Metamorphic modifications of the Muremera mafic–ultramafic intrusions, eastern Burundi, and their effect on chromite compositions

Journal of African Earth Sciences 101 (2015) 19–34

Contents lists available at ScienceDirect

Journal of African Earth Sciences journal homepage: www.elsevier.com/locate/jafrearsci

Metamorphic modifications of the Muremera mafic–ultramafic intrusions, eastern Burundi, and their effect on chromite compositions David M. Evans ⇑ Scientific Associate, Earth Science Department, Natural History Museum, Cromwell Road, London SW7 5BD, United Kingdom

a r t i c l e

i n f o

Article history: Received 28 April 2014 Received in revised form 31 August 2014 Accepted 1 September 2014 Available online 16 September 2014 Keywords: Karagwe–Ankole Belt Kibaran tectonomagmatic event Mafic–ultramafic intrusion Alteration Serpentinization Ferritchromite

a b s t r a c t The Muremera mafic–ultramafic intrusions were emplaced into metasedimentary rocks of the Karagwe– Ankole Belt in eastern Burundi, as part of the Mesoproterozoic Kibaran tectonomagmatic event. Igneous minerals of the Muremera intrusions have been partly altered to hydrous and carbonated metamorphic assemblages, although in most cases, the original igneous textures are well-preserved. Rounded, subhedral cumulus olivine has been partially and pseudomorphically replaced by lizardite–magnetite meshrim and lizardite–brucite mesh-centre assemblages, while anhedral interstitial plagioclase has been replaced by chlorite–tremolite. A later and localized event results in prograde alteration to antigorite– magnetite–chlorite–talc–carbonate and talc–carbonate–chlorite assemblages. The rocks are inferred to have undergone at least three separate metamorphic/alteration events resulting in: AS1 – an early alteration assemblage (mesh-rim lizardite–magnetite) characterized by very low fluid/rock ratios and widespread distribution; AS2 – a later, widespread low-temperature retrogressive (mesh-centre lizardite– brucite) assemblage associated with abundant close-spaced parallel veins; AS3 – later, prograde (antigorite–magnetite) and AT4 (talc–chlorite–carbonate) assemblages associated with more localized shearing and higher fluid/rock ratios. The AS1 assemblage most likely represents deuteric alteration that occurred soon after intrusion and cooling. The AS2 assemblage may relate to a continuation of this cooling, or may be correlated with the regional upright D2 folding event, while the AS3 and AT4 alteration assemblages are most likely correlated with the N–S oriented D3 faulting episode linked to the distal East African Orogeny. Euhedral to subhedral chromite grains are essentially unaltered where enclosed in primary unaltered olivine, pyroxene or plagioclase, as well as in AS1 lizardite–magnetite and AS2 lizardite–brucite altered olivine or pyroxene. In samples which show alteration to AS3 antigorite–magnetite and AT4 talc–carbonate– chlorite assemblages, the chromite grains are zoned to Mg–Al-poor ferritchromite rims. The width and the composition of these alteration rims are related to the degree of alteration of the silicate assemblages. However, it is concluded that chromite found in rocks that have avoided late/localized AS3–AT4 alteration preserve their magmatic to late-magmatic chemical and textural characteristics, and can therefore be used in regional geological studies and exploration. Ó 2014 Published by Elsevier Ltd.

1. Introduction The Kibaran belt (sensu latu) has long been proposed as a key tectono-stratigraphic domain in the Precambrian geological evolution of central Africa (Cahen et al., 1984; Clifford, 1970), and particularly that of the proto-Congo craton (de Waele et al., 2008). Recent geological investigations in the north-east part of this belt (the Karagwe–Ankole Belt or KAB) have focussed on its stratigraphy and structure (Fernandez-Alonso et al., 2012), its igneous activity (Buchwaldt et al., 2008; Duchesne et al., 2004; Tack ⇑ Address: 21 rue Jean de la Bruyère, 78000 Versailles, France. Tel.: +33 139209995. E-mail address: [email protected] http://dx.doi.org/10.1016/j.jafrearsci.2014.09.004 1464-343X/Ó 2014 Published by Elsevier Ltd.

et al., 2010) or its metallogeny (Deblond and Tack, 1999; Dewaele et al., 2010; Maier and Barnes, 2010; Maier et al., 2010; Pohl et al., 2013). These studies have tended to confirm the intraplate setting for orogeny and igneous activity as originally proposed by Klerkx et al. (1987) and Tack et al. (1994), as opposed to the convergent margin setting proposed by Rumvegeri (1991) and Kokonyangi et al. (2006), at least in the KAB. However, a better understanding of the metamorphic history of the KAB is needed to complement the early structural studies and in order to place the proto-Congo craton in the context of plate reconstructions of Precambrian supercontinents (Pisarevsky et al., 2014). Relatively few studies of the metamorphism of the KAB have been carried out since the regional mapping work in Burundi (Sintubin, 1989; Tack and Deblond, 1990). These earlier studies

20

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

concentrated on the initial high grade peak of metamorphism in the KAB, whereas the more recent work by Dewaele et al. (2010, 2011) has examined the later phases of metamorphism in the context of hydrothermal mineralization. There is thus a need for further detailed studies of the full metamorphic evolution within structurally well-constrained rock units within the KAB. Ultramafic rocks are highly sensitive to metamorphic conditions when a fluid is present and they tend to preserve textural evidence from different generations of fluid ingress (Evans, 1977; Wicks et al., 1977). This fluid may have originated either from within the crystallizing rock itself (residual hydrous melt), or from outside, gaining access to the rock via through-going fractures (veins and shears) and/or by grain boundary infiltration. Wicks et al. (1977) and Wicks and Whittaker (1977) have proposed a textural and compositional framework for the interpretation of serpentine alteration of ultramafic rocks. In particular, they draw attention to the probable timing relationship of the alteration of olivine along its margins and its internal cooling cracks (which they term mesh-rim serpentinization) and alteration of the remaining olivine between these mesh-rims (mesh-centre serpentinization). By careful study of these textures and of the serpentine polymorphs involved in the hydration reactions, the metamorphic evolution of the ultramafic rocks can be reliably deduced. For example, Beard et al. (2009) have been able to distinguish two different generations of serpentinization in partly altered troctolites of the oceanic crust, the first formed in a rock-dominated system at T > 300 °C, the second in a fluid-dominated system at T < 300 °C. The chromite occurring in ultramafic rocks is also potentially useful as an indicator of both igneous and metallogenetic processes (Averill, 2011; Irvine, 1965), but only if its modification by subsequent metamorphism and alteration is well understood (Abzalov, 1998; Barnes, 2000; Merlini et al., 2009). There have been many studies on the effects of metamorphism on primary igneous chromite (Abzalov, 1998; Barnes, 2000; Burkhard, 1993; Kimball, 1990; Wylie et al., 1987) and its involvement in metamorphic reactions within ultramafic rocks (Evans and Frost, 1975). It has generally been observed that igneous chromite is only weakly affected by metamorphism up to the lower to mid-greenschist facies, but that it is generally strongly altered in upper greenschist facies and above. To a great extent, the degree of alteration of the chromite, especially at lower temperatures is controlled by the fluid to rock volume and by the type of immediately adjacent minerals. For example, chromite within continental layered intrusions that have undergone greenschist facies or lower conditions of metamorphism with relatively low water–rock ratios are relatively unaffected by metamorphism (Abzalov, 1998; Cameron, 1975; Wilson, 1982), whereas chromite within layered intrusions of oceanic affinity that have been subject to pervasive sea floor alteration at greenschist facies with high water–rock ratios can be significantly modified (Kimball, 1990; Merlini et al., 2009). Those spinels adjacent to or involved with reactions with Al-bearing alteration phases (e.g. hornblende and chlorite) may experience significant changes of their Cr/(Cr + Al) ratio, whereas those spinels that are only surrounded by serpentine experience only changes of Mg/ (Mg + Fe2+) ratio (Kimball, 1990). The access of CO2-bearing fluids resulting in talc–carbonate assemblages is also cited as resulting in greater compositional change to chromite (Barnes, 2000; Burkhard, 1993). The most obvious change to igneous chromite during metamorphism is the development of ferritchromite rims replacing the outer zone of the original grain (Spangenberg, 1943; Merlini et al., 2009). The progressive alteration of chromite with prograde metamorphism is thought to commence with overgrowth of secondary magnetite rims on the igneous grain during lowtemperature serpentinization (Abzalov, 1998; Barnes, 2000). This is followed at higher temperature by commencement of a

dissolution–precipitation reaction, initially at the interface between chromite and magnetite and progressing inwards, involving loss of Al and Mg to surrounding Al-bearing silicates, and their replacement by Fe3+ and Fe2+ (Barnes, 2000; Wylie et al., 1987). Merlini et al. (2009) propose that ferritchromite forms in two stages; first with the retrograde serpentinization of adjacent olivine, followed by prograde reaction of the aluminous chromite and adjacent serpentine in the presence of water and oxygen to form ferritchromite (replacing the spinel) and chromian chlorite (replacing serpentine). In any case, the result is usually a zoned crystal with an aluminous core and iron-rich rim, or in the case of complete reaction to an equilibrium assemblage, a homogeneous Al-depleted spinel grain. The composition of the equilibrium textured spinel (ferritchromite or chromian magnetite) or of the outer margin of the ferritchromite rim against magnetite in zoned grains can indicate the temperature conditions of the metamorphism (Barnes, 2000; Evans and Frost, 1975). This study investigates the effects of deformation and metamorphism on the relict igneous minerals (chromite in particular) in small mafic–ultramafic intrusive bodies of the KAB in eastern Burundi. The results of this study are of relevance for the understanding of regional deformation and metamorphic events on the proto-Congo craton, and will be of practical use in regional metal exploration using chromite as an indicator mineral.

2. Regional geology The KAB is a Palæo- to Mesoproterozoic tectonostratigraphic domain that covers most of Burundi, Rwanda, and parts of northwest Tanzania, southwest Uganda, the north and south Kivu provinces of the Democratic Republic of Congo (Tack et al., 2010). It comprises two main superposed elements: 1 – an underlying Archaean to Palæoproterozoic metamorphic basement making up part of the proto-Congo craton, which is overlain by 2 – a siliciclastic, shallow marine to epicontinental sedimentary sequence (Baudet et al., 1988; Fernandez-Alonso et al., 2012). Based on U– Pb dating of both volcanic and detrital zircons in the sedimentary sequence, sedimentation occurred between 1780 Ma and 1400 Ma (Fernandez-Alonso et al., 2012). The KAB sedimentary sequence has been deformed by three main tectonic events: an early bedding-parallel extensional D1 event characterized by mylonitic zones, a compressional D2 event resulting in open, upright folding (F2), and a late N–S oriented transpressional faulting event (D3) that has reactivated older structures (Fernandez-Alonso et al., 2012; Klerkx et al., 1987; Tack, 1990). The D1 mylonitic extension is closely associated with the Kibaran tectonomagmatic event, which comprises widespread within-plate igneous activity on the broader proto-Congo craton in central Africa between 1400 Ma and 1350 Ma (Klerkx et al., 1987; Tack et al., 2010). The D2 compressional episode (D2a event of Klerkx et al., 1987) has been demonstrated to post-date the Kibaran intrusive event (Evans et al., 2000) and it is now assumed to be caused by the distal effects of the Irumide orogeny at the southern margin of the proto-Congo craton between about 1050 and 1000 Ma (Fernandez-Alonso et al., 2012). These authors and Dewaele et al. (2011) suggest that the late transpressional D3 event (D2b event of Klerkx et al. (1987)) is related to far-field effects within the proto-Congo craton of the East African orogeny to the east at around 590–550 Ma. Note, however, that recent detailed structural observations in the Nyakahura-Biharamulo area of Tanzania and their synthesis by Koegelenberg and Kisters (2014) suggest that all three deformation events may have their origin in a southeasterly-verging fold-thrust belt that records overall convergent tectonics between the Congo and Tanzania cratons. Evidently, further structural and metamorphic studies are needed in the KAB.

