Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The fluid evolution in metamorphic terrains may be recorded by fluid inclusions and mineral chemical equilibria. Several studies have shown that scapolite is a mineral indicator of chlorine (Cl) concentration in coexisting fluid (e.g., Orville 1975; Ellis 1978; Vanko and Bishop 1982; Oterdoom and Wenk 1983; Mora and Valley 1989; Moecher and Essene 1991; Harley et al. 1994). Unlike other chlorine-bearing minerals, such as those in the mica and amphibole groups, scapolite contains little OH; as such, its Cl content at a given pressure and temperature conditions reflects the NaCl activity in the fluid, independent of the ƒH2O, thereby making scapolite an ideal indicator of Cl composition in the fluid (Ellis 1978). Scapolite-group minerals are generally considered as the salt-bearing analogues of (Na, Ca)-feldspars, which are stable under conditions of high volatile activity. Compositions within the scapolite group are represented generally as 3(Ab, An, Ksp).(CaCO3, NaCl, CaSO4), where Ab, An, and Ksp refer to the components NaAlSi3O8, CaAl2Si2O8, and KAlSi3O8, respectively (Rebbert and Rice 1997). The end-member formulae were originally derived by analogue with plagioclase and may be written as marialite (Ma) = 3NaAlSi3O8. NaCl (or Ab3. NaCl) and meionite (Me) = 3CaAl2Si2O8. CaCO3 (An3. CaCO3) (Evans et al. 1969; Teertstra and Sherriff 1997). Beside Silvialite (Si) = 3CaAl2Si2O8. CaSO4 (An3. CaCO4) Marialite and meionite are the only two currently accepted species (Bayliss 1987).

Scapolite is a common metamorphic mineral in the Lufilian–Zambezi belt. It is widespread in most marbles and calc-silicate lithologies in southern Zambia (Munyanyiwa 1985) and is also widespread in carbonate and calc-silicate rocks in central Zambia (Simpson et al. 1965; Drysdall and Stillman 1966; Katongo 1999), but rare in granitic gneisses and pelitic schists elsewhere in the belt. Small metagabbroic bodies intrusive into the supracrustal sequence in central and southern Zambia have abundant scapolite (Simpson and Stillman 1963; Vrána et al. 1975; John et al. 2003). Although rare in most granitic rocks, scapolite has been reported in granite gneisses in southern Zambia (Wilson et al. 1993; Hanson et al. 1994). In the Copperbelt region, scapolite is abundant in the Katangan metasediments, and metagabbros (Mendelsohn 1961; Cosi et al. 1992; Tembo et al. 1999).

In spite of the prevalence of scapolite in the Neoproterozoic Lufilian–Zambezi belt, good scapolite compositional data that could be used independently in regional evaluation of scapolitization are limited. Simpson et al. (1965) determined the composition of scapolite as “dipyre” (Na3Cal4Si8O24Cl) from physical properties such as specific gravity, density, refractive index etc., whereas Drysdall and Stillman (1966) found the composition of scapolite as sodic-meionite (Me55) by wet-chemical analyses. Munyanyiwa (1990) reported Cl-free scapolite (Me60–65) in calc-silicates from southern Zambia, but did not present mineral data or statement of how this mineral composition was determined. Similarly Cosi et al. (1992), Hanson et al. (1994), and Wilson et al. (1993) did not provide compositional data of scapolites from their study areas in the belt. The methods used in determining compositions of scapolite in some of these studies are grossly inaccurate and cast doubts on the reliability of the claimed compositions of scapolite. The only previous reliable scapolite compositional data (calcic marialite-Me25–38) in the belt are those that were obtained by electron microprobe analysis (Tembo 1994).

The origin of scapolite in most metamorphic terrains is generally attributed to evaporite precursors. In the Lufilian–Zambezi belt, in spite of limited scapolite compositional constraints, NaCl-rich fluids are generally considered by most authors to have been derived from evaporite beds that were present in metasedimentary successions.

In this paper, we present the first scapolite compositional data from the Munali hills granite gneiss, amphibolites, and metagabbros from the Munali hills area, in the southern part of the Lufilian–Zambezi belt. We also attempt to evaluate the source and composition of fluids that were associated with metamorphism in the belt. We have integrated compositional data of scapolite from the Copperbelt region in the northern part of the belt (Tembo 1994) with our data to assess the regional extent of the source of fluids, which caused extensive scapolitization in the Lufilian–Zambezi belt.

