Y. YAO,, * Department of Geology, University of the Witwatersrand, Private Bag 3, WITS 2050, Johannesburg, South Africa P. J.

Transcription

1 Economic Geology Vol. 96, 2001, pp Fluid Characteristics of Granitoid-Hosted Gold Deposits in the Birimian Terrane of Ghana: A Fluid Inclusion Microthermometric and Raman Spectroscopic Study Y. YAO,, * Department of Geology, University of the Witwatersrand, Private Bag 3, WITS 2050, Johannesburg, South Africa P. J. MURPHY,** Raman-Luminescence Laboratory and Economic Geology Research Unit, University of the Witwatersrand, Private Bag 3, WITS 2050, Johannesburg, South Africa AND L. J. ROBB* Department of Geology, University of the Witwatersrand, Private Bag 3, WITS 2050, Johannesburg, South Africa Abstract Fluid inclusions in vein quartz from 10 granitoid-hosted gold deposits and prospects in the Birimian terrane of Ghana, as well as from the Sansu mine (Ashanti shear zone type) and quartz veins in the nonmineralized Princess Town granodiorite, were studied by microthermometry and Raman microspectrometry. Fluid inclusions from the granitoid-hosted gold deposits are dominated by aqueous H 2 O-CO 2 -NaCl type 1 and liquid CO 2 -N 2 ±CH 4 type 2, with minor (<10%) aqueous H 2 O-NaCl type 3. Type 1 inclusions show large variations in CO 2 phase volume proportions (10 90%) at 25 C and have salinities commonly between 0 and 6 wt percent NaCl equiv. Their bulk densities fall in a major range from 0.62 to 1.11 g/cm 3. Type 2 inclusions show no visible H 2 O phase at room temperature and have bulk densities between 0.30 and 0.92 g/cm 3. Type 3 inclusions, containing 10 to 20 vol percent H 2 O vapor, have salinities of 1 to 8 wt percent NaCl equiv and bulk densities of 0.77 to 1.03 g/cm 3. In most cases, the three types of fluid inclusions coexist as groups in individual quartz grains of the samples studied. These fluid inclusions are interpreted to be trapped during phase separation of an originally homogeneous H 2 O-CO 2 fluid, with low salinity (<6 wt % NaCl equiv) and moderate to high density ( g/cm 3 ). The type 1 fluid inclusions are suggested to be heterogeneous mixtures of the two end members of type 2 and 3 inclusions. Fluid immiscibility is documented by petrographic characteristics and microthermometric results of the three types of inclusions. Trapping temperatures and pressures, estimated from microthermometry of type 3 inclusions, equation of state, and the P-T-X nature of the H 2 O-CO 2 -NaCl system, are mainly between 200 and 350 Cand 1 and 3 kbars for the gold deposits. By contrast, fluid inclusions in the Sansu mine at Ashanti mostly comprise the liquid CO 2 -N 2 ±CH 4 type 2. The inclusions are considered to represent postentrapment modifications of trapped fluids. In addition, fluid inclusions in barren vein quartz from the Princess Town granodiorite comprise the low-salinity (commonly <6 wt % NaCl equiv) H 2 O-NaCl type 3. Raman microspectrometry shows that gaseous compositions of fluid inclusions from both granitoid- and shear zone-hosted gold deposits are mainly composed of CO 2 (80 95 mol %), with significant amounts of N 2 (2 20 mol %) and CH 4 (0 10 mol %). The distinct low-salinity H 2 O-CO 2 -rich fluids from the granitoid-hosted gold deposits are comparable in composition to those from the major Ashanti and Tarkwaian types of gold deposits in the Birimian terrane of Ghana. These fluids are most likely to be metamorphic in origin and associated with the waning stages of the regional Birimian orogeny. Gold deposition within the Birimian granitoids was related to fluid phase separation and sulfidization of host rocks during hydrothermal alteration and mineralization. The present study, together with previous publications, suggests that fluid inclusions are characterized by H 2 O-CO 2 -NaCl and/or CO 2 -N 2 ±CH 4 types in mineralized areas, whereas H 2 O-NaCl fluid compositions dominate in barren areas. Fluid inclusion characteristics may, therefore, be a useful tool for regional gold exploration in the Birimian terrane of Ghana. Introduction THE BEST KNOWN and most productive gold deposits in the Birimian terrane of Ghana are represented by the Ashanti shear zone and Tarkwaian quartz-pebble conglomerate types. The Ashanti-type shear zone-hosted deposits have a total gold inventory of some 1,100 t (Milési et al., 1991), whereas the Corresponding author: , 065yyao@cosmos.wits.ac.za *Present address: Economic Geology Research Institute, School of Geosciences, University of the Witwatersrand, Private Bag 3, WITS 2050, Johannesburg, South Africa. **Present address: Centre for Earth and Environmental Science Research School of Earth Sciences and Geography, Kingston University, Kingston upon Thames, Surrey KT1 2EE, UK. Tarkwaian conglomerate-hosted deposits have gold reserves of around 550 t (Milesi et al., 1991; Gold Fields Ghana Ltd., pers. commun., 1997). A third type of gold deposit, increasingly receiving the attention of exploration in the region, are the granitoid-hosted gold deposits, which contain estimated gold reserves and/or resources of over 100 t metal at an average grade of 2 to 4 g/t (Ashanti Goldfields Co.; Mutual Ghana Ltd.; Takoradi Gold N.L., pers. commun., 1997). The present study focuses on these granitoid-hosted gold deposits and describes the nature of fluids associated with mineralization and their relationships to the shear zone-hosted gold deposits in the region. The Paleoproterozoic (Birimian) geology of Ghana is defined by five relatively evenly spaced northeast-southwest striking /01/3200/ $