21

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

30°30’E

Sintubin (1989) in one of the few published studies of the metamorphism of the KAB rocks recorded a rapid increase of metamorphic grade from southeast (almandine low grade) to northwest (almandine high grade) in the Kazingwe complex of south east Burundi, which he ascribed to overthrusting of a stack of nappes during either the D1 or D2 events. On the other hand, Tack and Deblond (1990) have drawn attention to the thermal metamorphic effects of the voluminous intrusion of granitic and gabbroic rocks of the Kibaran tectonomagmatic event on the enclosing KAB sedimentary rocks in the Nyabikere area of Burundi. They observed the sequential appearance in pelitic rocks of mineral assemblages including andalusite, staurolite, cordierite and sillimanite as one approaches the intrusions. They particularly noted the random fabric of the porphyroblasts, apparently overprinting earlier beddingparallel planar fabrics (D1), and this observation is also made on a wider scale in the explanations of geological map sheets of the eastern part of Burundi (for example Karayenga, 1987; Waleffe, 1981) and northwestern Tanzania (Grey, 1967). The later D2 folding and D3 shearing events seem to have either been too weak or too localized to have greatly modified or overprinted this thermal metamorphic event in the bulk of the sedimentary rocks. Nevertheless, a partial greenschist facies retrogression of the earlier, randomly oriented thermal metamorphic porphyroblasts has been noted in places (Evans et al., 2000). The 300 km long Kabanga–Musongati alignment (KMA) of Mesoproterozoic mafic–ultramafic intrusions are the primitive mantlederived element of the Kibaran tectonomagmatic event in eastern Burundi and northwestern Tanzania (Deblond, 1993; Tack et al., 1994, 2010). The Muremera group of intrusions are part of the KMA and are located in eastern Burundi, close to the national boundary with Tanzania. They comprise three main groups of individual mafic–ultramafic intrusive bodies, labelled here Muremera A, Muremera B and Rujungu, from south to north (Fig. 1). These small, olivine cumulate-dominated intrusions were brought to light in 1978 by a multilateral mineral exploration program funded by the United Nations Development Program (UNDP:

Granite Mafic sill Ultramafic

Rulenge

TAN

Fault (D3) Fold axis (F2)

ZAN

Bouzet, 1980). The Muremera intrusions have a lensoidal sill or ‘‘bean pod’’-like surface expression (Figs. 1 and 2), with a concentric arrangement of rock types from olivine–orthopyroxene– sulphide–chromite ortho- to mesocumulate rocks in the centre, to a gabbronoritic marginal zone extending laterally as sills, and showing clear evidence for assimilation of and hybridization with partially molten metasedimentary rocks (Niyondezo et al., 1982; Muhagaze, 1984). In most respects, the Muremera intrusions are very similar to those of the Kabanga area in adjacent Tanzania (Fig. 1), which have been recently described in detail by Maier et al. (2010) and Maier and Barnes (2010). Both the Kabanga and the Muremera intrusions have suffered pervasive but incomplete serpentinization of the olivine-bearing rocks and uralitization of gabbroic rocks (Evans et al., 2000; Maier et al., 2010). Adjacent metasedimentary rocks show clear thermal metamorphic effects, analogous to those observed by Tack and Deblond (1990), although often retrogressed to greenschist facies mineral assemblages (Evans et al., 2000). The Muremera B group is the most well-known and investigated of the intrusive bodies at Muremera. Here, the UNDP exploration program drilled four shallow diamond drill holes in the northern intrusive body (MURB-F and MURB-S series), and commercial companies (BHP-Billiton/Danyland Ltd.) have subsequently drilled a further four holes (MURB_D series: Fig 2). This study is based upon the detailed surface mapping, geological and geotechnical logging and petrographical examination of both UNDP and commercial drill cores carried out by the author during commercial exploration work (Dwyka Resources, 2007, 2008) and during subsequent follow-up research work carried out at the Natural History Museum, London. 3. Study methods As fresh ultramafic rocks do not generally outcrop in Burundi, the samples used in this study came from exploration drill holes sunk in the course of commercial geological exploration work at Muremera B. This work followed generally accepted methodologies and principles that are detailed in Appendix A (in Supplementary material). These samples represent varying stages of

IA

60

MURB-S7 MURB-F2

Kabanga

MURB-F1

BU -3°00’S

RU

58

MURB_D001

Mylonite (D1)

MURB-F3

MURB_D003 MURB_D002

ND

I

Rujungu

MURB_D004 60 65

Muremera B

Stream

Kigamba Muremera A

Inferred fault Ultramafic rock Gabbronorite Quartzite

10 km

Mica schist

1 km Fig. 1. Geological sketch map of the Kabanga–Muremera area in eastern Burundinorthwestern Tanzania showing interpreted major structures. Granites and major structures are from regional geological maps (Grey, 1967; Karayenga, 1987; Waleffe, 1981); mafic sills and ultramafic bodies are from author’s mapping and interpretation. Blank areas are occupied by metasedimentary rocks of the KAB.

Graphitic schist

Fig. 2. Interpreted geological map of the Muremera B intrusions based on author’s drill core logging, field traverses and interpretation of magnetic images. The location and sub-surface extent of drill holes is shown as a dot with arrow.

22

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

deformation and alteration with respect to one of the carbonated shear zones. All samples were examined petrographically by the author to characterize structural and alteration elements. In-situ micro-X-ray diffraction (XRD) analysis of serpentine minerals from two samples was carried out at the Natural History Museum, London. The X-ray beam from a GeniX Cu source was focussed and reduced to a 0.1 mm diameter beam directly onto polished thin sections with a 10° 2h incidence angle. Further details of the XRD method are given in Appendix A of the Supplementary material. Chromite and silicate mineral grains from seven of the samples have been analyzed by wavelength-dispersive electron probe microanalysis (EPMA) using Cameca SX50 and SX100 instruments at the Natural History Museum, London. The major elements were calibrated using natural or synthetic mineral standards and the trace elements using pure metal, sulphide or natural oxide standards. The operating conditions for the EPMA analysis of chromite, olivine and pyroxene used a 20 kV accelerating voltage and a 20 nA beam current, the beam being focussed to a 1–2 lm point. Analysis of plagioclase, serpentine and other hydrous minerals used a 15 kV voltage, 20 nA beam current and a beam defocused to 10 lm diameter. Full details of the EMPA analysis methods and tables of the results are given in Appendix A and Tables A.1 and A.2 of the Supplementary material.

4. Metamorphism of ultramafic rocks at Muremera B 4.1. Geological structure from mapping and core logging Mafic–ultramafic rocks outcrop very poorly at Muremera B prospect, but occur in all of the eight drill holes that have been examined (Fig. 2). The dominant rock type in the centre of the intrusions is a feldspathic or sulphidic harzburgite (olivine–orthopyroxene–chromite cumulate rock). Other common rock types in the centres of the bodies include olivine melanorite, feldspathic orthopyroxenite and olivine orthopyroxenite, whereas biotitic quartz gabbro and melanocratic gabbronorite occur at the margins. All of these rock types have been altered to some extent or another and the intensity and type of this alteration is generally related to the presence and abundance of through-going faults, shears or vein systems. Three main types of planar structure are identified in ultramafic rocks at Muremera B:  Type 1 – a widespread foliation of very fine (<1 mm width), locally closely-spaced anastomosing veinlets of dark-green serpentine that have a consistent steep orientation (Fig. 3a).  Type 2 – discrete dilatant and transpressional veins generally filled with carbonate and/or pale serpentine, chlorite or other hydrous minerals (0.2–10 cm width: Fig. 3b).  Type 3 – discrete faults and shear zones with schistose textures (5–100 cm in width). The Type 1 foliation veinlets are distributed widely but irregularly throughout the rock mass and are not related to the Type 2 and 3 faults and veins. In fact, they are generally only seen in the absence of the abundant veining and minor shears related to fault zones. In hand specimens of drill core, the Type 1 serpentine veinlets are most easily visible where they cut across light-coloured interstitial sulphide aggregates (Fig 3a). Where reliable orientation of core has been carried out, they are seen to have a consistent relatively steep dip to the northwest at Muremera B. Both extensional and transpressional minor structures (Type 2) are more abundant in proximity to the larger faults or shears (Type 3). Preliminary structural analysis of the faults and larger veins in drill core reveals that they have a wide variety of orientations. In

detail, it is seen that there are several generations of planar extensional and transpressional structures, some cross-cutting or overprinting others and that they are irregularly distributed in the rock mass. At this stage, however, no further subdivision of these meso-scale planar structures can be made due to the lack of rigourous measurement of orientations. Logging of drill core and surface mapping has identified a major zone of faulting running approximately north–south and dipping moderately to the east in the Muremera B area (Fig. 2). In gabbroic rocks in the drill core, this Type 3 faulting is characterized by abundant schistose biotite and chlorite and complete destruction of primary textures. In olivine-rich cumulate rocks, Type 3 faulting is characterized by strongly-developed ductile shear textures in talc–carbonate–chlorite rock. Pale-coloured altered rocks with a relatively high abundance of smaller shears, veins and fractures (Type 2 structures), but in which the primary rock texture is still recognizable, occur for 2–10 m on either side of the strongly sheared zones (Fig. 3b). The broad distribution characteristics of the Type 2 structures indicate that they are related to and subsidiary to those of Type 3. 4.2. Distribution and intensity of alteration Based on drill core logging, hand specimen examination and routine petrographical examination, the intrusive rocks at Muremera B can be divided into three classes according to their degree of alteration of primary anhydrous minerals to secondary hydrous or carbonate minerals: Class (1) unaltered and very weakly altered; Class (2) partly-altered; and Class (3) completely altered. The third class of completely altered rocks is subdivided into those rocks that preserve their original magmatic textures (Class 3i) and those in which primary textures are obliterated (Class 3ii). These classes of alteration are summarized in Table 1 and further detail is given in Appendix B of the Supplementary material. Completely unaltered or very weakly altered rock is uncommon at Muremera B (Table 1). It occurs in those parts of the intrusions that are free of visible veins or veinlets, and distal from larger shears. In unaltered rock, primary magmatic minerals are easily identified. For example, unaltered olivine grains are recognizable by their greyish-brown colour and a pronounced schiller reflection in sunlight due to the presence of abundant minute crystallographically-controlled rod-like and symplectite exsolution inclusions. Orthopyroxene often occurs as poikilitic grains forming 10 mm sized pale green spots within the darker olivine cumulate groundmass. In the bulk of the rock, however, alteration of the igneous minerals is partial to complete but the macroscopic primary cumulus igneous textures are well-preserved, making identification of the pre-alteration igneous minerals possible (Classes 2 and 3i: Table 1). The alteration is mainly characterized by hydrous or carbonate minerals, and so is related to ingress or redistribution of fluids at submagmatic temperatures. These fluids access the rock by two main mechanisms: either along grain boundaries, or via dilatant fractures and veins. As the abundance or size of Type 2 veins and shears increases towards fault zones, the primary igneous minerals become more completely altered and overprinted until only a few highly resistant minerals such as chromite remain. However, even in these rocks, the original rock type can usually be deduced based on alteration mineral assemblage and the presence of faint ‘‘ghost’’ outlines of pseudomorphed igneous textures (Fig. 3b). Gabbroic and melanocratic gabbronorite rocks are characterized by uralitic and saussuritic alteration giving the rock its typically greenish colouration. Pyroxenites that do not contain any olivine have a similar uralitic alteration assemblage, which is dominated in this case by a paler-green tremolitic actinolite. Altered harzburgites are usually dark-coloured (dark grey to nearly black) being characterized by serpentinization that often preserves