2 Regional Geological Framework of the Lufilian–Zambezi Belt

Until recently, the Lufilian arc and the Zambezi belt (Fig. 18.1) were believed to form discrete orogenic belts separated by the Mwembeshi dislocation zone (MDZ), on the basis of presumed marked differences in metamorphic grade, structural vergence, and orogenic histories (De Swardt and Drysdall 1964; Unrug 1983; Coward and Daly 1984; Hanson et al. 1993). The MDZ is a prominent crustal-scale ENE trending shear zone across central Zambia, which is believed to extend to the northeast into Malawi and to the southwest into Namibia (De Swardt and Drysdall 1964; Coward and Daly 1984; Daly 1986; Porada 1989; Kampunzu and Cailteux 1999). Isotopic age data of key lithotectonic units, coupled with re-interpretations of the geology in the belts in Zambia and Zimbabwe, have shown that the two belts are coeval and form part of a network of Neoproterozoic-early Palaeozoic orogenic belts in central-southern Africa (e.g., Hanson et al. 1993; Dirks and Sithole 1999; Kampunzu and Cailteux 1999; Vinyu et al. 1999; Porada and Berhorst 2000; De Waele et al. 2009). The transcontinental network of belts separates the Congo and Kalahari cratons (Coward and Daly 1984; Unrug 1983; Johnson et al. 2007; De Waele et al. 2009) and formed during the reconstruction of Rodinia (Unrug 1996; Weil et al. 1998). The disposition and composition of metasedimentary rocks in the belt are regarded to indicate deposition in a rift basin (Wilson et al. 1993; Hanson et al. 1994; Porada and Berhorst 2000).

Fig. 18.1
figure 1_18

Simplified geological map of Zambia and parts of neighbouring countries. Symbols; *: sample locations; MHG: Munali hills granite gneiss and metagabbro and amphibolitic dykes; NG: Nchanga granite; LG: Lusaka granite. Reference towns indicated by ■; Kb: Kabwe; Ls: Lusaka and Kf: Kafue (After Porada and Berhorst 2000)

Full size image

A detailed review and synthesis of the evolutionary history for much of the Lufilian–Zambezi belt was provided by Porada and Berhorst (2000). Key et al. (2001) presented the most recent comprehensive interpretation of tectono-thermal evolution of the Lufilian arc based on their work in NW Zambia. An overview of the structure and geology of the Zambezi belt was given by Hanson et al. (1994). Here, we summarize previous work on the Lufilian arc and Zambezi belt to provide a regional geological framework for this study.

The Lufilian arc is an arcuate, northward convex-shaped geological structure covering eastern Angola, the southern Democratic Republic of Congo (DRC), northwestern Zambia, and is well known for its vast world-class copper–cobalt deposits (Fig. 18.1). In the Copperbelt region, which straddles northwestern Zambia and southern DRC, the stratigraphy is well known due to exploration and mining activities. The succession comprises basement and metasedimentary rocks of the Neoproterozoic Katanga Supergroup. In the Zambian Copperbelt region, the basement is composed of Lufubu schists and gneisses (Mendelsohn 1961) that are intruded by Eburnian (ca. 2,200–1,800 Ma) granites (Cahen et al. 1984). Key et al. (2001) dated basement rocks of Archaean age (ca. 2.6 Ga) in the western arm of the Lufilian arc. Overlying the Lufubu schists and gneisses, is a sequence of schists and quartzites of the Muva Supergroup, which are in turn intruded by early Neoproterozoic granites (e.g., Nchanga granite). The revised lithostratigraphy of the Katanga Supergroup (Porada and Berhorst 2000) consists, in stratigraphic succession, of arenaceous Roan Group, carbonaceous Mwashia Group with metavolcanic interbeds, and conglomeratic Kundelungu Group. It has long been observed by many authors that there are several horizons of evaporites in the Katangan succession (e.g., Mendelsohn 1961; Binda 1994; Cailteux et al. 1994). The former existence of evaporites in the Katangan metasediments is indicated by sabkha facies, crystals and pseudomorphs of gypsum and anhydrite, stratigraphic gaps underlain by collapse breccias, chloride inclusions in ores, and saline springs (Porada and Berhorst 2000; Jackson et al. 2003). Evaporites have played an important role in deformation (Cailteux and Kampunzu 1995) and formation of megabreccias (olistostromes)(Wendorff 2000) in the Katanga succession. In the Damara belt, the southwestern extension of the Lufilian arc, former evaporite deposits are characterized by high sodium contents, abundant albite pseudomorphs after evaporite minerals and concordant collapse breccia (Porada and Berr 1988).