2 1612 YAO ET AL. volcanic belts, separated by intervening sedimentary basins (Fig. 1). Volcanic and sedimentary rocks are thought to be coeval with the sediments representing lateral facies equivalents of the volcanic belts (Leube and Hirdes, 1986; Leube et al., 1990); these successions formed at about 2.0 to 2.2 Ga (Hirdes et al., 1992; Taylor et al., 1992; Oberthür et al., 1998). The volcanic belts are composed mainly of deformed basaltic metalavas, intruded by coeval tonalitic to granitic plutons (Belt-type granitoids: 2136 ± 12 to 2179 ± 2 Ma, see Birimian Granitoids and Their Associated Gold Deposits ), which are in turn overlain in certain of the belts by Tarkwaian sedimentary rocks (conglomerates, sandstones, and slates) Sample locality and (non-)mineralization type; = Basin granitoid-hosted gold deposits and prospects; 1= Ayankyerim; 2 = Yamensakrom; 3= Nhyiaso; 4 = Chirawewa East; 5 = Chirawewa West. = Belt granitoid-hosted gold deposits and prospects; 6= Mporhor; 7= Abore North; 8 = Mpesetia; 9= Dokrupe; 10= Sakpa. = Shear zone-hosted gold deposits (Ashanti type); 11= Sansu. = Non-mineralized Belt granitoid; 12= Princess Town. FIG. 1. Simplified geologic map of Ghana (modified after Hirdes and Leube, 1989; John et al., 1999), showing sample localities of Basin (1 5) and Belt (6 10) granitoid-hosted gold deposits and prospects in this study. Also shown are the Sansu gold deposit at Ashanti (11) and barren quartz veins from the nonmineralized Belt granitoid at Princess Town (12) /98/000/ $

3 FLUID CHARACTERISTICS IN Au DEPOSITS, BIRIMIAN TERRANE, GHANA 1613 deposited between 2095 ± 10 and 2132 ± 3 Ma (Hirdes and Nunoo, 1994). The intervening sedimentary basins (volcaniclastics, wackes, argillites) formed between 2135 ± 5 and 2184 ± 3 Ma (Davis et al., 1994) and are also intruded by areally extensive Basin-type granitoids (2086 ± 4 to 2125 ± 2 Ma, see Birimian Granitoids and Their Associated Gold Deposits ). Regional deformation and metamorphism of the Birimian terrane are related to the Eburnean tectonothermal event at about 2100 Ma (Eisenlohr and Hirdes, 1992; Oberthür et al., 1998). Major thrusts and shears generated during the deformation form channels for mineralizing hydrothermal fluids (Oberthür et al., 1994). The metamorphic grade of Birimian rocks is generally considered to be of greenschist facies (Leube et al., 1990; Oberthür et al., 1994, 1996). However, peak metamorphic conditions of the rocks reached amphibolite facies at about 500 to 650 C and 5 to 6 kbars for the Eburnean orogeny (John et al., 1999). The greenschist facies conditions (<450 C and <3 kbars) of Birimian rocks are retrograde, during which gold mineralization in the Birimian took place; the clockwise, retrograde P-T path of Birimian metamorphism indicates simultaneous cooling and decompression of the Birimian terrane during exhumation (John et al., 1999). Toward the end of the deformation the late kinematic Basin granitoids were intruded into the basin sediments, and the lateral southeast-northwest compression of the Birimian terrane was basically completed. Since then, the Birimian rocks have occasionally been affected by younger thermal-tectonometamorphic events (Eisenlohr and Hirdes, 1992; Hirdes et al., 1992; Taylor et al., 1992; Zitzmann et al., 1993; Oberthür et al., 1998). The Ashanti-type gold deposits occur in strongly folded and sheared volcanic rocks, mainly along the western margins of the Ashanti belt, and comprise quartz sulfide (mainly arsenopyrite) veins in steeply dipping shear zones, as well as disseminated sulfides (mainly pyrite) in the metavolcanic rocks. It has been suggested that these gold deposits are structurally controlled and synmetamorphic in origin (Oberthür et al., 1994). Tarkwa-type conglomerate-hosted gold deposits mainly occur in the southern parts of the Ashanti belt and are considered to be paleoplacers considerably modified by input of later fluids that have pervasively altered the original mineral assemblage (Milési et al., 1991; Klemd et al., 1993; Hirdes and Nunoo, 1994). Fluid inclusions in vein quartz from the Ashanti type and in quartz-pebble and vein quartz from the Tarkwaian type of gold deposits are dominated by high-density (up to 1.1 g/cm 3 ) CO 2 - N 2 ±CH 4 compositions, with minor aqueous H 2 O-CO 2 -NaCland H 2 O-NaCl-bearing fluid inclusions. The latter have variable salinities from 0 to 10.5 wt percent NaCl equiv for the Tarkwaian (Klemd et al., 1993) and from 0.4 to 25.3 wt percent NaCl equiv for the Ashanti gold deposits (Schwartz et al., 1992; Oberthür et al., 1994; Schmidt Mumm et al., 1997). The origin of these fluids is currently a topic of debate. Schmidt Mumm et al. (1997, 1998) argued that mineralizing fluids in the Ashanti belt were derived from a CO 2 -rich, nearly H 2 O- free hydrothermal system. By contrast, Schwartz et al. (1992), Klemd and Hirdes (1997), and Klemd (1998) suggested that CO 2 -rich liquid and aqueous H 2 O-rich fluid inclusions resulted from fluid immiscibility and/or grain boundary migration-recrystallization during the ductile-brittle deformation of the Birimian terrane. Schmidt Mumm et al. (1997) also reported that primary fluid inclusions in vein quartz from the Ayanfuri granite-hosted gold deposit mainly comprise CO 2 - rich liquid and aqueous H 2 O-rich types, of low salinity (<3 wt % NaCl equiv). It was suggested that these fluids were trapped during phase separation of mineralizing solutions (Schmidt Mumm et al., 1997). This paper presents the results of a comprehensive microthermometric (>2,600 measurements) and Raman spectroscopic (>90 analyses) study of fluid inclusions in vein quartz from 10 major granitoid-hosted gold deposits and prospects in the Birimian terrane of Ghana. In addition, the Sansu mine from the Ashanti shear zone system and barren quartz veins from the nonmineralized Princess Town granodiorite were also studied for comparative purposes. The samples studied cover a wide area from southeast to northwest Ghana (Fig. 1). The objectives of this study are to document the regional characteristics of mineralizing fluids depositing gold in the Birimian granitoids and to construct the P-T conditions of entrapment of the fluids. On the basis of fluid inclusion data, as well as geologic setting and geochemical characteristics of ores, an attempt is made to investigate the genetic link, if any, of ore-forming fluids between the shear zone- and granitoid-hosted gold deposits and to discuss the genesis of gold mineralization in the Birimian granitoids. The relevance of fluid inclusion characteristics as an indicator for regional gold exploration in the Birimian of Ghana is also discussed. Methodology Fluid inclusion microthermometry was carried out at the University of the Witwatersrand, South Africa, using a Linkham TH 600 heating-cooling stage and an OLYMPUS VANOX microscope with a ULWD MSPlan 50 lens. The SYNFLINC synthetic standards of CO 2 and aqueous H 2 O fluid inclusions were measured for calibration of the stage in this study. Most measurements were made at heating rates of 0.2 to 10 C/min. Clathrate melting temperatures (T m(clathrate) ) were determined by temperature cycling (Roedder, 1963; Diamond, 1990), with heating rates near T m(clathrate) of 0.2 to 0.5 C/min. The reproducibility of measurements was better than ±0.2 C below 30 C and ±2 C at total homogenization temperature when the chips were centered in the specimen holder. Calculations of salinity, mole fractions (X H2 O, X CO2, X NaCl ), density of carbonic phase and bulk fluids (ρ CO2, ρ b ) and pressures were made using published equations of state: Bowers and Helgeson (1983) for H 2 O-CO 2 -NaCl fluid inclusions; Holloway (1981) for carbonic liquid fluid inclusions; and Zhang and Frantz (1987) for H 2 O-NaCl fluid inclusions. Raman analyses of fluid inclusions were carried out at the University of the Witwatersrand, using a Jobin-Yvon T64000 spectrometer in a single spectrograph mode. The nm line of a Coherent Argon ion laser was used, with a source power of 700 mw (reduced to <100 mw at the sample surface) and a focused spot size of 1 µm. The laser was focused onto samples using an OLYMPUS BH-40 microscope and MPlan 100 lens. Spectra were acquired using a liquid nitrogen-cooled Spectrum One CCD detector with a pixel size of 27 µm (1, pixels). A grating of 1,800 g/mm gave a resolution of <2 cm 1. Integration time was generally 150 s /98/000/ $