23

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

(a)

(b)

Type 1

CV

Type 2

$ o

o

pl

o o

pl

1 cm

1 cm

Fig. 3. (a) Partly serpentine-altered sulphidic peridotite with fine Type 1 serpentine veinlets (black, parallel to arrow within circle) cutting across interstitial sulphides ($: light coloured) and serpentinised olivine (o: black) from bottom left to top right of image. Plagioclase (pl) also occurs interstitially to olivine. Striations from top-left to bottom-right of image are caused by sawing of the drill-core (MURB_D002 at 177.2 m); (b) highly-altered feldspathic peridotite, with talc–carbonate-altered olivine grains (o: pale, rounded) surrounded by chlorite–tremolite-altered feldspar (pl: darker, blue–green); a Type 2 carbonate vein (CV) cuts the rock in the upper part of the image (MURB_D002 at 119.3 m). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1 Division of alteration classes based on core logging of drill hole MURB_D002.

a b

Class

% of Rock massa

Relative intensity

% Alteration of rockb

Primary textures

Associated structures

1 2 3i 3ii

10 40 45 5

Very weak Partly-altered Complete Complete

<20 20–90 >90 >98

Preserved Preserved Part-modified Obliterated

Type 1 – (Type 2) Type 2 Type 3

Estimated percentage of intrusive rocks in the alteration class on scale of intrusion. Approximate volume percentage of alteration minerals in the rock on hand sample scale.

igneous textures remarkably well (Fig. 3a). More intense alteration of olivine-rich rocks adjacent to faults and shear zones results in bleaching of the rock to a more uniform light grey colour due to pervasive talc–carbonate–chlorite alteration (Fig. 3b). 4.3. Metamorphic petrology Detailed petrographical examinations and mineralogical analyses were undertaken on a limited number of representative samples of the three main classes of alteration described above, in order to investigate more thoroughly the alteration of the ultramafic rocks. The nine samples that were taken for more detailed petrography are listed with their characteristics in Table 2. The alteration of these rocks is described in more detail with reference to individual samples that best represent each alteration class in Appendix C of the Supplementary material. The samples all represent plagioclase and sulphide-bearing olivine(-orthopyroxene) orthocumulates (varying from plagioclase harzburgite to olivine melanorite lithologies) and in all except one sample, magmatic textures are largely preserved. From these textures and variably preserved primary igneous mineral remnants it can be deduced that the principal cumulus mineral in these samples was olivine (30–70 vol.%), which was rounded to subhedral, 1–4 mm in size and with composition Fo75–Fo87. Chromite (1–2 vol.%) was always present with olivine as a cumulus mineral, sometimes included within it as small, rounded grains, but mostly occurring as loose clusters of euhedral to subhedral 0.05–0.2 mm sized grains or as isolated coarser (0.2–0.5 mm) grains enclosed

in intercumulus minerals. In olivine-rich rocks (harzburgites) the intercumulus minerals include 6–15 mm sized oikocrysts of orthopyroxene (20–40 vol.%); En79–En89), and a medium-grained groundmass of 1–3 mm sized, xenomorphic (interstitial)-textured plagioclase, clinopyroxene and phlogopite (together, 5–30 vol.%). In less olivine-rich rocks (olivine pyroxenite and olivine melanorites), orthopyroxene was commonly a euhedral shaped cumulus mineral, with plagioclase, clinopyroxene, phlogopite and hornblende as the main interstitial minerals. Magmatic sulphide (pyrrhotite  pentlandite > chalcopyrite) also occurs interstitially to the olivine and chromite as intercumulus patches (2–20 vol.%) in all of these rock types. The least altered sample (Class 1) came from drill core with no visible veining either in hand specimen or under the microscope. In it, however, both cumulus olivine and interstitial plagioclase are partly altered. Olivine is mostly fresh but all of its internal cooling fractures are altered to narrow zones of a fine intergrowth of serpentine and magnetite (mesh-rim serpentine). Plagioclase has reacted with olivine, where they are in contact, to form a kelyphitic reaction rim texture of finely acicular tremolite and flaky chlorite (Fig. 4a). Plagioclase in contact with cumulate orthopyroxene is unaltered. Pyroxenes in this sample are largely unaltered, but may contain minor fine-grained flecks of actinolite–tremolite along fracture or cleavage planes. Chromite grains show no evidence of alteration (Fig. 5a). Partly altered samples of Class 2 were collected from drill core containing mainly Type 1 close-spaced fine serpentine veinlets although minor and wide-spaced Type 2 veins may occur nearby.

24

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

Table 2 Description of samples used in this study. Mineral abbreviations from Siivola and Schmid (2007).

a

Samplea

Texture

Rock name

Class

Alteration (%)

Alteration assemblage

MBD2-227 m MBF1-40 m MBF1-47 m MBF1-69 m MBF2-72 m MBF3-79 m MBF3-97 m MBF2-68 m MBF2-75 m

Orthocumulate Orthocumulate Orthocumulate Orthocumulate Orthocumulate Orthocumulate Orthocumulate Orthocumulate Schistose

Olivine melanorite Plagioclase-Peridotite Olivine melanorite Plagioclase-Peridotite Plagioclase-Peridotite Plagioclase-Peridotite Plagioclase-Peridotite Plagioclase-Peridotite Tlc–Cb–Chl schist

1 2 2 3i 3i 3i 3i 3i 3ii

15 70 90 96 98 99 95 100 100

Srp–Mag–Tr–Chl Srp–Tr–Chl–Mag–Ser Srp–Tr–Chl–Mag–Ser Srp–Chl–Tlc–Mag–Tr–Cb Srp–Chl–Tlc–Cb–Mag–Tr Srp–Chl–Tr–Mag–Tlc–Cb Srp–Tr–Chl–Tlc–Cb–Mag Srp–Mag–Chl–Tlc–Cb Tlc–Cb–Chl–Mag

Sample name is made up of abbreviated drill hole number and down-hole depth.

Fig. 4. Photomicrographs (all plane-polarized transmitted light) of textures and assemblages of altered ultramafic rocks at Muremera-B: (a) olivine grain (Ol) cut by dark serpentine–magnetite (Srp–Mag) along fractures with a chlorite–tremolite (Chl, Tr) reaction rim adjacent to plagioclase (Pl; sample MBD2-227 m); (b) completely altered olivine grain showing distinction of clear mesh-rim type serpentine (Srp + Mag) along fractures and yellow–brown mesh-centre serpentine (Srp–Brc) replacing olivine cores (sample MBF1-47 m); (c) two translucent dark brown chromite grains (Chr) within pale yellow bastite (Srp(Opx)) after interstitial pyroxene between completely altered olivine and plagioclase (sample MBF1-47); (d) group of subhedral chromite (Chr) within flaky chlorite (Chl(Pl)) and high relief pumpellyite (Pmp) after interstitial plagioclase (sample MBF1-69 m); (e) cumulus olivine grain now completely replaced by a wide rim of clear interpenetrating serpentine (Srp(Ol)) and a core of intergrown serpentine– tremolite–magnetite and pyrite (Py; sample MBF2-68 m); (f) granoblastic dolomite grains (Dol) within a lepidoblastic matrix of intergrown talc–chlorite (Tlc + Chl; various shades of green) with irregular-shaped opaque sulphides (Po; sample MBF2-75 m). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The Type 1 serpentine veinlets are seen under the microscope to be filled with a very fine pale yellow–green coloured serpentine aggregate, with pseudo-fibres arranged perpendicular or oblique to the vein orientation. Where these veinlets cut across cumulus olivine grains, the olivine is completely altered pseudomorphically to two types of serpentine aggregate: along internal cooling fractures the serpentine is colourless and is intergrown with fine magnetite (mesh-rim texture, similar to that in unaltered samples),

whereas the intervening areas of olivine are altered to a very fine, pale yellow–green to yellow–brown coloured aggregate of serpentine without secondary magnetite (mesh-centre texture), very similar in appearance to the veinlet-filling serpentine but without any obvious pseudo-fibre fabric (Fig. 4b). Poikilitic orthopyroxene is also altered to a very fine pale yellow serpentine aggregate (bastite) where it is adjacent to or cut by the fine veinlets (Fig. 4c). Interstitial plagioclase shows kelyphitic

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

25

Fig. 5. Back-scattered electron images of chromite grains from Muremera-B intrusions: (a) small euhedral chromite (Chr) surrounded by serpentine (Srp) of the mesh-rim type (sample MBD2-227 m); (b) euhedral unzoned chromites (Chr) within orthopyroxene oikocryst (Opx) are cut through, but unaltered by serpentine (Srp(opx)) veinlets (sample MBF1-40 m); (c) group of subhedral chromite grains (Chr) within chlorite–sericite altered plagioclase (Chl(pl)), showing very narrow higher reflectance rims (sample MBF1-40 m); (d) large interstitial chromite (Chr) with a narrow ferritchromite (Fecr) rim (sample MBF1-47 m); (e) group of interstitial chromite grains (Chr) with a wider, spongy-looking ferritchromite (Fecr) rim (sample MBF2-68 m); (f) subhedral interstitial chromite grains (Chr) now almost completely replaced by the ferritchromite (Fecr) rim and with a ragged Cr-magnetite outer rim in contact with talc–carbonate–chlorite (Tlc + Chl + Dol) assemblage (sample MBF2-75 m). Mineral abbreviations from Siivola and Schmid (2007), except Fecr: ferritchromite.