Metamorphism in the Lufilian arc increases gradually from prehnite–pumpellyite facies in the northern boundary of the Lufilian arc, up to amphibolite facies in the southern part of the Copperbelt region (Kampunzu and Cailteux 1999), but higher grades of up to eclogite facies have been locally recorded (Vrána et al. 1975; Cosi et al. 1992). From studies of mineral associations and phengite compositions in the ore-shale, in the Copperbelt region of Zambia, Moine et al. (1986) estimated the metamorphic conditions of 420–550°C and 2–6 kbar in the Copperbelt region. Deformation in the Lufilian arc mainly involved thin-skinned tectonics with northerly directed thrusting (Coward and Daly 1984; Daly 1986; Kampunzu and Cailteux 1999; Porada and Berhorst 2000; Key et al. 2001).

The east-west trending Zambezi belt stretches from central Zambia into northern Zimbabwe, where it merges with the north–south trending Mozambique belt (Fig. 18.1). In Zambia, the Zambezi belt consists of wide zones of remobilized Mesoproterozoic to Neoproterozoic crystalline basement, unconformably overlain by late Neoproterozoic supracrustal rocks (Hanson et al. 1988; Wilson et al. 1993). The Basement is composed of the Mpande gneiss, a megacrystic, K-feldspar, biotite augen gneiss and the Ngoma gneiss, a protomylonitic, K-feldspar gneiss. The two gneissic units form an extensive terrain in the central parts of the belt. The Munali hills granite gneiss is a small lensoidal body that is intrusive into the Mpande gneiss, and adjacent metasedimentary rocks (Smith 1963) Katongo et al. (2004) report an U–Pb zircon age of 1,090.1 ±1.3 Ma. The supracrustal sequence starts with a thick pile of rift-related bimodal volcanics, which are overlain by a thick sequence of psammites and pelites that are in turn succeeded by an extensive unit of marbles and calc-silicates rocks (De Swardt and Drysdall 1964; Wilson et al. 1993; Hanson et al. 1994). The supracrustal sequence in the Zambezi belt in Zambia has been correlated based on structural continuity and broad lithological similarities with the Makuti and Rushinga Groups in Zimbabwe (Broderick 1981; Barton et al. 1991). Evaporite horizons have not really been observed in the Zambezi belt, but former evaporites are indicated by relict chicken-wire structures, reflecting replacement of original anhydrite nodules by calcite in marbles at Nampundwe Mine, in central-southern Zambia (Burnard et al. 1993). Metamorphism in the Zambezi belt was primarily amphibolite facies (Barton et al. 1991; Hanson et al. 1994; Hargrove et al. 2003), but locally there are occurrences of tectonically exhumed high-pressure rocks comprising eclogites and talc-kyanite whiteschists in Zambia and northwest Zimbabwe (Vrána et al. 1975; Dirks and Sithole 1999; Johnson and Oliver 2000). Diopside-plagioclase assemblages are uniformly developed in the calc-silicate rocks and carbonate formations in southern Zambia. Garnet-biotite thermometry and plagioclase-hornblende thermobarometry on amphibolite layers in the Carbonate Formations indicate peak metamorphic conditions of 560–600°C and 5–6 kbars (Munyanyiwa and Hanson 1988; Munyanyiwa 1990). These P-T conditions are consistent with widespread occurrence of kyanite in the supracrustal sequence south of Kafue. Temperatures between 565°C and 580°C and pressures ranging from 6 to 8 kbars were indicated by the garnet–hornblende thermobarometry on garnet amphibolites in the Chisamba area north of Lusaka (Katongo 1999). Occurrences of several small bodies of eclogite (Vrána et al. 1975; Vrána 1978), and talc-kyanite-whiteschists (Vrána and Barr 1972; Cosi et al. 1992; John et al. 2003) indicate that local medium to high pressure were attained during metamorphism in most parts of the Zambezi belt. Scapolite metamorphism also affected metagabbros, and eclogites in the southeastern extension of Zambezi belt in northern Zimbabwe (Dirks and Sithole 1999; Munyanyiwa et al. 1997). In contrast, to the Lufilian arc, deformation in the Zambezi belt is characterised by thick-skinned tectonics, involving both supracrustal and basement rocks (Coward and Daly 1984).