4 1614 YAO ET AL. for each gas range (with 2 accumulations averaged and cosmic rays rejected). Data manipulation was carried out using the Jobin-Yvon-Spex SpectraMax software and peak areas calculated using peak-fitting routines in the Microcal Origin program. The frequency ranges scanned meant that any other possible components, such as other hydrocarbons or H 2, would have been detected if present. Where nitrogen was detected, an analysis under identical conditions was carried out in quartz near the inclusion, at the same focal depth, to test for contamination by atmospheric nitrogen. When atmospheric nitrogen was detected the appropriate peak area was subtracted from the nitrogen peak for the inclusion, and this usually represented no more than 1 or 2 mol percent N 2. Molar proportions were calculated using the relative Raman scattering cross sections given by Burke (1994), i.e., 2.5 for CO 2 (1, ,388 cm 1 bands), 7.5 for CH 4 (2,917 cm 1 bands), and 1 for N 2 (2,331 cm 1 bands). The instrumental response for CO 2 and N 2 was tested using gas standards supplied by A. M. van den Kerkhof. No instrumental correction factor was found to be necessary, and the reproducibility was measured as ±5 mol percent N 2 at low N 2 concentrations and ±1.5 mol percent N 2 at higher N 2 concentrations. Detection limits for N 2 in the carbonic phase of fluid inclusions in the present study were typically around 1 mol percent, but this was better for CH 4 which is a stronger Raman scatter. The Raman analyses were conducted to obtain quantitative gaseous compositions of primary fluid inclusions. Gaseous molar volumes of fluid inclusions (V g) were estimated by interpolation and/or extrapolation in the T m -T h -contoured VX diagrams of Thiéry et al. (1994). As discussed by Hall and Bodnar (1990), melting temperatures of carbonic phases (T m(co2 )) for carbonic liquid fluid inclusions are insensitive to the vapor density. Thus, the T h(co2 ) and not the T m(co2 ) were used to estimate gaseous molar volumes, in combination with Raman analytical results of the corresponding fluid inclusions. Birimian Granitoids and Their Associated Gold Deposits Junner (1940) divided the Paleoproterozoic Birimian granitoids into two major types: an older granite (G1: Cape Coast type) and a younger granite (G2: Dixcove type). This classification was widely used in geologic maps of Ghana. Since the 1980s, the Birimian granitoids have been reclassified into Basin and Belt types on the basis of their geologic association (Mauer, 1986, 1990; Leube et al., 1990). From an economic point of view, both Basin and Belt types of granitoids are further divided into nonmineralized and mineralized plutons (Yao and Robb, 1998a). Nonmineralized Belt granitoids crop out within volcanic belts as small- and medium-sized plutonic bodies (Fig. 1) and are typically tonalitic to granodioritic in composition. The granitoids are characterized mainly by a metaluminous character, and their mafic minerals are hornblende and/or biotite. By contrast, the nonmineralized Basin granitoids were emplaced within basin sediments as large batholiths, and they are much more restricted in chemical composition, mainly comprising granodiorite and granite. These rocks dominantly show a marginally peraluminous character. Biotite is the main mafic mineral, and hornblende is rarely observed. Accessory minerals of the Basin granitoids are similar to those of the Belt granitoids and consist of epidote and titanite, with minor zircon, apatite, and opaque minerals (Yao and Robb, 1998b). U-Pb zircon and monazite isotope dating indicates that the Belt granitoids (2136 ± 12 to 2179 ± 2 Ma ) were emplaced 10 to 90 m.y. before the Basin granitoids (2086 ± 4 to 2125 ± 2 Ma; Hirdes et al., 1992, 1993; Loh and Hirdes, 1996; Oberthür et al., 1998; Yao and Robb, 1999). The age differences between both granitoid types may reflect a progressive accretion of magmatic areas from the southeast to the northwest of Ghana (Hirdes et al., 1992; Yao and Robb, 1999). On the basis of petrographic and geochemical characteristics, as well as whole-rock Rb-Sr, Sm-Nd, and Pb-Pb isotope signatures (Taylor et al., 1992), it is suggested that the Belt granitoids are I-type granites and formed by partial melting of basaltic rocks with compositional variations being controlled largely by amphibole fractionation. By contrast, the Basin granitoids are more akin to S-type granites and are considered to have generated by minimum melting of crustal materials with addition of a minor mantle component (Yao and Robb, 1998b; and references therein). Over 20 granitoids from both Belt and Basin types are now known to host gold mineralization (Yao and Robb, 1998a). Granitoids now represent a significant new target for regional gold exploration in the Birimian orogeny. Mineralized Belt and Basin granitoids occur as small steeply dipping intrusive stocks and are concordant with regional northeast-striking faults and shear zones. These plutons exist over a wide area in southwest and west-central Ghana but are concentrated mainly in the Ashanti, Asankrangwa, and Bole-Navrongo belts, as well as the Kumasi basin (Fig. 1). The mineralized Belt and Basin granitoids are dominantly granodioritic to granitic in composition and are comparable to their nonmineralized counterparts in terms of geological, mineralogical, geochemical, and petrogenetic characteristics. Typically, however, the mineralized granitoids occur as smaller intrusions (Yao and Robb, 1999). Yao and Robb (1999) have shown that gold mineralization within both Belt and Basin granitoids is similar, and that no specific granitoid type in the Birimian terrane of Ghana is associated with gold mineralization. The granitoids are considered to constitute favorable sites for mineralization because they provide preferential conduits for fluid flow due to their brittle character during regional deformation. The mineralization is characterized by both quartz veins and/or stockworks and pervasive alteration zones within granitoids. Quartz veins and veinlets commonly fill extension-related fractures and are locally deformed. They are frequently observed to crosscut each other from individual gold deposits, indicating that quartz veining took place in multiple stages of mineralization. Crosscutting relationships among the quartz veins, granitoids, and sheared wall rocks suggest that gold mineralization postdated the granitoid emplacement and was related to relatively late brittle structures developed during the Birimian orogeny. The ore mineral assemblage in all granitoid-hosted deposits is similar and dominated by pyrite and arsenopyrite in the early stage, with minor chalcopyrite, sphalerite, galena, native gold, and rutile in the late stage and locally developed secondary hematite. The gold is closely associated with arsenopyrite and pyrite and also fills microfractures in quartz grains, indicating that gold deposition was developed from early to late stages during mineralization /98/000/ $