reaction rims where it is in contact with olivine as in the leastaltered sample, but where it is cut by the yellow–green serpentine veinlets it is altered to a very fine clouding of saussurite. Chromite again shows no visible effects of alteration when completely surrounded by bastite serpentine (Fig. 4c), or when it is cut by serpentine veinlets (Fig. 5b). In samples of Class 3i alteration, taken from drill core variably affected by Type 2 veins and shears, the primary silicate minerals are almost completely altered. In samples with few or thin Type 2 veins, olivine mostly shows alteration to colourless mesh-rim serpentine associated with magnetite, and magnetite-free yellow–brown mesh-centre serpentine (Fig. 4d). Pyroxenes are partly to mostly altered to either pale yellow bastite, or colourless tremolite. Interstitial plagioclase is wholly altered to aggregates of a granular high relief mineral (possibly pumpellyite), set in flaky chlorite and acicular tremolite (Fig. 4d). In these samples, chromite grains that are hosted within interstitial plagioclase or sulphide are zoned to narrow, opaque, higher-reflectance ferritchromite rims (Fig. 5c and d). These ferritchromite rims often contain small (0.5–5 lm) blebs and lamellæ of a bireflectant mineral, presumed to be ilmenite, as well as inclusions of flaky silicates, possibly chlorite. These inclusions are more common towards the outer margin of the grain. The boundary between the inner low-reflectance core

and outer high-reflectance rim is usually sharp and has a rounded shape. In those Class 3i samples that are more affected by Type 2 veins (larger, more frequent), two notable differences in mineral assemblages are noted. The former olivine grains are now replaced by an apparently coarser, flaky, interpenetrating aggregate of colourless serpentine, in which the texture of internal cooling cracks has been lost (Fig. 4e). This colourless serpentine is associated with magnetite, but this is distributed either around the margin of the former olivine, or as an aggregate, intergrown with tremolite and valeriite in the centre of the former olivine (Fig. 4e). The second difference with previously described samples is that the former interstitial plagioclase is partially to totally replaced by aggregates of granular carbonate, talc and flaky chlorite. Chromites enclosed in these carbonate–chlorite aggregates have well-developed high-reflectance ferritchromite rims (Fig. 5e; Table 3). In the one sample representing Class 3ii alteration, taken from drill core within a Type 3 shear zone, the original magmatic textures are no longer discernible. There is no distinction between former olivine and former plagioclase areas. It is completely and homogeneously altered to a granular to schistose aggregate of carbonate–talc–chlorite, with patches of sulphide and magnetite hinting at its probable original texture of a sulphide-bearing olivine

26

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

cumulate rock (Fig. 4f). Chromite in this strongly altered sample is almost completely altered to ferritchromite, only a few of the larger grains retaining a small low-reflectance core (Fig. 5f; Table 3). 4.4. Microanalysis of alteration minerals Two samples containing three different serpentine textural types were chosen for in-situ micro-XRD analysis. Sample MBF140 m (Class 2) contains fine-grained clear serpentine along former fractures (mesh-rim serpentine), and abundant yellow–brown serpentine aggregates, replacing both former olivine cores between fractures (mesh-centre serpentine) and orthopyroxene (bastite) adjacent to veins. The diffraction spectrum obtained on a relatively large, homogeneous patch of yellow–brown mesh-centre serpentine aggregate is that of lizardite 1T with a strong {2 0 2} peak at 1.49 Å. In this spectrum, there is evidence for the minor presence of brucite (Fig. A.1 of Supplementary material). Diffraction patterns were also obtained on a relatively wide mesh-rim (50 lm) of clear serpentine along an olivine fracture and on a wide yellow bastite vein in orthopyroxene in this sample. Both these spectra also indicated the dominant presence of lizardite (Fig. A.2 of Supplementary material). Sample MBF2-68 m (Class 3i) contains abundant interpenetrating sheaves of a clear-coloured coarser serpentine within former olivine grains. Several analysis spots on large patches of clear serpentine, free of visible magnetite grains were undertaken. These diffraction patterns indicate the dominant presence of antigorite, which is demonstrated by the {2 0 2} peak occurring at 1.52 Å (Figs. A.3 and A.4 of Supplementary material). Electron probe microanalysis (Tables 4 and A.1 of the Supplementary material) of the serpentine aggregates in six samples using a defocused beam shows that these aggregates are composed of mixtures of varying proportions of serpentine (Mg3Si2O5(OH)4) with other Fe-rich, Si-poor phases, probably either magnetite (Fe3O4) or ferroan brucite ((Mg,Fe)(OH)2). The generally small width of alteration along olivine fractures relative to the beam diameter and the universal presence of visible thin magnetite bands along the centre of the fracture mean that the majority of analyses of mesh-rims represent mixtures of serpentine and magnetite and the presence of brucite is uncertain. The areas of yellow–brown mesh-centre serpentine aggregates, however, are much larger and are free of visible magnetite grains (at least to the resolution of the back-scattered electron imaging system of the microprobe – 0.1–0.2 lm), therefore the microprobe analyses can give a better estimate of the proportion of serpentine to brucite. The analyses show that the mesh-centre serpentine aggregates are composed of a mixture of a low-alumina Cl and Fe-bearing serpentine and a hydrous low-silica phase, most likely ferroan brucite. Individual analyses of mesh-centre serpentine generally have low totals (75–84%) and plot along a

straight-line mixing trend between serpentine and ferroan brucite with a Mg/(Mg + Fe2+) ratio of approximately 0.4 (Fig. 6, Table 4). The coarser, flaky serpentine replacing former olivine with an interpenetrating texture in sample MBF2-68 m has a relatively consistent composition corresponding to a pure serpentine containing minor amounts of structural Al2O3 (0.5–2 wt%) and FeO (3–6 wt%), low Cl and F and with an apparent structural excess of Si cations to divalent cations, a characteristic of the corrugated antigorite structure (Table 4, Fig. 6). Chlorite flakes with clear colouration occupying interstitial spaces after plagioclase have compositions varying from clinochlore to penninite, with Mg/(Mg + Fe) ratios between 0.9 and 0.93. Defocussed beam analyses of very fine grained blueish-green coloured aggregates between the clear chlorite flakes are either non-stoichiometric or contain extraneous elements such as K: the aggregates probably contain a very fine intergrowth of chlorite with either serpentine or sericite respectively. Tremolite that forms the reaction rim between olivine and plagioclase in the less-altered samples has Mg/(Mg + Fe) ratios between 0.93 and 0.94. 5. Mineral chemistry of chrome spinels at Muremera B Chromite grains in the least-altered samples (MBD2-227 m and MBF1-40 m) have compositions typical for layered intrusions derived from Mg-rich basaltic magma (Tables 5 and A.2 in Supplementary material). In particular, they have Cr/(Cr + Al) ratios of around 0.6 and low to very low ratios of Fe3+/(Al + Cr + Fe3+) from 0.01 to 0.09. Within each sample, the atomic ratios of Cr/(Cr + Al) and Mg/(Mg + Fe2+) vary systematically according to their igneous mineral hosts. Cumulate chromite hosted within large orthopyroxene oikocrysts of the harzburgites tends to have the highest values of both Cr/(Cr + Al) and Mg/(Mg + Fe2+) and lowest trace element values (TiO2, MnO, ZnO), whereas chromite within interstitial plagioclase, clinopyroxene and sulphide has lower values of major element ratios and higher trace elements (Table 5). As has been noted above, chromite grains are generally zoned to higher-reflectance rims in the more altered samples. Microprobe analysis (EPMA) of the low-reflectance cores of the zoned chromites in each altered sample shows that these cores largely overlap in compositional space with the chromites in least-altered samples (Table 5; Fig. 7). On a trivalent cation diagram, the cores are all relatively poor in the Fe3+ ion and have Cr/(Cr + Al) ratios between 0.5 and 0.6 (Fig. 7). The chromite cores of altered samples show the same degree of variation with enclosing host mineral as the chromites from least-altered samples. It seems therefore that the process of alteration of chromite to ferritchromite does not involve a significant change in the core composition. In contrast, the high-reflectance ferritchromite rims of chromites in different samples occupy distinct ranges within the trivalent cation space (Fig. 7). In samples MBF1-47 m, MBF1-69 m and

Table 3 Relative proportions of alteration assemblages and characteristics of the ferritchromite (FeCr) rims on chromites in samples. Abbreviations: Mod moderate; N.a. not analyzed; N.p. not present.

a b

Sample

AS1

AS2

AS3/AT4

FeCr locationa

% FeCr rimb (%)

Secondary magnetite

Fe3+/Cr3+ ratio

MBD2-227 m MBF1-40 m MBF1-47 m MBF1-69 m MBF2-68 m MBF2-72 m MBF2-75 m MBF3-79 m MBF3-97 m

Mod Mod Mod Mod Weak Weak N.p. Mod Mod

Weak Strong Strong Strong Mod Strong N.p. Mod Weak

None V.weak Weak Weak Strong Mod V.Strong Weak Mod

N.p. Plagioclase Plagioclase Plagioclase All All All Plagioclase All

0 2 20 20 90 40 98 30 70

N.p N.p. Weak Weak Mod Weak Strong Weak Weak

N.p. N.a. 0.27 0.30 0.55 0.33 0.74 N.a. N.a.

Refers to the mineral association of chromites with FeCr rims. Refers to the average volume % of FeCr rims on chromite grains.

27

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34 Table 4 Representative EPMA analyses of serpentine aggregates. Type

Mesh-rim

Sample Spot/grain Min. assoc. No. anal.

MBD2-227 m Sp06-C Serp–mag 1

MBD2-227 m Sp06-J Serp–mag 1

Weight percent Na2O MgO Al2O3 SiO2 K2O CaO TiO2 V2O3 Cr2O3 MnO FeOT NiO F Cl Total

0.01 37.73 0.00 39.58 0.01 0.01 0.02 0.01 0.01 0.09 8.50 0.00 0.00 0.13 86.09

0.00 37.18 0.03 38.10 0.02 0.02 0.01 0.00 0.01 0.09 11.91 0.02 0.02 0.11 87.56

Cation proportions Si Ti Al V Cr Fe00 Mg Ca Mn Ni Na K F Cl Mg/(Mg + Fe2)

on basis of 28 oxygen equivalents 7.76 7.52 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 1.40 1.96 11.03 10.93 0.00 0.00 0.02 0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.04 0.04 0.888 0.848