3 Field Relations and Petrography of Scapolite Bearing Rocks

In this study, scapolite was identified in the Munali hills granite gneiss, amphibolites, and metagabbros in the Munali hills area (Fig. 18.1). Scapolite occurs as pervasive replacement of plagioclase in all the studied lithologies, as shown in microphotos (Fig. 18.2) and field occurrences (Fig. 18.3).

Fig. 18.2
figure 2_18

(a) Photomicrograph showing characteristic partial replacement of plagioclase by scapolite in sample MHG2. (b) Photomicrograph showing of coronal intergrowth of scapolite and quartz in sample MHG9. Note: photograph was taken with quartz in partial extinction to enhance contrast against scapolite. (c) Back-scattered electron (BSE) image of coarse porphyroblastic scapolite with inclusions of quartz, biotite, and amphibole. (d) NW trending veinlet in central part of BSE image of relatively coarse-grained intergrowth of scapolite and biotite. Scap: scapolite; Biot: Biotite; Amp: amphibole; qtz: quartz; Plag: plagioclase: Kfsp: K-feldspar

Full size image
Fig. 18.3
figure 3_18

Field pictures of the scapolite-bearing rocks in the Kafue area. (a) Leukocratic, medium- to fine-grained variety of the Munali Hill granite, dark spots are biotite-rich areas. (b) Deformed mafic dyke (now amphibolite) intruding into a medium-grained Munali Hill granite

Full size image

3.1 Munali Hills Granite Gneiss

The Munali hills granite gneiss forms a small part of the Munali hills, a WNW trending elongate ridge, underlain mainly by granitic gneiss of the Mpande Formation (Smith 1963; Mallick 1966). The southern boundary of the granite gneiss is marked by an alternating sequence of kyanite schist, quartzite, calc-silicates, and marbles. The Munali hills granite gneiss is a porphyritic K-feldspar-rich granite gneiss composed of a range of fine- to coarse-grained varieties, varying from leucocratic to melanocratic types and displaying variable degrees of deformation. The granite gneiss is composed of pink microcline porphyroclasts set in fine-medium grained felsic groundmass. Isolated, dark, medium-grained irregularly shaped, biotite-rich enclaves are common. Deformation is generally weak, and exhibited by a crude foliation defined by aligned biotite warping around slightly elongated porphyroclasts of feldspar and quartz aggregates. The gneisses are composed of microcline, quartz, plagioclase/scapolite, biotite and calcite ± muscovite. The microcline crystals are generally poikiloblastic, hosting smaller grains of other mineral constituents, non-perthitic and slightly kaolinized. Accessory minerals include sphene, epidote, apatite, zircon, and ±tourmaline.

The granite displays varying degrees of alteration, ranging from chloritization of biotite, sericitization and ubiquitous scapolitization of plagioclase. Scapolite constitutes up to 30 modal % and invariably intergrown with plagioclase. The replacement of plagioclase by scapolite is mainly along grain margins, cleavage planes and less commonly in the cores (Fig. 18.2a). In some grains, scapolite hosts fine inclusions of twinned plagioclase, sericite, epidote, and carbonate, probably indicating replacement after slight retrogression of plagioclase. In a few porphyroblasts of scapolite, calcite occurs along cleavage or fractures planes. A few grains of scapolite display symplectite-like intergrowths with quartz (Fig. 18.2b). Undulose extinction in quartz, kink bands in biotite and deformation twins in plagioclase indicate low temperature deformation.

3.2 Amphibolites

The Munali hills granite gneiss is intruded by several steep, WNW trending amphibolitic dykes, measuring up to 5 m in width. At outcrop scale, the amphibolitic dykes exhibit a strong shear fabric, suggesting emplacement in extensional shear fractures. The amphibolites are dark green, and fine-grained, with white plagioclase phenocrysts and scapolite porphyroblasts, and are composed mainly of blue green amphibole and subordinate biotite, plagioclase, scapolite and minor calcite. Plagioclase is partly retrogressed to sericite. The scapolite is up to 40 modal % and partially or completely replaces plagioclase. Scapolite in the amphibolite displays various textures ranging from large porphyroblasts to fine grains and coarse-grained veinlets (Fig. 18.2c, d). The veinlets are composed of coarse-grained scapolite and biotite in stable equilibrium. Most scapolite crystals host inclusions of plagioclase, calcite, and epidote. Accessory minerals include epidote, sphene, carbonates, and opaque minerals. The opaque minerals are chalcopyrite and pyrite.