5 FLUID CHARACTERISTICS IN Au DEPOSITS, BIRIMIAN TERRANE, GHANA 1615 The alteration mineral assemblage from the various granitoid-hosted deposits is also similar and comprises quartzsericite(muscovite)-carbonate-pyrite-arsenopyrite ± chlorite ± tourmaline. Early-stage pyrite, arsenopyrite, and quartz ± tourmaline are replaced by muscovite, which is in turn replaced by late-stage carbonates and chlorite. High-temperature alteration minerals, such as K feldspar, biotite, and amphibole, were not found, suggesting that the alteration was formed in typically mesothermal P-T conditions of about 200 to 400 C and 1 to 3 kbars (McCuaig and Kerrich, 1998). The similarity of sulfides between the ores and alteration zones suggests that the alteration was produced by the same mineralizing fluids. The lack of sulfates and clay minerals in the alteration zones suggests that the fluids were reduced, with near-neutral ph (about 5 6) during the mineralization process (Mikucki and Ridley, 1993). Chemical mass-balance calculations for altered versus unaltered granite samples from individual deposits demonstrate that most of the major oxides and trace elements (Rb, Sr, Ba, V, Cr, Co, Ni, Zr) were buffered by the original granitoid composition, suggesting that the alteration was developed in a rock-buffered system with respect to these components. The U-Pb isotope age of mineralization, derived from hydrothermal rutile, is between 2110 ± 32 and 2133 ± 21 Ma for a Belt-type granodiorite from the Sakpa mine in the Bole- Navrongo belt (U-Pb zircon intrusive age: 2137 ± 10 Ma; Yao and Robb, 1999). Mineralization in the Basin-type granitoids of the Obuasi and Ayanfuri mines of the Kumasi basin is between 2086 ± 4 and 2098 ± 7 Ma (U-Pb zircon intrusive age: 2105 ± 2 Ma; Oberthür et al., 1998). These ages are about 5 to 30 m.y. younger than the ages of corresponding granitoid emplacement, confirming field observations that the mineralization postdated granitoid intrusion. The age range of mineralization overlaps with that of Basin-type granitoid emplacement and, therefore, coincides with the second of the major magmatic pulses (e.g., Basin-type granitoid intrusion) of the Birimian orogeny. Mineralization, therefore, appears to be associated with a relatively late stage of the Birimian orogeny. Fluid Inclusions Petrography Fluid inclusions in vein quartz from five Basin granitoids at Obuasi (Nhyiaso, Ayankyerim, Yamensakrom) and Ayanfuri (Chirawewa East, Chirawewa West) in the Kumasi basin and from five Belt granitoids at Mporhor in the Ashanti belt, at Mpesetia and Abore North in the Asankrangwa belt, and at Bole (Dokrupe, Sakpa) in the Bole-Navrongo belt (Table 1) were studied by microthermometry and Raman microspectroscopy. For the purpose of comparison, fluid inclusions in vein quartz from the Sansu mine at Ashanti and in barren vein quartz from the nonmineralized Belt-type Princess Town granodiorite (Table 1) were also studied. The localities of these areas studied are shown in Figure 1. Forty-six doubly polished fluid inclusion sections ( µm) were examined under the microscope, of which 16 samples were selected to carry out the fluid inclusion study (Table 1). Vein quartz within granitoids in the studied samples forms euhedral and subhedral crystals and aggregates. Except for crosscutting microfractures observed in some quartz grains, the quartz shows only rare recrystallization and weak undulose extinction, which suggests that the quartz formed in a low-stress brittle regime, as indicated by field observations. In the case of the Sansu mine at Ashanti, however, quartz grains are highly orientated, but their axis directions were not determined. The quartz shows strong undulose extinction and common recrystallization, and microfractures crosscutting the elongated quartz grains are well developed. These characteristics indicate that the vein quartz at Sansu was formed by multistage ductile and brittle deformation. The structural features of quartz optically observed in this study suggest that primary fluid inclusions from the granitoid-hosted gold deposits represent original fluid entrapment without postentrapment modifications, whereas fluid inclusions from the Sansu mine at Ashanti have been subjected to postentrapment modifications, which are attributed to grain boundary migration-recrystallization during ductile to brittle deformation. The latter features have also been discussed by Schwartz et al. (1992), Klemd et al. (1996), Klemd and Hirdes (1997), and Klemd (1998). In most cases, the growth zones of quartz crystals were not clearly observed, although quartz grain boundaries are clearly recognized. Fluid inclusions that occur as isolated individuals, random groups, and planar arrays in intragranular quartz crystals were inferred to be primary, whereas fluid inclusions aligned along microfractures in transgranular trails were designated as secondary (Figs. 2 5). Some fluid inclusions from the granitoid-hosted gold deposits show necking down and leakage textures at room temperature (Fig. 2G, H). Such modified fluid inclusions can be identified by careful optical observation, and their microthermometric data were not included in the present dataset. There is no textural evidence for inclusion stretching at room temperature, and heating rates were kept low during total homogenization of fluid inclusions to avoid stretching of fluid inclusions (Roedder, 1984). Fluid inclusion types Based on the petrographic characteristics of fluid inclusions at room temperature and their microthermometric characteristics, the fluid inclusions in vein quartz from the granitoidhosted and Ashanti-type gold deposits, as well as in barren quartz veins in the Princess Town granodiorite, can be divided into the following types (Table 2, Figs. 2-5): Type 1: Primary (1P) and secondary (1S) 2- or 3-phase H 2 O-CO 2 ± NaCl fluid inclusions. CO 2 phase volume proportions (V CO2 ), estimated at 25 C, commonly show low V CO2 (10 40%) and high V CO2 (60 90%) individuals. In a given inclusion group, the type 1 fluid inclusions with low and high V CO2 values appear to be bimodal in occurrence (Figs. 2B, D; 3A, B, D, F). Hence, subtypes, defined by the V CO2, comprise (V CO2 <50%), 1P-b (V >50%), 1S-a (V CO 2 CO 2 <50%), and 1S-b (V CO2 >50%) fluid inclusions (Figs. 2A-D, F; 3). Type 2: Primary (2P) and secondary (2S) monophase liquid CO 2 -N 2 ±CH 4 fluid inclusions (Figs. 2D-E; 3D, F, H; 4A-B). These inclusions are generally dark and show no visible aqueous phase at 25 C. Type 3: Primary (3P) and secondary (3S) 2-phase aqueous H 2 O-NaCl fluid inclusions with vapor bubble volume proportions of commonly less than 10 percent (Figs. 2F; 4C-D). No /98/000/ $