Bastite

Mesh-centre

MBF1-40 m Sp03-A Serp–mag 1

MBF1-40 m Sp03-B Serp–brc 1

MBF1-69 m Sp03-B Serp–brc 1

MBF1-69 m Sp03-C Serp–brc 1

MBF2-68 m Sp01-A Serp 1

MBF2-68 m Sp03-A Serp 1

MBF2-68 m Sp04-A Serp 1

0.01 38.49 0.00 41.74 0.00 0.05 0.01 0.00 0.03 0.08 5.87 0.01 0.48 0.12 86.89

0.02 32.02 1.18 39.73 0.00 0.07 0.08 0.01 0.65 0.36 11.53 0.00 0.36 0.16 86.18

0.01 32.66 0.43 32.33 0.00 0.02 0.02 0.01 0.04 0.26 12.50 0.01 0.24 0.16 78.71

0.01 32.41 0.46 30.49 0.00 0.02 0.02 0.00 0.03 0.25 14.94 0.00 0.21 0.14 78.98

0.00 36.27 2.05 43.08 0.01 0.03 0.02 0.01 0.47 0.05 5.52 0.01 0.03 0.03 87.62

0.01 36.29 1.61 43.42 0.00 0.02 0.01 0.02 0.43 0.03 5.85 0.00 0.05 0.03 87.78

0.00 35.92 2.01 42.92 0.00 0.00 0.00 0.01 0.63 0.02 5.57 0.02 0.03 0.04 87.17

7.96 0.00 0.00 0.00 0.00 0.94 10.94 0.01 0.01 0.00 0.00 0.00 0.29 0.04 0.921

7.89 0.01 0.28 0.00 0.10 1.92 9.48 0.02 0.06 0.00 0.01 0.00 0.23 0.05 0.832

7.22 0.00 0.12 0.00 0.01 2.34 10.87 0.00 0.05 0.00 0.00 0.00 0.17 0.06 0.823

6.92 0.00 0.12 0.00 0.00 2.84 10.97 0.00 0.05 0.00 0.00 0.00 0.15 0.05 0.795

8.08 0.00 0.46 0.00 0.07 0.86 10.14 0.00 0.01 0.00 0.00 0.00 0.02 0.01 0.922

8.14 0.00 0.36 0.00 0.06 0.92 10.14 0.00 0.00 0.00 0.00 0.00 0.03 0.01 0.917

8.10 0.00 0.45 0.00 0.09 0.88 10.10 0.00 0.00 0.00 0.00 0.00 0.02 0.01 0.920

Class 1 mesh-rim

15

Other mesh-rim

Mg+Fe cations per 28 O

Mesh-centre 14

Bastite

Interpenetrating 13

Layer Lz accommodation

Al,Cr substitution

12

Atg 11

10 6.5

7.0

7.5

8.0

8.5

Si cations per 28 O Fig. 6. Plot of (Mg + Fe) cations against Si cations per 28 O-equivalents of the serpentine aggregates analyzed, showing the main causes of variation from the ideal lizardite composition (Lz: filled round symbol): mixing with finely-divided brucite (Brc) or magnetite (Mag), substitution of Si by trivalent ions, and accommodation of layer bending in antigorite (Atg). The mesh-rim serpentine from the least-altered sample (MBD2-227 m) is shown separately from mesh-rim serpentine of other samples.

MBF2-72 m, which are only incipiently affected by Type 2 fracturerelated alteration, the narrow Fe-rich rims retain a significant amount of Al3+ and seem to form a near-continuum of compositions

Interpenetrating

with the cores (Fig. 7 a). Samples MBF2-68 m and MBF2-75 m, both significantly affected by talc–carbonate alteration related to Type 2 veins and Type 3 shearing, have chromites with wide ferritchromite rims that have lost most of their Al3+, and plot on or close to the Cr–Fe3+ join (Fig. 7b and c). The spread of values along the Cr– Fe3+ join of the rims of MBF2-68 m (Fig. 7b) is related to the TiO2 content, which will be discussed below. The ferritchromite rims of the most strongly altered and deformed sample (MBF2-75 m) plot in a tight group closer to the Fe3+ end member than those of MBF2-68 m (Fig. 7b and c). It can be noted that as alteration progresses, the ferritchromite rims become steadily more depleted in Al, and then enriched in Fe3+, to reach values of Fe3+/(Al + Cr + Fe3+) of about 0.5 in the most altered sample. It can be also noted from Fig. 7 that the less altered samples have chromites with more variable ferritchromite rim compositions, whereas the most thoroughly altered sample has chromite rims with much more consistent composition. The similarity of core compositions and distinctness of the altered rim compositions can be clearly seen on cation ratio diagrams such as Fig. 8. From less altered to more altered, the ferritchromite rims show trends of increasing Fe3+/(Al + Cr + Fe3+) and Cr/(Cr + Al) and decreasing Mg/(Mg + Fe2+) ratios, while the core compositions stay more or less the same within the range of compositions of chromite from least-altered samples. The variation between the core compositions of grains hosted in different silicate phases is greater than that between the different altered samples (Fig. 8). The Mg/(Mg + Fe2+) ratios of chromite cores of the most altered sample (Fig. 8c) are at the lower end of the range for Mg/ (Mg + Fe2+) of least-altered samples, but their trivalent ion ratios are normal.

28

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

Table 5 Representative EPMA analyses of chromite grains from weakly altered rock. Sample

MBD2-227 m

MBF1-40 m

Spot/grain Encl. by Size (lm) No. anal.

Gr02-A Olivine 40 1

Gr15-A Plagioclase 500 1

Gr04-A Sulphide 550 1

Gr04-A O’pyroxene 300 1

Gr04-B O’pyroxene 150 1

Gr05-A Plagioclase 90 1

Gr01-M Olivine 60 1

Gr02-R Plagioclase 180 3

Gr03-B Sulphide 190 5

Gr04-A Sulphide 175 2

Weight percent MgO Al2O3 SiO2 CaO TiO2 V2O3 Cr2O3 MnO FeOT CoO NiO ZnO Fe2Ocorr 3 Totala

3.48 19.16 0.05 0.00 1.50 0.74 38.47 0.41 34.16 0.00 0.00 2.08 0.676 100.72

3.83 19.37 0.12 0.00 2.93 0.88 39.60 0.31 32.47 0.00 0.00 0.49 0.332 100.33

4.67 21.92 0.04 0.00 2.04 1.05 40.70 0.25 28.79 0.00 0.03 1.77 0.138 101.38

9.04 19.83 0.09 0.00 0.46 0.38 48.61 0.22 22.36 0.01 0.00 0.07 0.042 101.12

8.58 21.06 0.06 0.00 0.52 0.36 46.87 0.24 23.02 0.00 0.00 0.12 0.031 100.86

5.06 21.83 0.03 0.00 1.83 0.49 40.93 0.28 30.54 0.01 0.00 0.55 0.274 101.82

5.61 21.66 0.04 0.02 0.71 0.41 41.12 0.50 29.50 0.08 0.00 0.39 0.361 100.40

5.01 19.73 0.06 0.03 0.62 0.36 44.11 0.46 29.00 0.06 0.00 0.52 0.21 100.18

4.39 18.75 0.04 0.01 1.15 0.46 42.05 0.49 32.27 0.07 0.00 0.47 0.50 100.67

4.90 18.99 0.03 0.03 1.90 0.49 41.75 0.49 31.32 0.07 0.00 0.48 0.44 100.87

0.087 0.023 5.907 0.077 9.713 0.080 4.646 3.405 0.000 0.048 0.003 0.000 0.014 0.423 0.005 0.622

0.098 0.016 6.275 0.073 9.366 0.058 4.807 3.232 0.001 0.052 0.000 0.001 0.022 0.402 0.004 0.599

0.351 0.007 6.571 0.101 8.266 0.524 5.999 1.926 0.000 0.060 0.002 0.000 0.104 0.243 0.034 0.557

0.138 0.010 6.594 0.085 8.397 0.701 5.671 2.160 0.006 0.109 0.017 0.000 0.074 0.276 0.045 0.560

0.122 0.016 6.097 0.076 9.144 0.408 5.952 1.959 0.009 0.103 0.013 0.000 0.101 0.248 0.026 0.600

0.228 0.011 5.814 0.097 8.747 0.984 6.116 1.722 0.004 0.110 0.016 0.000 0.092 0.220 0.063 0.601

0.373 0.008 5.843 0.102 8.615 0.869 5.967 1.907 0.008 0.108 0.014 0.000 0.092 0.242 0.057 0.596

Cation proportions on basis of 32 oxygen equivalents Ti 0.299 0.580 0.394 Si 0.013 0.032 0.012 Al 5.977 6.004 6.640 V 0.157 0.185 0.216 Cr 8.049 8.236 8.268 000 1.344 0.656 0.267 Fe 00 6.216 6.487 5.920 Fe Mg 1.372 1.501 1.788 Ca 0.000 0.000 0.000 Mn 0.092 0.069 0.054 Co 0.000 0.000 0.000 Ni 0.001 0.000 0.006 Zn 0.406 0.096 0.336 Mg# 0.181 0.188 0.232 Fe3# 0.088 0.044 0.018 Cr# 0.574 0.578 0.555 a

MBF1-47 m

Total includes a correction for Fe2O3 calculated by stoichiometry (Carmichael, 1967).

The trace element diagrams (Fig. 9) show some interesting variations between core and rim compositions depending on the grain’s mineral host and degree of alteration. The cores of chromites from all samples generally have moderate levels of TiO2 (0.5–3.0 wt%), V2O3 (0.35–0.55 wt%) and MnO (0.4–0.55 wt%). In general, the ferritchromite rims have equal or greater abundances of these elements, however, TiO2 can be extremely variable, particularly for chromites hosted in plagioclase and sulphide. The moderately and strongly altered samples (without textural destruction and elimination of serpentine: Fig. 9a and b) have ferritchromite rims that may locally contain levels of TiO2 between 3 and 17 wt%, indicating the presence of exsolved ilmenite lamellae and blebs within the ferritchromite (as has been observed microscopically). These analysis points with excessively high TiO2 values (indicating overlap onto ilmenite blebs) have correspondingly low major element values of Fe3+, Al3+ and Mg resulting in scatter along the Cr3+–Fe3+ join of the ternary (Fig. 7a and b). The core compositions of these grains are relatively low in TiO2 and V2O3 compared to chromite in weakly-altered samples (Fig. 9a and b), suggesting that some of these elements have migrated out of the chromite core at the time of ilmenite bleb formation. (Note that for clarity, analyses with high TiO2 (>3.0 wt%) have been omitted from Fig. 9.) On the other hand, Mn shows much less variation between mineral hosts within a particular sample, implying that Mn has equilibrated between chromite grains on a thin section scale. Mn is always higher in the ferritchromite rim than in the chromite core. The most altered sample that has suffered textural destruction and serpentine breakdown shows much more consistent ferritchromite rim compositions in both major and trace elements (Figs. 8c and

9c), implying that sample-scale chemical equilibrium was reached in this sample.

6. Discussion 6.1. Interpretation of metamorphic sequence and conditions at Muremera B Petrographical study indicates that the Muremera B samples have undergone several alteration or metamorphism events: it is pertinent to ask in what sequence these events occurred and what were their approximate pressure–temperature conditions. Wicks and Whittaker (1977) and Wicks et al. (1977) have discussed the replacement of olivine by serpentine or serpentine–brucite in terms of the retrogression or progression of temperatures, and with or without accompanying strain. They consider that pseudomorphic replacement textures are generally related to retrogressive fluid-ingress without accompanying strain (their Type 1 and Type 3 serpentinization textures). Non-pseudomorphic replacement textures are generally related either to alteration with accompanying strain (Types 2, 4, 6 and 8), or to higher-temperature prograde alteration reactions (Type 5 and 7 of Wicks and Whittaker, 1977). At Muremera B, at least four separate alteration mineral assemblages are recognized. These are firstly defined by their dominant mineralogy (serpentine or talc), and secondly by their textures (pseudomorphic or non-pseudomorphic). These can be listed as follows:

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

Cr

MBF1-47 MBF2-72 MBF1-69 Al

MBF2-68

Fe3+ 50.00

MBF2-75

Cr100.00 80

(a)

Cr 40 Fe3+

Cr 40 Al Cr 80

Ti-rich grains

(b)

Cr 40 Fe3+

Cr 40 Al Cr 80 40.00

60.00

80.00

100.00

(c)

Cr 40 Al

Cr 40 Fe3+

Fig. 7. Plots of trivalent ions in chromite: (a) strongly-altered samples with wellpreserved textures; (b) strongly serpentinized sample with partial textural destruction; (c) strongly talc–carbonate altered sample. Dashed line represents the compositional field of chromites from weakly altered samples MBD2-227 m and MBF1-40 m.