3.3 Metagabbros

3.3.1 Munali Hills Area

The metagabbro is one of the varieties of a metagabbroic composite body that is associated with nickel sulphide mineralization. The sulphide mineralization consists mainly of pyrrhotite with lesser amounts of pyrite, pentlandite, ­chalcopyrite, and accessory magnetite. The metagabbro is dark green, medium- to coarse-grained and consists mainly of amphiboles, augite, plagioclase, and biotite, which have undergone various degrees of alteration. Plagioclase (20–30 modal %) is invariably partly replaced by scapolite. Other secondary minerals include chlorite, tremolite, epidote, sericite and zoisite.

3.3.2 Copperbelt Region

Detailed descriptions of metagabbros in the Copperbelt region are provided by Tembo et al. (1999). The metagabbros are composed of calcic amphiboles, sodic plagioclase, and variable amounts of scapolite and have preserved primary igneous textures. The metagabbros are invariably affected by hydrothermal alteration. As in other scapolite bearing rocks in the Munali hills area, scapolite in these metagabbros occurs as a pervasive replacement of plagioclase. Textural relationships among the minerals in the metagabbros indicate that scapolitization post-dated amphibolitization (Tembo 1994).

4 Analytical Procedures

Mineral analyses on scapolite, and associated minerals such as plagioclase, amphiboles, and biotite from a few selected samples, were conducted on a four-spectrometer CAMECA SX100 microprobe at the University of Vienna, Department of Geological Sciences. Counts were obtained simultaneously from four spectrometers, using a 15 keV accelerating voltage, a 1–2 μm beam diameter, and a beam current of 20 nA. Natural and synthetic standards were used for calibrations and the PAP correction (Pouchou and Pichoir 1991) was applied to the data.

Mineral compositional data of scapolite by Tembo (1994) from the Copperbelt region were obtained on a scanning electron microscope (SEM)-Cambridge Stereoscan 250 MK3-equipped with an energy dispersive (ED) analytical system, at the University of Göttingen, Germany. Operating conditions for the quantitative analyses were 15 keV accelerating potential, 20 nA probe current and 100 s counting time. Natural and synthetic standards were used to calibrate the analyses. Matrix effects on elements and their oxides weight percentages were corrected for and calculated using the Cambridge online program MACRO ZAF. Representative results of mineral analyses are presented in Tables 18.1–18.5.

Table 18.1 Representative mineral compositions of scapolite from the Munali hills area
Full size table
Table 18.2 Representative mineral composition of scapolite from metagabbros in the Copperbelt region, Zambia (After Tembo 1994)
Full size table
Table 18.3 Representative mineral compositions of plagioclase from the Munali hills area
Full size table
Table 18.4 Representative mineral compositions of amphiboles in amphibolites from the Munali hills area
Full size table
Table 18.5 Representative mineral compositions of biotite from the Munali hill area
Full size table

5 Results

5.1 Scapolite

Scapolite compositions are commonly reported in terms of equivalent anorthite content (EqAn = 100 * (Al − 3)/3), where Al is calculated on the basis of tetrahedral 16 cations (Evans et al. 1969; Teertstra and Sherriff 1997), mole fractions of ­chlorine (XCl = Cl atoms per formula unit (a.p.f.u)), CO2 (XCO2 = 1 – Cl – S a.p.f.u), and meionite composition­ (Me % = Ca/(Ca + Na + K)). Scapolite from the Munali hills area has a range of compositions from 31–46 EqAn % and 0.37 to 0.50 XCl. The scapolites are sodium-rich with 27–47 Me % compositions that fall within calcian marialite scapolite compositional group (15< %Me < 50, Fig. 18.4a; Rebbert and Rice 1997). The highest XCl contents are recorded in the Munali hills granite gneiss sample (MHG10). The compositions of scapolite plot close mid-way between the ideal end-members marialite and meionite, indicating compositions represented by solid solutions between these two end-members and a possible silvialite component Ca4[Al6Si6O24]SO4 (Fig. 18.4b).