6 1616 YAO ET AL. TABLE 1. Petrographic Characteristics of Fluid Inclusions in Vein Quartz from Granitoid- and Shear Zone-Hosted Gold Deposits and from Nonmineralized Granodiorite in Ghana Deposit/ Location Host rock Wall rock Mineral. Metamorphic Structure Sample no. Sample description Fluid inclusion prospect type grade style Size (main) Major type Minor type Measured Modification (µm) number Basin granitoids from the Kumasi basin Ayankyerim Obuasi Granodiorite Phyllite GT Greenschist Extension GH-AR-QV-10 Quartz vein from granitoid 5-68 (10-25), 1P-b, 2P 1S-a, 1S-b, 2S 332 Some Mod. GH-AR-QV-14 Quartz vein from granitoid, 4-42 (8-20), 3p 1P-b, 2P 127 extended to phyllite near contact zone Yamensakrom 1 Obuasi Granodiorite Phyllite GT Greenschist Extension GH-YM-QV-16 Quartz vein from granitoid, 5-62 (8-20) 1P-b, 2P 145 extended tophyllite near contact zone Nhyiaso Obuasi Tonalite, Phyllite GT Greenschist Extension GH-NS-QV-9 Quartz vein from granitoid 3-56 (6-25), 1P-b, 2P, 2S 1S-a 307 trondhjemite, GH-NS-GR-31 Quartz vein with dissemi (6-15), 1P-b, 2P 1S-a, 1S-b, 2S, 3P 312 Some Mod. monzogranite nated sulfides from granitoid Chirawewa Ayanfuri Tonalite, Phyllite GT Greenschist Extension, GH-CE-QV-19 Smoky quartz vein with dis (8-20), 1P-b, 2P 1S-a, 3P, 4P 253 East trondhjemite deformed seminated sulfides from granitoid Chirawewa Ayanfuri Tonalite, Phyllite GT Greenschist Extension, GH-CW-QV-22 Quartz vein with dissemi (6-20), 1P-b, 2P 1S-a, 3P 280 West trondhjemite deformed nated sulfides from granitoid Belt granitoids from the Ashanti (AT), Asankrangwa (AS), and Bole-Navrongo (BN) belts Mporhor 1 Mporhor Granodiorite Meta- GT Greenschist Extension? GH-MP-QV-6 Quartz vein from granitoid 5-55 (8-25), 1P-b, 2P 1S-a, 2S 203 (AT) volcanics Abore North Abore (AS) Diorite, Phyllite GT Greenschist Extension OA-PQ-13NO19 Quartz vein with disseminated 5-48 (10-30), 3P 1P-b, 2P 25 granodiorite, OA-PQ-8NO6 sulfides from granitoid 5-54 (8-20), 1S-a 3P, 3S 73 monzogranite Mpestia 1 Mpestia Granodiorite Phyllite GT Greenschist Extension GH-MT-QV-31 Quartz vein with disseminated 5-53 (8-25), 1P-b 2P, 1S-a, 1S-b 162 (AS) sulfides from granitoid Dokrupe Bole (BN) Granodiorite Phyllite GT Greenschist Extension GH-DU-QV-2 Quartz vein with disseminated 5-15 (8-10) 1S-a 18 GH-DU-QV-27 sulfides and visible gold grains 5-38 (8-20) 1P-b, 2P, 1S-a 124 from granitoid Sakapa Bole (BN) Granodiorite Phyllite GT Greenschist Extension GH-SA-QV-29 Quartz vein with disseminated 5-26 (7-15) 1P-b, 1S-a 48 sulfides from granitoid Shear zone-hosted (Ashanti-type) gold deposit from the Ashanti belt Sansu Ashanti Metavolcanics, Meta- ST Greenschist? Ductile- GH-SU-QV-2 Quartz vein with disseminated 4-22 (5-15) 2P 2S 96 Post-Mod. shear zones volcanics brittle, sulfides from shear zones, sheared zone showing lamination Barren quartz vein within Belt granitoid from the Ashanti belt Princess Town Princess Granodiorite Metavocanics NG Greenschist Extension VGR-QV-1 Quartz vein from granitoid 5-56 (8-25) 3P, 4P 3S 132 Town 1 Prospect Note: Mineral. = mineralization, GT = granitoid-hosted type, ST = shear zone-hosted type (e.g., Ashanti type), NG = nonmineralized granitoid;, -b and 1S-a, -b = primary and secondary H 2 O-CO 2 -rich ± NaCl fluid inclusions with VCO2 <50% and >50% at room temperature, respectively; 2P, 2S = primary and secondary monophase liquid CO 2 -N 2 ±CH 4 fluid inclusions, respectively; 3P, 3S = primary and secondary 2-phase H 2 O ± NaCl fluid inclusions, respectively; 4P = primary monophase H 2 O ± NaCl fluid inclusions; Mod. = modifications of fluid inclusions showing evidence of leakage and/or necking down under microscope at room temperature; Post-Mod. = postentrapment modifications of fluid inclusions in the ductile and brittle transition, respectively /98/000/ $

