreicher et al., 2000). The association made should certainly warrant further attention. (See section 2.4 for a description of the Phalaborwa deposit). The Schiel Complex, also located in the Northern Province, South Africa, is another large syenitic complex with subordinate carbonatite, foskorite, and syenogabbro. The deposit of apatite, associated with magnetite and vermiculite was discovered in Subsequent exploration revealed ore reserves of 36 million tonnes at 5.1% P2O5 in the weathered zone to a depth of 39.6 m (Verwoord 1986,

17 Figure 7.6. Distribution of alkaline rocks and carbonatite around the Bushveld Complex (after Woolley, 2001; Crocker et al., 2001). Bushveldage alkaline intrusives shown in red, Pilanesberg-age alkaline intrusives shown in green and undated alkaline intrusives shown in blue

18 1993). Copper mineralisation is low-grade with the best known intersection of 0.23 % Cu over 1.5 m (Wilson, 1998). The determined age of 2, million years (Walraven et al., 1992) is the same as the Bushveld age within error. It should be borne in mind that only recent geochronology of the Phalaborwa carbonatite indicated an age coeval to the Bushveld Complex. A reassessment of the age of the Schiel Complex may provide insight into the abundant alkaline intrusives spatially associated with the Bushveld Complex. In close proximity to the study area are three similar carbonatite intrusions, namely Kruidfontein, Tweerivier and Nooitgedacht, which lie on an approximate north-south line (Figure 7.7). The Kruidfontein Carbonatite Complex is the largest of the three, consisting principally of pyroclastic rocks in a 5 km diameter volcanic caldera. The Complex consists of ringed lithologies comprising an inner carbonatite bedded tuff and outer silicate tuffs and breccias. Subsidiary flows of trachyte, phonolite, rhyolite and other lavas have been identified (Woolley, 2001). Breccia fragments consist of a combination of volcanic and country rock lithologies, namely trachytic, phonolitic, rhyolitic and sövitic rocks, banded ironstones, quartzite, dolomite, schist and altered basic rocks. Extensive replacement by quartz, carbonate and fluorite has taken place and intense K-metasomatism is recognized (Clarke & Le Bas, 1990). Anomalous values of REE, Au, Mn, and Y have been obtained by chemical analysis (Clarke et al., 1994; Clarke & Le Bas, 1990; Pirajno et al., 1995). The Tweerivier Carbonatite Complex is the only of the three hosted in Bushveld granites, which have been locally fenitised. The geology of the Complex comprises two remarkably different halves. The northern half is composed predominantly of dolomitic carbonatite and the southern half is composed of gabbros and anorthositic gabbros, with cross-cutting sövite sheets and a radioactive, silicified ferruginous rock (Woolley, 2001). The sövite sheets include

19 accessory phlogopite, apatite (but may constitute up to 50 % of some rocks), magnetite, calcite, dolomite, pyrite, rare olivine and baddeleyite. In the Nooitgedacht Complex, carbonatite is the most abundant rock in outcrop with minor fenites, sövites and nepheline syenites. Planar layers of magnetite and apatite occur with phlogopite, pyrite, pyrochlore, chondrodite and fluorite (Woolley, 2001). Sulphides reported are galena, pyrite, pyrrhotite and chalcopyrite. Substantial soil Cu-anomalies suggest copper enrichment at depth in places (Harmer, 2000 b). 27 o o o 00 Nooitgedacht Detailed Study Area Kruidfontein Transvaal Sediments Lebowa Granite Suite Karoo Sediments 25 o 15 Tweerivier Figure 7.7. Distribution of some carbonatite and alkaline intrusions near the study area. Base map is taken from 1: Geological Sheet, South African Council for Geoscience

20 7.3. Summary and Conclusions The evaluation of the farms near Ruigtepoort has provided insight into the development of geological characteristics consistent with IOCG-type mineralisation. The geological mapping has determined the distribution and extent of mineralisation on the farms Ruigtepoort, Blokspruit, Slipfontein, Elandslaagte and Doornfontein and characterised the morphologies, mineral assemblages, alteration haloes and mineral potential of many of these deposits and occurrences. These geological characteristics were compiled with regional results for comparison and compliment the existing inventory of known mineralisation A clear relationship between mineralisation and alteration has been established from both petrographical and geochemical observations, and the distribution patterns of alteration types around mineralisation may permit vectoring towards mineralisation. Certainly, the styles of alteration are indicative of the mineralising system in operation, such that the recognition of certain alteration types may be sufficient to identify prospective terranes. In terms of IOCG deposits, this entails the recognition of regional albitisation, broad but intense K-metasomatism, and sericitisation, silicification and chloritisation in the immediate vicinity of mineralisation. The identified characteristics of the deposits and occurrences of the study area appear consistent with those of IOCG deposits. On a regional scale, they are related to anorogenic, A-type granite magmatism that is Palaeoproterozoic in age. The bodies are strongly structurally-controlled forming veins, breccia-pipes, and sub-horizontal mantos, which occur near or in the intersections of structural elements that may be splays from regional structural features. Mineralisation is dominated by a late-stage alteration assemblage of Fe-oxide (hematite and magnetite)-quartz-fluorite. The primary assemblage is considered to have comprised ferroactinolite-magnetite-fluorite-apatite (britholite) and is preserved in the Ysterkop North deposit. This assemblage closely resembles the primary