Fig. 18.4
figure 4_18

(a) Classification of scapolites from the Lufilian–Zambezi belt on the basis of meionite component: Me% = 100 * (Ca/Ca + Na + K). All the scapolite data points plot in the calcian marialite group (After Rebbert and Rice 1997). (b) Plot of XCl versus EqAn(scap) and XAn(plag). Scapolite data points from the Munali hills area plot between the three ideal end member tie lines, whereas some data points from the Copperbelt region lie above the ideal compositions. Symbols: ♦MHG10; □ MHG2; +- MHG5a; ▲ MHG5c; ×- MHG9; ∆- NDG; CB- metagabbros; ○ Plag coexisting with scapolite; ● LG-plag; ■- NG-plag

Full size image

Mineral compositions of scapolite in metagabbros from the Copperbelt region (Tembo 1994) have a restricted range between 29 and 38 EqAn % comparable to the compositions of scapolite from the Munali hills area. Compared to those of the Munali hills area, the scapolites from the Copperbelt region have higher XCl values ranging from 0.61 to 0.84 and data points lie on and above the tie line between the marialite-meionite ideal end-member compositions.

5.2 Plagioclase

Plagioclase coexists with scapolite in all samples. The mineral formula was calculated on the basis of 32 oxygens. The plagioclase has a wide compositional range from 13 to 61 anorthite content with lower contents (An23–37–oligoclase-andesine) in granite gneisses and higher contents (An36–61 –andesine-labradorite) mainly in the amphibolite samples (Fig. 18.4b, Table 18.3). The anorthite contents of plagioclase overlap with coexisting scapolite (Fig. 18.4b). Plagioclase from the non-scapolite bearing granites in the belt, e.g., Lusaka and Nchanga granites, are more sodium-rich with a compositional range of An4–16.

5.3 Amphiboles

The amphiboles are Cl-bearing ferroan paragasite and magnesian hastingsite. Chlorine contents in pargasite are rather uniform at about 0.26 XCl, whereas that in hastingsite are slightly higher, ranging from 0.34 to 0.40 XCl. Similarly, XMg values range from 0.32 to 0.34 and 0.41 to 0.42 in pargasite and hastingsite, respectively.

5.4 Biotite

Biotite mineral formulae were calculated on the basis of 22 oxygens. Biotite was classified into two groups on the basis of Al(IV) a.p.f.u contents and XFe (Fig. 18.5a). Most biotites have high annite components, except some from one amphibolite sample (MHG5a), which are rich in the phlogopite component. On the XCl-XMg diagram (Fig. 18.5b), the biotite (annite-rich) from the metagabbros have the highest­ Cl contents ranging from 0.6 to 0.9 wt%, whereas biotites from amphibolites­ have the highest XMg values. Biotites from non-scapolite bearing rocks such as Lusaka granite also have compositions and Cl contents similar to those coexisting in scapolite-bearing rocks.

Fig. 18.5
figure 5_18

(a) Classification of biotite from scapolite and non-scapolite bearing rocks from the Lufilian–Zambezi belt. Most biotites plot in the annite–siderophyllite field, except some from the amphibolitic dyke sample MHG5a. (b) Plot of XCl versus XMg of biotites from scapolite and non-scapolite bearing rocks from the Lufilian–Zambezi belt. Metagabbros have biotites with the highest Cl contents, whereas some from the amphibolitic dykes have high Mg (the phlogopite-rich biotites in (a)). Symbols: ♦ Munali hills granite gneiss; ▲ Amphibolite; ∆ Metagabbro; □ Lusaka granite

Full size image

6 Discussion

6.1 Scapolite and Fluid Compositions

The NaCl content of a fluid can be estimated from scapolite coexisting with calcite according to the reaction (Ellis 1978):

$${\left({\rm NaCl}\right)}_{{\rm scp}}+{\left({{\rm CaCO}}_{3}\right)}_{{\rm cc}}={\left({{\rm CaCO}}_{3}\right)}_{{\rm scp}}+{\left({\rm NaCl}\right)}_{{\rm fluid}}$$

Ellis (1978) investigated this fluid exchange equilibrium at 750°C and 4 kbar, consistent­ with metamorphic conditions in the granulite facies. Mora and Valley (1989) outlined the limitations underlying Ellis’s expression that are mainly because of thermodynamic mixing properties and inappropriate P–T conditions done by Ellis (1978). In the absence of thermodynamic data different from that of Ellis (1978), we used XNaCl curves of Ellis (1978) to estimate, as a first approximation, the NaCl salinity of coexisting fluids in the belt. The XNaCl in scapolite from the Munali hills area range from 0.2 to 0.5 mol, whereas those from the Copperbelt region (Tembo 1994) are higher and vary from 0.5 to 1 (Fig. 18.6). Salinities in this range are known from areas where evaporite sequences occur (Rich 1979; Roeder 1984; Behr and Horn 1982).