7 FLUID CHARACTERISTICS IN Au DEPOSITS, BIRIMIAN TERRANE, GHANA 1617 A B B 1P-b 1S-a A: GH-CW-QV-22, WV=0.6 mm B: GH-CW-QV-22, WV=0.12 mm C D 4P 2P 2P 2P 1P-b 1P-b 2P 1p-a 2P C: GH-YM-QV-16, WV=0.12 mm D: GH-NS-QV-9, WV=0.12 mm E F 2P 3P 2P 2P 4P E: GH-NS-QV-9, WV=0.12 mm F: GH-CE-QV-19, WV=0.12 mm FIG. 2. Photomicrographs of the major fluid inclusion types in vein quartz from Basin-type granitoid-hosted gold deposits and prospects in the Kumasi basin, Ghana. WV = width of view., 1P-b = primary 2- or 3-phase H 2 O-CO 2 -rich ± NaCl fluid inclusions with V CO2 <50 and >50 percent, respectively; 1S-a = secondary 2-phase H 2 O-CO 2 -rich ± NaCl fluid inclusions with V CO2 <50 percent; 2P = primary monophase liquid CO 2 -N 2 ±CH 4 fluid inclusions; 3P = primary 2-phase aqueous H 2 O ± NaCl fluid inclusions; 4P = primary monophase aqueous liquid H 2 O ± NaCl fluid inclusions. (Cont., p. 1618) /98/000/ $

8 1618 YAO ET AL. G H 2P 1P-b G: GH-AR-QV-10, WV=0.12 mm H: GH-AR-QV-10, WV=0.12 mm FIG. 2. (Cont.) Note a 2P fluid inclusion in (G) showing microfractures that may represent paths of H 2 O leakage and textures of a 1P-b fluid inclusion in (H), possibly resulting from necking down or decrepitation of fluid inclusions. Samples: AR = Ayankyerim mine at Obuasi, CE = Chirawewa East at Ayanfuri, CW = Chirawewa West at Ayanfuri, GH = Ghana, NS = Nhyiaso mine at Obuasi, QV = quartz vein, YM = Yamensakrom prospect at Obuasi. TABLE 2. Summary of Petrographic and Microthermetric Data for Fluid Inclusions from Granitoid-Hosted Gold Deposits in Ghana Type System Occurrence Size (µm) Microthermometric data H 2 O-rich-CO 2 -NaCl 3 to 62 T m(co2 ): 62.7 to 57.3 C; T m(clathrate) : 2.4 to 14.2 C; T h(co2 )(L): 11.8 to C; T h(co2 )(V): 15.6 to C; T h(l) : 161 to 379 C (mostly decrepitate); salinity: 0 to 12 (commonly 0 6) wt % NaCl equiv; ρ b : 0.62 to 1.02 g/cm 3 ; ρ c : 0.05 to 0.97 g/cm 3 ; X CO2 : to P-b H 2 O-CO 2 -rich-nacl 5 to 50 T m(co2 ): 61.6 to 57.4 C; T m(clathrate) : 3.1 to 11.2 C; T h(co2 )(L): 41.6 to C; T h(co2 )(V): 5.5 to 27.5 C; T h(v) : 161 to 387 C (mostly decrepitate); salinity: commonly 0 to 6 wt % NaCl equiv; ρ b : 0.23 to 1.11 g/cm 3 ; ρ c : 0.11 to 1.12 g/cm 3 ; X CO2 : 0.17 to S-a H 2 O-rich-CO 2 -NaCl 5 to 54 T m(co2 ): 61.0 to 57.8 C; T m(clathrate) : 8.1 to 11.2 C; T h(co2 )(L): 10.2 to C; T h(l) : 121 to 310 C (mostly decrepitate); salinity: 0 to 3.8 wt % NaCl equiv; ρ c : 0.68 to 0.98 g/cm 3 ; X CO2 : 0.01 to P CO 2 -N 2 -CH 4 3 to 56 T m : 97.1 to 57.0 C (commonly 60.0 to 58.0 C); T h(co2 )(L): 14.8 to C; T h(co2 )(V): 58.3 to C; ρ b : 0.30 to 0.92 g/cm 3 ; CO 2 *: 63 to 97 mol % 2S CO 2 -N 2 -CH 4 4 to 68 T m : 59.6 to 57.8 C; T h(co2 )(L): 36.8 to C 3P H 2 O-NaCl 4 to 30 T e : 53.2 to 16.7 C; T m(ice) : 4.8 to 0.5 C; T h(l) : 181 to 316 C (some decrepitate); salinity: 0.9 to 7.6 wt % NaCl equiv; ρ b : 0.77 to 1.03 g/cm 3 ; X H2 O: 0.98 to P H 2 O-NaCl 5 to 15 T m(ice) : 3.0 to 1.2 C; salinity: 2.0 to 5.0 wt % NaCl equiv; ρ b : 1.01 to 1.03 g/cm 3 ; X H2 O: 0.99 Notes: T m(co2 ) = melting temperature of CO 2 phase; T m(clathrate) = dissolution temperature of CO 2 -clathrate; T h(co2 )(L), (V) = homogenization temperature of CO 2 phase into the carbonic liquid and vapor phases, respectively; T h(l), (V) = total homogenization temperature into the aqueous fluid and carbonic vapor phases, respectively; T m = melting temperature of solid CO 2 phase for carbonic liquid fluid inclusions; T e = first melting temperature of eutectic phases; T m(ice) = final melting temperature of ice; ρ c = density of carbonic phase; ρ b = density of bulk fluid inclusions; X CO2, X H2 O = mole fractions of CO 2 and H 2 O, respectively; CO 2 * = CO 2 in mol % estimated from Raman analyses of 2P fluid inclusions; 4P fluid inclusions are rarely found in the samples studied from granitoid-hosted gold deposits /98/000/ $