AS1 – pseudomorphic replacement of olivine by lizardite–magnetite, along mesh-rims. AS2 – pseudomorphic replacement of olivine by lizardite–brucite, in mesh-centres. AS3 – non-pseudomorphic replacement of olivine (or earlier serpentine) by antigorite–magnetite, with interpenetrating texture. AT4 – non-pseudomorphic replacement of olivine or earlier serpentine by talc–carbonate–magnetite, with obvious schistose texture. The excellent preservation of igneous olivine textures by both mesh-rim (AS1) and mesh-centre (AS2) serpentinization and their generally widespread distribution at Muremera indicates an early retrogressive hydrous alteration without accompanying strain (Type 3 serpentinization textures of Wicks and Whittaker, 1977). Both AS1 mesh-rim and AS2 mesh-centre serpentines contain appreciable amounts of Cl, indicating either an igneous or a sea-water source for the fluid associated with these alterations. The narrow serpentine–magnetite mesh-rim alteration along olivine rims and fractures evidently predates the alteration of

29

mesh-centres, as these fractures would have been the initial means of entry of hydrous alteration fluid. Also, all olivine-bearing samples contain AS1 mesh-rim alteration, but do not necessarily contain AS2 mesh-centre alteration. AS2 mesh-centre alteration has never been observed alone, without the presence of AS1 mesh-rim alteration. This relationship implies that AS1 mesh-rim alteration predates and is more widespread than AS2 mesh-centre alteration. The AS1 mesh-rim alteration of olivine is texturally associated with narrow reaction rims of tremolite–chlorite on olivine–plagioclase grain boundaries. The generally narrow width of the serpentine mesh-rims and tremolite–chlorite reaction rims (<60 lm) implies a limited fluid supply and therefore a rock-dominated alteration system, but their universal presence in all olivine-bearing samples that have not been subject to later prograde alteration implies a very widespread alteration event. The AS1 serpentine in the mesh-rims is identified by XRD analysis as dominantly lizardite that is associated with magnetite that lines the original olivine fractures. The association of lizardite with magnetite suggests that the reaction occurred in the zeolite/prehnite–pumpellyite facies below about 300 °C (Evans, 1977). The AS2 mesh-centre serpentine alteration affects the centres of olivine grains that have been left unaltered by the event that formed the AS1 assemblage. This is an almost perfectly pseudomorphic alteration, as evidenced by the preservation of very fine (<1 lm) symplectite intergrowths of spinel or magnetite along crystallographic planes of the olivine (Fig. 4b). The mesh-centre alteration is related to very fine, close-spaced through-going veins filled with a similar type of serpentine. These veins also partially affect ortho- and clinopyroxene resulting in bastite alteration. The main characteristic of this serpentinization event, apart from its near-perfect preservation of original magmatic textures, is the green to yellow–green colouration of the serpentine and the absence of magnetite both in the olivine mesh-centres and in the through-going veinlets. Probe analyses of the coloured meshcentre aggregates plot along a mixing line between pure serpentine and a ferroan brucite (Fig. 6). XRD analyses of both olivine meshcentres and bastite after orthopyroxene indicate the dominance of lizardite in this assemblage, indicating a low-temperature of reaction for this, the AS2 event. In the mesh-centre serpentine aggregates, XRD analyses also indicate the presence of minor brucite. The association with brucite, the very-fine grain size of the aggregate and the perfect preservation of igneous textures including micron-scale exsolutions of symplectites confirms that this was a low-temperature event with no associated strain (again in the zeolite/prehnite–pumpellyite facies: Evans, 1977). It can be inferred that the AS2 alteration possibly occurred at a lower temperature than the AS1 alteration. Firstly, where both types occur together, the AS2 alteration does not overprint and recrystallize the earlier AS1 alteration assemblage. They are distinguishable on colour, with AS1 serpentine remaining colourless, whereas the AS2 serpentine aggregate is typically coloured in thin section (Fig. 4b). Secondly, the AS1 serpentine alteration is generally associated with magnetite along the original sites of the fractures, whereas the AS2 serpentine is generally not associated with magnetite. The alteration of an olivine with a composition of Fo88 to a mixture of serpentine and ferroan brucite without magnetite formation implies both low temperatures and low activities for O2 and SiO2 (Frost and Beard, 2007; Beard et al., 2009). This again suggests that alteration occurred at low fluid/rock ratios with intrinsic conditions governed by the rock mass, rather than by incoming fluids. The AS3 alteration assemblage, which is associated with the nearby presence of macroscopic veins or shears, is generally nonpseudomorphic and is characterized by coarser, interpenetrating bundles of a clear serpentine replacing olivine or earlier serpentine assemblages, associated with abundant magnetite. Although the

30

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

0.5

Fe3+/(Fe3++Al+Cr)

(a)

(b)

Sulphide

0.4

(c)

Sulphide

Plagioclase

Plagioclase

Clinopyroxene

Clinopyroxene

Sulphide Plagioclase

0.3 0.2 0.1 0

Cr/(Al+Cr)

0.9 0.8 0.7 0.6 0.5 0

0.1

0.2

0.3

0

0.1

Mg #

0.2

0.3

0

0.1

Mg #

0.2

0.3

0.4

Mg #

Fig. 8. Major elements shown as atomic ratios for Muremera chromites enclosed within interstitial minerals: (a) strongly-altered sample MBF1-47 m with well-preserved textures; (b) strongly-altered sample MBF2-68 m; (c) strongly-altered and sheared sample MBF2-75 m. Low-reflectance cores are shown with solid symbols, their more reflective (ferritchromite) rims shown by unfilled symbols. Analyses with high TiO2 (overlap on ilmenite blebs) have been excluded. The dashed line represents the field of chromite compositions from weakly altered samples MBD2-227 m and MBF1-40 m.

6.0

Sulphide Plagioclase Clinopyroxene

(a)

TiO2 %

5.0

Sulphide Plagioclase Clinopyroxene

(b)

Sulphide Plagioclase

(c)

4.0 3.0 2.0 1.0 0.0

V2O3 %

0.7 0.6 0.5 0.4 0.3 0.2

MnO %

1.0 0.8 0.6 0.4 0.2 0.0

0

0.1

0.2

Mg #

0.3

0

0.1

0.2

Mg #

0.3

0

0.1

0.2

0.3

0.4

Mg #

Fig. 9. Trace elements as weight % oxide values in Muremera chromites enclosed within interstitial minerals: (a) strongly-altered sample MBF1-47 m with well-preserved textures; (b) strongly altered sample MBF2-68 m; (c) strongly altered and sheared sample MBF2-75 m. Low-reflectance cores are shown with solid symbols, their more reflective (ferritchromite) rims shown by unfilled symbols. Analyses with high TiO2 (overlap on ilmenite blebs) have been excluded. The dashed line represents the field of chromite compositions from weakly altered samples MBD2-227 m and MBF1-40 m.

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

gross igneous texture (cumulus olivine shapes) may be preserved on a 1–2 mm scale, the finer textures such as the internal fracture pattern and the crystallographically-controlled exsolution of symplectites are lost in this event. The interpenetrating bundles of serpentine are antigorite, as deduced both from XRD results and from electron probe analyses that have a slight excess of Si cations relative to divalent and trivalent cations, a characteristic of antigorite due to layer accommodation (Wicks and Plant, 1979). Other associated minerals of this event are talc, chlorite, tremolite and ferroan dolomite that replace interstitial clinopyroxene– plagioclase–phlogopite assemblages. This assemblage is characterized by (1) the destruction of primary magmatic and earlier serpentine textures; (2) the replacement of the AS2 lizardite–brucite-assemblage by the AS3 antigorite–magnetite assemblage; (3) partial replacement of interstitial plagioclase– phlogopite by a talc–chlorite–dolomite assemblage; and (4) the presence of ferritchromite rims on all chromite grains. Taken together these features indicate a greenschist facies of metamorphism at a higher temperature than the AS1 and AS2 serpentinization events and with a higher pCO2. The extent of access of fluid in the rock was also much greater than in the AS1 and AS2 events, resulting in the nearly complete breakdown of all primary igneous minerals or their earlier alteration products to equilibrium metamorphic mineral assemblages, with relatively few relics of the earlier events. However, this alteration is relatively localized spatially to within 3–5 m of larger veins and faults, beyond which limit, the earlier AS1 and AS2 alteration assemblages are present and wellpreserved (Fig. 10). A final alteration assemblage (AT4) is recognized that is present only within and immediately adjacent to the larger Type 3 veins and shears. This is characterized by the complete destruction of original magmatic textures at the expense of granoblastic to lepidoblastic (sheared) textures. The rock is uniformly altered to a talc–carbonate–chlorite assemblage with original chromite almost completely altered to ferritchromite to chrome-magnetite. The silicate–carbonate mineral assemblage, the alteration of chromite to ferritchromite and the evidence of ductile shearing all indicate greenschist facies metamorphic conditions (>300 °C) with accompanying deformation. Thus this is a prograde (relative to AS1 and AS2) metamorphic alteration by a CO2-bearing fluid introduced during the shearing/faulting event. It resulted in overprinting and recrystallization of earlier fine-grained alteration assemblages,

Fig. 10. Sketch of the relationship of metamorphic alteration assemblages to associated structural elements within ultramafic rocks at Muremera B intrusion. Fine solid lines denote Type 1 fine veinlets; thick solid lines represent Type 2 extensional veins and minor shears; central diagonal zone represents a Type 3 shear zone; dashed lines represent gradational boundaries between domains of different alteration assemblages, as labelled.