Fig. 18.6
figure 6_18

Plot of XCl,Scap-EqAn of scapolites from the Lufilian–Zambezi belt, superimposed on the NaCl salinity curves of Ellis (1978). Data points of scapolite from the Munali hills area indicate salinities ranging from 0.2–0.5 mol, whereas those from the Copperbelt region are unrealistically high. Same symbols as in Fig. 18.4

Full size image

Fluid inclusion studies from carbonate hosted Kabwe Pb–Zn deposit in central Zambia indicate H2O–NaCl–CaCl2 and H2O–MgCl2–CaCl2 fluid compositions; salinities ranging from 11 to 31 wt% NaCl equivalent (EQ) and homogenisation temperatures varying from 257°C to 305°C (Kamona 1993). In the Chisamba area, north of Lusaka, H2O-NaCl fluid inclusions in quartz from vein quartz yielded fluid salinities ranging from 34 to 44 wt% NaCl EQ, and temperatures between 250°C and 350°C (Katongo 1999). These independent fluid inclusion studies indicate relatively high NaCl salinities at conditions consistent with greenschist facies metamorphism, but the host rocks are not scapolite bearing.

Scapolites from both the Munali hills area and Copperbelt region have strikingly similar calcian-marialitic compositions (23–48 Me%), but this may well just be a coincidence, as the two areas are several hundreds of kilometres apart. However, similar compositions were also found with the crude analytical techniques mentioned earlier. However, Munyanyiwa (1990) reported scapolite of mizzonite composition from southern Zambia without providing mineral compositional data.

Experiments by Vanko and Bishop (1982) showed that marialitic scapolite could be formed in fluids with little or no calcite. A minimum salinity for the production of marialitic scapolite in the absence of calcite is about 0.5 mol NaCl at 700–750°C (Vanko and Bishop 1982), which falls within the salinity range of our results (0.2–0.5 mol NaCl), estimated from the salinity curves of Ellis (1978). Some of the higher XNaCl values in our data set are unrealistic and underscore the limitation of applying Ellis’s experimental results to systems of different thermodynamic conditions, as pointed-out by several other authors (e.g., Mora and Valley 1989).

The composition of scapolite in the Munali hills area is independent of the lithology and coexisting mineral phases. The composition of coexisting plagioclase (An23–48), which lies within the same compositional range as the anorthite equivalent of the scapolite (EqAn29–48), appears to be an important factor in favoring scapolitization, because plagioclase from non-scapolite bearing Lusaka and Nchanga granites is more albite-rich (An5–20) than that coexisting with scapolite. Biotite is most commonly used in estimating halogen contents of fluids (Munoz and Swenson 1981) but we found no direct correlation between scapolite and biotite in the Munali hills granite gneiss, especially that Cl-bearing biotite also occurs in non-scapolite bearing Lusaka and Nchanga granites. However, phlogopite-rich biotite, which occurs in equilibrium with scapolite in veinlets in some amphibolites, was evidently formed under similar metamorphic conditions with scapolite. Similarly, the metamorphic Cl-bearing amphiboles also occur in equilibrium with scapolite and probably formed during the same metamorphic event.

6.2 Mineral Assemblages and Metamorphism in the Lufilian–Zambezi Belt

In the Lufilian–Zambezi belt, scapolites occur in mineral assemblages that are indicative of amphibolite facies metamorphism. In southern Zambia, amphibolite facies metamorphism is reflected by the scapolite–tremolite–diopside assemblage in calc-silicates, and scapolite–garnet–hornblende assemblage in garnet amphibolites (Simpson et al. 1965; Munyanyiwa 1990; Katongo 1999). However, in the Copperbelt region, Tembo (1994) reported greenschist facies assemblages comprising scapolite–chlorite–epidote–sericite in metagabbros. However, elsewhere around the world, scapolites occur in association with mineral assemblages that are stable at metamorphic conditions ranging from amphibolite to granulite facies (e.g., Vanko and Bishop 1982; Markl and Piazolo 1998; Faryad 2002). In our sample suite, scapolite formed mainly at the expense of plagioclase, especially in the Munali hills granite gneiss, however the scapolite–plagioclase assemblage is stable over a wide range of metamorphic conditions. Breakdown of plagioclase to sericite is indicative of greenschist facies metamorphism, which post-dated scapolitization. In amphibolites, however, scapolite is in equilibrium with both Cl-bearing biotite and amphiboles, suggesting that the assemblage formed during amphibolite facies metamorphism.