9 FLUID CHARACTERISTICS IN Au DEPOSITS, BIRIMIAN TERRANE, GHANA 1619 A B 1P-b 1P-b 1P-b A: GH-MP-QV-6, WV=0.12 mm B: GH-MT-QV-31, WV=0.12 mm C D 1S-b 1S 2S 1S-b 2S C: OA-PQ-13No19, WV=0.6 mm D: OA-PQ-13No19, WV=0.12 mm E F 2P 2P F 1P-b E: OA-PQ-14No4, WV=0.6 mm F: OA-PQ-14No4, WV=0.12 mm FIG. 3. Photomicrographs of the major fluid inclusion types in vein quartz from Belt-type granitoid-hosted gold deposits and prospects in the Birimian of Ghana. 1S-b = secondary 2-phase H 2 O-CO 2 -rich ± NaCl fluid inclusions with V CO2 >50 percent; 2S = secondary monophase liquid CO 2 -N 2 ±CH 4 fluid inclusions; MP = Mporhor prospect in the Ashanti belt. MT = Mpesetia prospect in the Asankrangwa belt; OA-PQ-13No19, OA-PQ-14No4 = borehole samples from the Abore North prospect in the Asankrangwa belt; DU = Dokrupe mine in the Bole-Navrongo belt; SA = Sakpa mine in the Bole-Navrongo belt. Abbreviations as in Figure 2. (Cont., p. 10) /98/000/ $

10 1620 YAO ET AL. G H 2P 2P G: GH-DU-QV-27a, WV=0.12 mm H: GH-SA-QV-29, WV=0.12 mm FIG. 3. (Cont.) A B 2P 2S B 2P 2P A: GH-SU-QV-2, WV=0.8 mm B: GH-SU-QV-2, WV=0.08 mm C D 3P 4P 4P 4P 3P 4P 3P 3P C: GH-VGR-QV-1, WV=0.12 mm D: GH-VGR-QV-1, WV=0.12 mm FIG. 4. Photomicrographs of the major fluid inclusion types in vein quartz from the Sansu (SU) mine at Ashanti and in barren vein quartz from the nonmineralized granodiorite at Princess Town in the Birimian of Ghana. VGR = Belt-type granitoid in meta-volcanic rocks at Princess Town. Abbreviations as in Figures 2 and /98/000/ $

11 FLUID CHARACTERISTICS IN Au DEPOSITS, BIRIMIAN TERRANE, GHANA 1621 CO 2 phase can be detected microthermometrically, although a very small amount of CO 2 may be present in certain of the fluid inclusions (Rosso and Bodnar, 1995). Type 4: Primary (4P) monophase aqueous liquid H 2 O-NaCl fluid inclusions with no vapor bubbles observed at 25 C (Figs. 2D, F; 4C-D). These inclusions are rarely found in samples from the gold deposits studied, and they may represent entrapment of metastable stretched water solutions (Roedder, 1984) and/or necked type 3 fluid inclusions. Types 1 and 2 constitute the majority of fluid inclusions in vein quartz within mineralized granitoids, whereas types 3 and 4 were mainly observed to occur as individual groups in barren vein quartz within the nonmineralized Princess Town granodiorite. By contrast, fluid inclusions in vein quartz from the Sansu mine are dominated by type 2. Primary fluid inclusions in all the studied samples commonly occur as randomly scattered groups comprising both type 1 and 2 inclusions (Figs. 2A-F; 3A-B, E-H; 5), although, in a few cases, type 1 or type 2 inclusions are also present in isolation (Fig. 3G). Type 1 fluid inclusions typically show a wide range of CO 2 volume proportions (V CO2 ), although some and 1P-b fluid inclusions display relatively constant V CO2 (10 30% for the ; 70 90% for the 1P-b; Fig. 5). The fluid inclusions are both regular and irregular in shape, and some show negative crystal forms (Figs. 2 5). The sizes of primary fluid inclusions range from about 3 to 62 µm (commonly 8 25 µm; Tables 1, 2). Secondary fluid inclusions of types 1 and 2 occur as small aggregates (Fig. 3C-D). They are commonly less than 15 µm in size, but some of them are as large as 68 µm (Tables 1, 2). Crosscutting relationships among primary fluid inclusion types were not observed in the 16 samples studied, and the chronological sequence of primary fluid inclusion types cannot be recognized. As shown in Figures 2 to 5, fluid inclusions of types 1P and 2P as well as 3P and/or 4P coexist in individual arrays and groups, implying that the different compositional types of primary fluid inclusions were probably coeval. Similarly, different types of secondary fluid inclusions are also observed to coexist in trails (Fig. 3C, D), FIG. 5. Microscopic sketch map showing distributions of 1P, 2P, and 4P fluid inclusions in individual inclusion groups of representative chips from the Nhyiaso and Yamensakrom gold deposits at Obuasi. Microthermometric results of numbered fluid inclusions are listed in Table 3. Abbreviations as in Figure /98/000/ $

12 1622 YAO ET AL. suggesting that they were trapped simultaneously or in a short time interval during mineralization. Microthermometric data A summary of microthermometric data for major fluid inclusion types from the granitoid-hosted gold deposits is presented in Table 2. Microthermometric results for representative fluid inclusion groups from the selected samples are shown in Table 3, and the detailed microthermometric data for fluid inclusions measured in the present study are presented in Tables 4 to 5 and plotted in Figures 6 to 12. Granitoid-hosted gold deposits: Fluid inclusions in individual samples mainly comprise types 1 and 2, with only minor types 3 and 4 present. Type 1P fluid inclusions are present in all the studied samples and show various CO 2 volume proportions (V CO2 ), estimated at 25 C, with X CO2 values ranging from to 0.21 for the and from 0.17 to 0.87 for the 1P-b. Samples from the Yamensakrom and Sakpa gold deposits, however, principally display fluid inclusions with X CO2 values of 0.02 to 0.20 and 0.01 to 0.19, respectively. Melting temperatures of the CO 2 phase (T m(co2 )) are between 62.7 and 57.3 C for the, and 61.6 and 57.4 C for the 1P-b (Table 2, 4, Fig. 6A, G). These temperatures are about 0.7 to 6 C lower than the pure CO 2 melting point ( 56.6 C), indicating the presence of additional volatile species such as N 2 and CH 4 (see Raman results). Melting temperatures of CO 2 clathrate (T m(clathrate) ) are between 2.4 and 14.2 C for the and between 3.1 and 11.2 C for the TABLE 3. Microthermometric Results for Primary Fluid Inclusions in Quartz Grains from Representative Chips at Obuasi FI no. Chip Size FI type Phase V CO2 T m T mco2 T m(ice) T m(clathrate) T h(co2)(l) T h(co2)(v) T h(l) T h(v) T d (µm) (%) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) Nhyiaso (GH-NS-QV-9) 20 B1-a B1-a B1-a B1-a B1-a B1-a B1-a B1-a n.m. n.m. 30 B1-a B1-a B1-a B1-a B1-a B1-a B1-a 17 1P-b B1-a 15 1P-b B1-a 7 1P-b B1-a 8 1P-b B1-a 7 1P-b Nhyiaso (GH-NS-QV-9) 319 B2-c n.m. n.m. n.m. 322 B2-c n.m B2-c B2-c B2-c B2-c B2-c B2-c B2-c 10 1P-b B2-c 12 1P-b B2-c 13 1P-b B2-c 16 2P B2-c 25 2P B2-c 10 2P B2-c 6 2P n.m. n.m. 320 B2-c 8 2P n.m. n.m. 321 B2-c 12 2P B2-c 13 2P B2-c 14 2P B2-c 20 2P n.o. n.o. 330 B2-c 11 2P B2-c 15 2P n.o. n.o /98/000/ $