31

complete destruction of primary textures and ultimately the superimposition of schistose textures and homogenization of mineral assemblages at the thin-section and hand-specimen scale. Its association with the same veins and shears as the AS3 event suggests that it is coeval and simply represents the more complete alteration by a late, localized deformation event, mainly involving shearing and access of oxidized, CO2-bearing fluids into the rocks in the greenschist facies. The relationship of the four alteration assemblages/events to each other and to the associated structural elements is schematically illustrated in Fig. 10. 6.2. Correlation of metamorphic events with known regional events Sedimentary rocks adjacent to the Muremera intrusions have undergone thermal metamorphism, resulting in the presence of cordierite and sillimanite within 5 m of the contacts and growth of andalusite prophyroblasts further away. The thermal metamorphism of the sedimentary rocks is due to the intrusion of the Kibaran bimodal igneous suite that includes the Muremera B ultramafic intrusions (Tack and Deblond, 1990). It has been observed that these high temperature minerals are often retrogressed to quartz–sericite–chlorite assemblages in the greenschist facies (Evans et al., 2000). It is likely that the retrogression of these high-temperature hornfels minerals is related to one or more of the alteration events described here. In the absence of obvious through-going veins that may have supplied hydrous fluid to the rock, it is proposed that the olivine mesh-rim and plagioclase reaction rim alteration (AS1 assemblage) was caused by localized migration of deuteric fluid (derived from residual hydrous melt). This fluid, which is likely to have been Cl-bearing, would have initially concentrated in pockets of residual melt (that are now occupied by interstitial phlogopite–hornblende), and migrated locally along plagioclase–olivine and sulphide–olivine grain boundaries, ultimately penetrating within the olivine grains along their cooling fractures. As such, the AS1 alteration event is probably coeval with, or immediately postdates the peak thermal metamorphic event in the sedimentary rocks. The ubiquitous presence and moderate abundance (1–5 modal %) of dark brown igneous phlogopite and pargasitic hornblende in the intercumulus spaces of the ultramafic cumulate rocks indicates that the primary magma contained moderate amounts of fluid, possibly derived from assimilation of sediments (Maier et al., 2010). The derivation of the alteration fluid from the magma itself would explain the presence of Cl in serpentine, the ubiquitous nature of the AS1 alteration, and its limited intensity due to the very low fluid/rock ratios. The AS2 alteration event is strongly associated with closespaced, through-going serpentine veinlets in ultramafic rock, which in general, have a relatively steep orientation parallel to the regional structural fabric. These would have enabled external fluids to access the interior of the intrusions. As argued above, the fluid responsible for the AS2 alteration was both reduced and of low temperature. One possible scenario is that the fluid for the AS2 alteration event was mainly derived from the adjacent hornfelsed sulphidic and graphitic metasediments as hydrothermal fluid convection cells driven by the heat of the intrusion collapsed back into the cooling intrusion itself (Bird et al., 1986; Norton and Taylor, 1979). If this was the case, this event would have followed closely after the AS1 event, being part of the final post-intrusive cooling process. An alternative scenario for the event associated with AS2 alteration is that it occurred much later during the D2 folding event. The steep orientation of the associated serpentine veining appears to coincide with weakly developed cleavage seen in metasediments distal from the intrusions, which is axial-planar with respect to upright F2 folding. This would imply that regional

32

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

metamorphism during D2 deformation occurred within the prehnite–pumpellyite facies at Muremera, and involved the circulation of metamorphic fluids that were derived locally from the adjacent sulphidic and graphitic sediments. This main D2 structural fabric, which is seen throughout the Karagwe–Ankole Belt and is characterized by moderate to tight, upright folding and development of a weak cleavage in sedimentary rocks, is thought to be due to the far-field effects of compression at the margin of the proto-Congo craton to the south, along the Irumide/southern Irumide orogenic belt (1050–1020 Ma; de Waele et al., 2008; Fernandez-Alonso et al., 2012). However, the detailed structural studies that could substantiate this hypothesis in the Muremera area have not yet been carried out. At the Muremera B intrusions, the AS3 and AT4 alteration is distinctly associated with carbonate-filled shear veins that surface mapping and drill core logging has indicated are oriented roughly north–south. Tack (1990) has described the presence of a regionally extensive zone of N–S faulting and thrusting to the west of Muremera. Tack (1990), Tack et al. (2010) and Fernandez-Alonso et al. (2012) have postulated that these N–S faults may be more widespread on the regional scale, and that they were activated during the East African orogenesis (570–530 Ma; de Waele et al., 2008) involving collisions from the east (their D3 deformation event). It is quite likely the N–S faults mapped at Muremera B and seen in drill core are related to these regional-scale N–S faults of the D3 deformation event. It thus seems probable that the AS3/ AT4 alteration is a prograde greenschist facies metamorphism related to CO2-bearing metamorphic fluid ingress during D3 transpressional faulting. The CO2-bearing fluids were most likely derived from the deeper crust and were pumped up by the transpressional tectonics during the East African orogenic event.

6.3. The effect of metamorphism on chromite and implications for petrological studies The most obvious effect of metamorphism on the chromites of the Muremera intrusions has been the development of ferritchromite rims on distinctly zoned chromite grains. Such rims have only been found in those samples showing some evidence of the AS3 and AT4 alteration assemblages. The thickness and composition of the ferritchromite rims depends on the degree of AS3 or AT4 alteration suffered by the sample (Table 3). Those samples that have been affected by AT4 alteration leading to complete recrystallization of early low-temperature serpentine and of relic olivine to the talc–carbonate–chlorite assemblage have chromites that are almost completely replaced by ferritchromite and by chromian magnetite (Table 3 and Fig. 5f). The composition of these ferritchromite rims is consistent from grain to grain within a sample and tends to be closer to the Fe3+ end of the trivalent cation plot (Table 3, Fig. 7c). Those samples with strong development of AS3 alteration assemblages contain chromite grains that are similarly almost completely altered to ferritchromite rims (Fig. 5e), but these ferritchromite rims are more variable and are less Fe-rich in composition (Fig. 7b). On the other hand, those samples with only weak development of AS3 alteration (minor talc-carbonate alteration of interstitial areas) contain chromites within these interstitial areas that have thin ferritchromite rims (Fig. 5c and d), which have very variable, more aluminous compositions (Fig. 7a). Moreover, these rims are only developed on chromite grains hosted within interstitial areas (plagioclase, sulphide and phlogopite) – chromite grains within olivine and in pyroxene oikocrysts in the same sample do not have such alteration rims. The sympathetic variation of the development of ferritchromite rims and the degree of AS3/AT4 alteration is undoubtedly related to the degree of penetration of metamorphic fluids during the D3

deformation event, and hence to the proximity to D3 deformation features such as shear zones and related veins. At Muremera B, there is no evidence for ferritchromite rim development on chromite during either of AS1 or AS2 alteration events (Table 3). All chromite grains that are hosted by either or both of AS1 or AS2 alteration assemblages in either former olivine, pyroxene or interstitial areas are unzoned in reflected light and have typical igneous euhedral to subhedral shapes (Fig. 5a and b). The trivalent ion proportions and most divalent ion proportions in chromite are identical in those samples exposed to significant AS1 and AS2 alteration compared to those samples with minimal alteration in general (Figs. 7 and 8). It can therefore be stated that the AS1 and AS2 low-temperature alteration events have had no effect on shape and texture and only minor effect on the divalent cation composition of the chromites. Therefore chromites found in samples distal from known D3 deformation features (shear zones, faults and related proximal veining) and which do not show any textural overgrowths or distinct metamorphic zoning in reflected light can be treated as having compositions that were determined during igneous and early post-magmatic (cooling) processes. Such chromites can therefore be used for studies of igneous and immediately post-magmatic processes, and would be suitable as indicator minerals for magmatic sulphide mineralization (Averill, 2011; Groves et al., 1983). 6.4. Implications of chromite compositions for igneous and early postmagmatic processes at Muremera B intrusions Chromites in those samples that have undergone only AS1 and AS2 alteration have compositions that are variable depending on their mineral hosts (Tables 5 and A.2 in the Supplementary material). Chromite grains hosted in orthopyroxene oikocrysts are more Mg and Cr-rich and poorer in Ti than those hosted in olivine or in other interstitial minerals. Such variability and the dependence of composition on mineral host is typical for euhedral chromite grains in small to large layered intrusions, and can be explained by both the reaction of these grains with late-crystallizing evolved interstitial melt, and by sub-solidus equilibration of the chromite with enclosing ferromagnesian minerals such as olivine and orthopyroxene (Cameron, 1975; Roeder and Campbell, 1985; Wilson, 1982). The former process leads to enrichment of the chromite in the Fe3+ and Ti4+ components, and is most developed in orthocumulate rocks that contain a high proportion of trapped evolved melt. On the other hand, the process of subsolidus equilibration leads mainly to the decrease of the Mg# ratio in the chromite and leaves the trivalent ions unchanged (Wilson, 1982). At Muremera B, the relatively high contents of TiO2 (>0.70 wt%, Table 5) in all chromite grains except those hosted in oikocrystic orthopyroxene indicates that they have reacted quite extensively with evolved late interstitial melts, which is not surprising as the samples examined are all orthocumulate in nature. Curiously, and unlike the chromite in the majority of layered intrusions where such reactions have been documented (Roeder and Campbell, 1985; Wilson, 1982), at Muremera this enrichment with Ti is not matched by a similar increase in the Fe3+ cation. In fact, the Fe3+# ratio for all Muremera chromites, including the cores of metamorphically modified zoned grains from weakly AS3 altered samples (Fig. 8a) is relatively low (<0.1). This observation was also made for chromite grains from the adjacent Kabanga intrusions in Tanzania by Maier et al. (2010), who postulated that assimilation of sulphidic and graphitic metasedimentary rocks might have drastically reduced fO2 in the parent magma. They also postulate that this assimilation may have either triggered the formation of Ni– Cu sulphides in the magma, or greatly modified their composition. Assimilation of graphitic sedimentary rock and reduction of the magma is a likely explanation for the low Fe3+# ratios in Muremera

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

B chromites, and this feature could be developed as an exploration tool in the search for Ni–Cu sulphide deposits in the intrusions of the KMA in Burundi. 7. Conclusion The small mafic–ultramafic intrusions at Muremera B in eastern Burundi have undergone at least three different alteration events, related to either thermal or deformational episodes affecting the rocks of the Karagwe–Ankolean Belt. The earliest event associated with AS1 assemblages is caused by small-volume deuteric hydration of cumulus olivine and intercumulus plagioclase along grain boundaries and cooling fractures during initial magmatic cooling. A later, low temperature hydration event (AS2 assemblages) related to close-spaced upright serpentine veining resulted in further alteration of olivine and orthopyroxene to a very fine-grained serpentine–brucite intergrowth. The AS2 alteration may been a post-magmatic cooling event, or it may have occurred during late Mesoproterozoic folding. Both AS1 and AS2 events have had little if any effect on chromites in the intrusions. The later event associated with AS3/AT4 assemblages is a prograde greenschist facies hydration and carbonation event localized by north–south oriented faults and associated vein sets. This event probably occurred during intracontinental transpressional tectonics related to the East African Orogeny. The effect of the AS3/AT4 alteration on chromite is significant, resulting in progressive growth of metamorphic ferritchromite rims at the expense of relict igneous cores. These cores remain chemically relatively unaltered, until their complete elimination due to metamorphic equilibration at greenschist facies with high fluid/rock ratios. Acknowledgements I am greatly indebted to Mr. John Spratt and Dr. Jens Najorka of the Imaging and Analysis team of the Natural History Museum, London. They have set up the instrumentation and calibrations on the Cameca SX100 electron microprobe and XRD equipment respectively, and ensured their smooth running. John has guided me in the day to day use of the probe for this study and Jens has carried out all of the analyses on the XRD. I also acknowledge the preparation of superb quality polished thin sections by Tony Wighton of the same team at the NHM. The management of Danyland (Burundi) Ltd. is thanked for originally inviting me to Burundi in 2007 and allowing me to study both the original UNDP drill cores as well as their own cores. I am grateful to Dr. Philippe Muchez (University of Leuven) and Dr. Luc Tack (Royal Museum of Central Africa), who provided thoughtful commentaries on an early draft of the manuscript, and to Dr. J.-C. Duchesne and an anonymous reviewer for their thorough reviews of the submitted manuscript. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jafrearsci.2014. 09.004. These data include Google maps of the most important areas described in this article. References Abzalov, M.Z., 1998. Chrome-spinels in gabbro-wehrlite intrusions of the Pechenga area, Kola Peninsula, Russia: emphasis on alteration features. Lithos 43, 109– 134. Averill, S.A., 2011. Viable indicator minerals in surficial sediments for two major base metal deposit types: Ni–Cu-PGE and porphyry Cu. Geochem.: Explor. Environ., Anal. 11, 279–291. Barnes, S.J., 2000. Chromite in komatiites, II. Modification during greenschist to mid-amphibolite facies metamorphism. J. Petrol. 41, 387–409.