The symplectite-like intergrowth of scapolite with quartz (Fig. 18.2b) observed in the Munali hills granite gneiss is not common in most scapolite bearing rocks. Because scapolite invariably replaced plagioclase in most rocks, this texture could have formed in two ways. Myrmekite-like intergrowths between scapolite and quartz adjacent to K-feldspar are common in most samples, therefore the first and most likely possibility is that scapolite replaced plagioclase, which had already formed myrmekite intergrowth with quartz. The second possibility is that scapolite replaced calcium-rich plagioclase. Since scapolite has less SiO2 in the crystal lattice than calcium-rich plagioclase, the excess SiO2 was exsolved after replacement and formed worm-like blebs as inclusions in scapolite, resulting in the symplectite-like intergrowth. A replacement of K-feldspar by scapolite was not found in the investigated samples.

6.3 Origin of Scapolite and NaCl-Rich Fluids

The distribution of scapolite in the field is often the best evidence for the origin of this mineral. Several modes of field occurrence that indicate the possible origin of scapolite are proposed. Scapolite distributed as fine-scale interbedded units of scapolite-bearing beds in meta-evaporites suggests strongly for an in-situ salt precursor (e.g., Hietanen 1967; Mora and Valley 1989; Markl and Piazolo 1998). In southern Zambia, most scapolite-bearing marble and calc-silicate rocks probably represent in-situ former evaporitic sequences because metamorphism in most cases was essentially isochemical (Munyanyiwa 1990; Munyanyiwa and Hanson 1988). Any other mode of occurrence is attributed to metasomatism (e.g., Vanko and Bishop 1982). Metasomatic introductions of NaCl fluids are associated with scapolite occurrences in meta-igneous rocks, which otherwise do not contain scapolite. In the Copperbelt region, evidence of former evaporites provides unambiguous precursor for the external source of NaCl-rich fluids for the origin of the scapolite. In the Munali hills and surrounding areas, there is no hard evidence indicating the occurrence of former evaporites in the area. However, Cl-rich scapolite in the Munali hills area, coupled high NaCl salinities of coexisting fluids determined in this study, are indicative of an external source of NaCl-rich fluids from former evaporite horizons in the supracrustal rocks that correlate well with the Katangan metasediments in the Copperbelt region. Thus, the most credible source for NaCl fluids in the Lufilian–Zambezi belt, as previously suggested by many authors, is the dissolution of evaporite horizons in metasedimentary sequences during regional amphibolite facies metamorphism that channeled through fault and shear zones in the area. These investigations show that the study area is of great importance to understand the relation between evaporites and scapolites.

7 Summary and Conclusions

We conducted fieldwork in the Munali hills area, in the southern part of the Lufilian–Zambezi belt, and performed petrographic and mineral analyses of ­scapolite, amphibole, biotite and plagioclase from the Munali hills granite gneiss, amphibolites and metagabbros, in an effort to evaluate the nature of fluids, their source and attending metamorphic conditions. Results of this study show that:

  1. 1.

    The scapolites from the Munali hills area and from the Copperbelt region have strikingly similar Ca-rich marialite compositions (28–46 Me%), indicating that the mineral formed from hydrothermal fluids from precursor rocks of similar composition.

  2. 2.

    The scapolite in the Munali hills area occurs as pervasive replacement of plagioclase (21–48 XAn%) by metasomatic processes and occurs in stable equilibrium with mineral assemblages that are indicative of amphibolite facies metamorphism. The abundance of scapolite throughout the belt suggests that the grade of metamorphism in most parts of the belt was at least amphibolite facies, at which scapolite is more stable than plagioclase in the presence of NaCl-rich fluids.

  3. 3.

    The source of NaCl-rich fluids for the formation of scapolite in the Copperbelt region is evidently from evaporite horizons in the Katangan metasedimentary succession. In the southern part of the belt, including the Munali hills area, there is no direct evidence of the former existence of the evaporite beds. The prevalence of Cl-rich scapolite in the Munali hills area and high NaCl salinities of coexisting fluids, indicate former extensive evaporite successions in the supracrustal rocks that correlate well with the Katangan metasediments in the Copperbelt region.