21 assemblage at Vergenoeg, which consists of fayalite-magnetite-fluorite-ree minerals (Borrok et al., 1989). The difference between these assemblages is principally of the presence of ferroactinolite and fayalite, which may represent a variation in fluid characteristics or an as yet unrepresented alteration of fayalite to ferro-amphibole. A secondary, intermediary assemblage is identified at Vergenoeg consisting of magnetite-siderite-quartz-apatite-sulphides-ree minerals. A similar secondary assemblage may exist for the Ruigtepoort-type deposits but has been completely overprinted by subsequent alteration. Intense chloritisation has completely replaced actinolite-dominated assemblages, including the immediate granite country rocks, such that it may be extremely difficult to determine intermediary assemblages. Anomalous metal concentrations in each of the occurrences and deposits appear consistent with expected metal associations in IOCG deposits. Ruigtepoort possesses anomalous Au, Cu and LREE; Slipfontein anomalous Cu LREE and Mo; Blokspruit anomalous Cu and LREE; Elandslaagte anomalous LREE and Mo; and Doornfontein anomalous LREE. Metal concentrations of the occurrences however, remain far too low to be economic and may be a consequence of insufficient sulphur in the system at the time of mineralisation. Alteration haloes around mineralisation appear to follow proposed IOCG models with intense K-metasomatism evident in many granites of the study area, intense sericitisation-silicification in closing proximity to mineralisation, and pervasive chloritisation of ore zones, affecting in particular of ferro-amphibole assemblages. Hematisation is observed localised around fractures and ore zones affecting magnetite and locally staining country rocks (beyond the normal deuteric alteration). The fluid characteristics for Ruigtepoort-type deposits has been inferred from fluid inclusion studies of the Vergenoeg deposit. Two populations are recognised,

22 the first a high-temperature, high salinity magmatic fluid and the second a lowertemperature, lower salinity fluid with a meteoric component. Mineralisation is modelled as having developed in the zone of fluid mixing. The similarities between the occurrences of the study area and the Vergenoeg deposit are convincing, and these deposits are regarded as having been formed under similar conditions. Equally, the similarities that exist between Vergenoeg, Ruigtepoort and the external examples of Olympic Dam and Salobo, indicate that each of these deposits formed under similar geological and chemical conditions with the capacity to form IOCG deposits. The proposed model for Ruigtepoort and related deposits is illustrated in Figure 7.8. The principal feature corresponds to the level of formation. Vergenoeg exhibits features of exhalative activity, whereas Ruigtepoort exhibits only extensive fracturing, consistent with a deeper level of formation. The local stratigraphy near Ruigtepoort is understood and indicative of a lower level of formation, with ascending stratigraphic units occurring northwards from Ruigtepoort towards Rooiberg. The recognised alteration patterns and the fluid characteristics are consistent with the magma-derived fluid model of Barton & Johnson (2004) (c.f. Figure 1.3). In this model, early magmatic fluids are only capable of producing low-level, barren ironstones. Regional albitisation is apparent but may be removed from the sites of mineralisation. Cu-Au mineralisation is developed in high-level zones, in particular, in the zone of fluid mixing between the original magma-derived fluids and surface waters. An external sulphur source is required to concentrate metals. In this respect, the sulphur-poor granites represent a poor prospect for significant metal accumulations. The volume of ore contained in the Vergenoeg deposit indicates that the Bushveld granites are indeed prospective for IOCG-type mineralisation of economic proportions. The surrounding Transvaal sediments and sedimentary inliers may present more suitable metal trap sites, in particular, sulphur-rich lithologies

23 The Bushveld-type mineralisation examined in this study is considered to be consistent with other IOCG deposits and the proposed models for their formation. Careful consideration of the salient characteristics may yet yield a world-class IOCG deposit in arguably the largest anorogenic terrane in the world. The mineralised occurrences of the study area are too small to be economic and are too deeply eroded. By analogy to other deposits in the Bushveld Complex, any larger ore-bodies of the IOCG-type that may have been formed in this area are likely to have been eroded away. Figure 7.8. Schematic model of level of formation for some Bushveld-type Fe-F deposits in conjunction with alteration and fluid characteristics of an IOCG model where magmatic fluids dominated (c.f. Figure 1.3) (modified after diagrams of Crocker et al., 1988; and Barton & Johnson, 2004)