13 FLUID CHARACTERISTICS IN Au DEPOSITS, BIRIMIAN TERRANE, GHANA P-b. The majority of T m(clathrate) values for the 1P fluid inclusions, however, fall between 7.0 and 11.0 C (Table 2, 4, Fig. 6B, H). From these results, the estimated salinities fall in a range from 0 to 12 wt percent NaCl equiv, clustering at 0 to 6 wt percent NaCl equiv for the 1P fluid inclusions. Average salinities of 1P inclusions from individual deposits are 3.4 ± 1.6 (1σ, n = 113) at Mporhor, 2.2 ± 1.1 (1σ, n = 300) at Ayanfuri, 1.7 ± 0.9 (1σ, n = 259) at Nhyiaso, 2.5 ± 1.0 (1σ, n = 283) at Ayankyerim, 0.03 ± 0.1 (1σ, n = 130) at Mpesetia, and 2.6 ± 1.1 (1σ, n = 104) wt percent NaCl equiv at Dokrupe. The T m(clathrate) values higher than that of the pure CO 2 clathrate (10 C) may result from a mixture of carbon dioxide and methane clathrates (CO H 2 O; CH 4 7H 2 O) and suggest relatively low salinities (cf. Hagemann et al., 1992). The presence of additional gases in the carbonic phase indicates that the salinities estimated from the T m(clathrate) values represent minima (Diamond, 1994). The carbonic portion of most of the and 1P-b fluid inclusions homogenized into the liquid phase, and their T h(co2 )(L) values span a range from 11.8 to C for the and from 41.6 to C for the 1Pb (Fig. 6C, I, one value of 41.6 C in Fig. 6I not shown), respectively. In some cases, however, the carbonic part of 1P inclusions homogenized into the vapor phase, with T h(co2 )(V) values in the range from 15.6 to C for the (n = 36, Fig. 6D) and from 5.5 to 27.5 C for the 1P-b (n = 22, Fig. 6J). Estimated CO 2 densities cover a wide range from 0.05 to 0.97 g/cm 3 for the and from 0.11 to 1.12 g/cm 3 for the 1P-b, respectively. More than 90 percent of the CO 2 densities TABLE 3. (Cont.) FI no. Chip Size FI type Phase V CO2 T m T mco2 T m(ice) T m(clathrate) T h(co2)(l) T h(co2)(v) T h(l) T h(v) T d (µm) (%) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) 333 B2-c 13 2P B2-c 32 2P B2-c 7 2P n.m. n.m. 338 B2-c 7 2P B2-c 12 2P B2-c 13 4P 2.5 C: Yamensakrom(GH-YM-QV-16) 1 A1-a A1-a A1-a A1-a A1-a A1-a n.m A1-a n.m A1-a A1-a n.m A1-a n.m A1-a A1-a A1-a A1-a A1-a n.m A1-a n.m A1-a n.m A1-a A1-a n.m. n.m. n.m. 20 A1-a A1-a n.m A1-a A1-a A1-a A1-a A1-a A1-a A1-a A1-a A1-a A1-a A1-a Notes: FI = fluid inclusion; and 1P-b = primary 2- or 3-phase H 2 O-CO 2 ± NaCl fluid inclusions with V CO2 <50 and >50%, respectively; 2P = primary monophase liquid CO 2 -N 2 ±CH 4 fluid inclusions without a visible aqueous phase at room temperature; 4P = monophase H 2 O ± NaCl fluid inclusions; V CO2 = carbonic volume percent estimated at room temperature; T m = final melting temperature of solid CO 2 phase for 2P fluid inclusions; T m(co2) = melting temperature of CO 2 phase; T m(clathrate) = dissolution temperature of CO 2 clathrate; T m(ice) = final melting temperature of ice; T h(co2)(l), (V) = homogenization temperature of CO 2 phase into the carbonic liquid and vapor phases, respectively; T h(l), (V) = total homogenization temperature into the aqueous fluid and carbonic vapor phases, respectively; T d = decrepitation temperature; n.m. = not measured; n.o. = not observed /98/000/ $

14 1624 YAO ET AL. TABLE 4. Summary of Microthermometric and Raman Spectrometric Data for Fluid Inclusions from Mineralized Basin Granitoids at Obuasi and Ayanfuri in the Kumasi Basin Salinity Deposit/prospect FI type Issue V CO2 /V b T i T mco2 / T e T m(ice) T m(clathrate) T hco2(l) T hco2(v) T h(l) T h(v) T d (wt % ρ CO2 ρ b X H2O XNaCl X CO2 CO 2 * N 2 * CH 4 * (sample) (%) T m (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) (ºC) NaCl) (g/cm 3 ) (g/cm 3 ) (mol %) (mol %) (mol %) Nhyiaso, Min Obuasi Max (GH-NS-GR-31 + N GH-NS-QV-9) 1P-b Min Max N P Min Max N S-a Min Max N S-b Min Max N S Min Max N P RQ Min Max N RQ Min Max N Yamensakrom,* Min Obuasi Max (GH-YM-QV-16) N P-b Min Max N P Min Max N 1 1 Ayankyerim, Min Obuasi Max (GH-AR-QV-10 + N GH-AR-QV-14) 1P-b Min Max N P Min Max N S-a Min Max N Ayankyerim, 1S-b Min Obuasi Max (GH-AR-QV-10 + N GH-AR-QV-14) 2S Min Max N /98/000/ $