33

Baudet, D., Hanon, M., Lemonne, E., Theunissen, K., 1988. Lithostratigraphie du domaine sédimentaire de la chaîne Kibarienne au Rwanda. Ann. Soc. Géol. Belgique 112, 225–246. Beard, J.S., Frost, B.R., Fryer, P., McCaig, A., Searle, R., Ildefonse, B., Zinin, P., Sharma, S.K., 2009. Onset and progression of serpentinization and magnetite formation in olivine-rich troctolite from IODP Hole U1309D. J. Petrol. 50, 387–403. Bird, D.K., Rogers, R.D., Manning, C.E., 1986. Mineralized fracture systems of the Skaergaard intrusion, East Greenland. Medd. Grønland – Geosci. 16, 68pp. Bouzet, P., 1980. Rapport de Fin de Mission: prospection géologique, géochimique et géophysique dans le secteur de Muremera. Programme des Nations Unies pour le Développement – Projet de Recherches Minières BDI/77-003, 23pp. Buchwaldt, R., Toulkeridis, T., Todt, W., Ucakuwun, E.K., 2008. Crustal age domains in the Kibaran belt of SW-Uganda: combined zircon geochronology and Sm–Nd isotopic investigation. J. Afr. Earth Sc. 51, 4–20. Burkhard, D.J.M., 1993. Accessory chromian spinels: their coexistence and alteration in serpentinites. Geochim. Cosmochim. Acta 57, 1297–1306. Cahen, L., Snelling, N.J., Delhal, J., Vail, J.R., Bonhomme, M., Ledent, D., 1984. The Geochronology and Evolution of Africa. Clarendon Press, Oxford, 512pp. Cameron, E.N., 1975. Postcumulus and subsolidus equilibration of chromite and coexisting silicates in the Eastern Bushveld Complex. Geochim. Cosmochim. Acta 39, 1021–1033. Carmichael, I.S.E., 1967. The iron–titanium oxides of salic volcanic rocks and their associated ferromagnesian silicates. Contrib. Mineral. Petrol. 14, 36–64. Clifford, T.N., 1970. The structural framework of Africa. In: Clifford, T.N., Gass, I.G. (Eds.), African Magmatism and Tectonics. Oliver & Boyd, Edinburgh, pp. 1–26. De Waele, B., Johnson, S.P., Pisarevsky, S.A., 2008. Palaeoproterozoic to Neoproterozoic growth and evolution of the eastern Congo Craton: its role in the Rodinia puzzle. Precambr. Res. 160, 127–141. Deblond, A., 1993. Etude géologique des massifs gabbroïques de l’est du Burundi. Unpubl. PhD Thesis, Université de Liège, vol. 4, 962pp. Deblond, A., Tack, L., 1999. Main characteristics and review of mineral resources of the Kabanga–Musongati alignment in Burundi. J. Afr. Earth Sc. 29, 313–328. Dewaele, S., De Clercq, F., Muchez, P., Schneider, J., Burgess, R., Boyce, A., FernandezAlonso, M., 2010. Geology of the cassiterite mineralisation in the Rutongo area, Rwanda (central Africa): current state of knowledge. Geol. Belg. 13 (1–2), 91–112. Dewaele, S., Henjes-Kunst, F., Melcher, F., Sitnikova, M., Burgess, R., Gerdes, A., Fernandez-Alonso, M., De Clercq, F., Muchez, P., Lehmann, B., 2011. Late Neoproterozoic overprinting of the cassiterite and columbite–tantalite bearing pegmatites of the Gatumba area, Rwanda (Central Africa). J. Afr. Earth Sc. 61, 10–26. Duchesne, J.-C., Liégeois, J.-P., Deblond, A., Tack, L., 2004. Petrogenesis of the Kabanga–Musongati layered mafic–ultramafic intrusions in Burundi (Kibaran Belt): geochemical, Sr–Nd isotopic constraints and Cr–Ni behaviour. J. Afr. Earth Sc. 39, 133–145. Dwyka Resources, 2007. Muremera Nickel Project Update. Company Press Release Dated 30 July 2007. (27.08.14). Dwyka Resources, 2008. Progress at Muremera Nickel Project. Company Press Release Dated 24 April 2008. (27.08.14). Evans, B.W., 1977. Metamorphism of alpine peridotite and serpentinite. Annu. Rev. Earth Planet. Sci. 5, 397–447. Evans, B.W., Frost, B.R., 1975. Chrome-spinel in progressive metamorphism – a preliminary analysis. Geochim. Cosmochim. Acta 39, 959–972. Evans, D.M., Boadi, I., Byemelwa, L., Gilligan, J.M., Kabete, J., Marcet, P., 2000. Kabanga magmatic nickel sulphide deposits, Tanzania – morphology and geochemistry of associated intrusions. J. Afr. Earth Sc. 30, 651–674. Fernandez-Alonso, M., Cutten, H., de Waele, B., Tack, L., Tahon, A., Baudet, D., Barritt, S.D., 2012. The Mesoproterozoic Karagwe–Ankole Belt (formerly the NE Kibara Belt): the result of prolonged extensional intracratonic basin development punctuated by two short-lived far-field compressional events. Precambr. Res. 216–219, 63–86. Frost, B.R., Beard, J.S., 2007. On silica activity and serpentinization. J. Petrol. 48, 1351–1368. Grey, I.M., 1967. Geological Map with Explanation, Quarter Degree Sheet 29 and 29W, Ngara; 1:125,000. 1967. Compiled by Grey, I.M. Mineral Resources Division, Dodoma, Tanzania. Groves, D.I., Barrett, F.M., Brotherton, R.H., 1983. Exploration significance of chrome-spinels in mineralized ultramafic rocks and nickel–copper ores. Spec. Publ. Geol. Soc. South Africa 7, 21–30. Irvine, T.N., 1965. Chromian spinel as a petrogenetic indicator. Part 1, Theory. Can. J. Earth Sci. 2, 648–672. Karayenga, D., 1987. Feuille Ruyigi S4/30 – NW Geological map at 1: 100,000 scale published by Département de Géologie et de Minéralogie du Musée royal de l’Afrique centrale (Tervuren, Belgique) and Ministère des Travaux Publics, de l’Energie et des Mines du Burundi. Kimball, K.L., 1990. Effects of hydrothermal alteration on the compositions of chromian spinels. Contrib. Mineral. Petrol. 105, 337–346. Klerkx, J., Liegeois, J.-P., Lavreau, J., Claessens, W., 1987. Crustal evolution of the northern Kibaran Belt, Eastern and Central Africa. In: Kröner, A. (Ed.), Proterozoic Lithospheric Evolution. American Geophysical Union, Washington, pp. 217–233. Koegelenberg, C., Kisters, A.F.M., 2014. Tectonic wedging, back-thrusting and basin development in the frontal parts of the Mesoproterozoic Karagwe–Ankole belt in NW Tanzania. J. Afr. Earth Sc. 97, 87–98.

34

D.M. Evans / Journal of African Earth Sciences 101 (2015) 19–34

Kokonyangi, J., Kampunzu, A.B., Armstrong, R., Yoshida, M., Okudaira, T., Arima, M., Ngulube, D.A., 2006. The Mesoproterozoic Kibaride belt (Katanga, SE D. R. Congo). J. Afr. Earth Sc. 46, 1–35. Maier, W.D., Barnes, S.-J., 2010. The Kabanga Ni sulfide deposits, Tanzania: II. chalcophile and siderophile element geochemistry. Miner. Deposita 45, 443– 460. Maier, W.D., Barnes, S.-J., Sarkar, A., Ripley, E., Li, C., Livesey, T., 2010. The Kabanga Ni sulfide deposit, Tanzania: I. Geology, petrography, silicate rock geochemistry, and sulphur and oxygen isotopes. Miner. Deposita 45, 419–441. Merlini, A., Greico, G., Diella, V., 2009. Ferritchromite and chromian-chlorite formation in mélange-hosted Kalkan chromitites (Southern Urals, Russia). Am. Mineral. 94, 1459–1467. Muhagaze, L., 1984. Résultats des forages dans les zones B et C du périmètre de Muremera. Unpublished report for République du Burundi: Projet de recherches et développement minier (BDI/81-007), 16pp. Niyondezo, S., Rukwaya, A., Muhagaze, L., 1982. Rapport sur les recherches géologiques effectuées sur l’indice de mineralisation sulfurée à Muremera – Zone A. Unpublished report for République du Burundi: Projet de recherches et développement minier (BDI/81-007), 22pp. Norton, D., Taylor, H.P., 1979. Quantitative simulation of the hydrothermal systems of crystallizing magmas on the basis of transport theory and oxygen isotope data: an analysis of the Skaergaard Intrusion. J. Petrol. 20, 421–486. Pisarevsky, S.A., Elming, S.-Å., Pesonen, L.J., Li, Z.-X., 2014. Mesoproterozoic paleogeography: supercontinent and beyond. Precambr. Res. 244, 207–225. Pohl, W.L., Biryabarema, M., Lehmann, B., 2013. Early Neoproterozoic rare metal (Sn, Ta, W) and gold metallogeny of the Central Africa Region: a review. Appl. Earth Sci. (Trans. Inst. Min. Metall. B) 122, 66–82. Roeder, P.L., Campbell, I.H., 1985. The effect of postcumulus reactions on compositions of chrome-spinels from the Jimberlana intrusion. J. Petrol. 26, 763–786. Rumvegeri, B.T., 1991. Tectonic significance of Kibaran structures in Central and Eastern Africa. J. Afr. Earth Sc. 13, 267–276. Siivola, J., Schmid, R., 2007. A Systematic Nomenclature for Metamorphic Rocks: 12. List of Mineral Abbreviations. Recommendations by the IUGS Subcommission

on the Systematics of Metamorphic Rocks. (21.04.14). Sintubin, M., 1989. Characterization of the Kibaran metamorphism in the Kazingwe Complex, SW Burundi. Musée royale d’Afrique centrale, Tervuren (Belg), Département de Géologie et Mineralogie. Rapport Annu. 1987–1988, 123–137. Spangenberg, K., 1943. Die chromitlaagerstatte von tampedal in Zobten. Z. praktische Geol. 51, 13–35. Tack, L., 1990. Late Kibaran structural evolution in Burundi. I.G.C.P. No. 255 Newslett./Bull. 3, 77–79. Tack, L., Deblond, A., 1990. Intrusive character of the Late Kibaran magmatism in Burundi. I.G.C.P. No. 255 Newslett./Bull. 3, 81–87. Tack, L., Liègeois, J.P., Deblond, A., Duchesne, J.C., 1994. Kibaran A-type granitoids and mafic rocks generated by two mantle sources in a late orogenic setting (Burundi). Precambr. Res. 68, 323–356. Tack, L., Wingate, M.T.D., de Waele, B., Meert, J., Belousova, E., Griffin, B., Tahon, A., Fernandez-Alonso, M., 2010. The 1375 Ma ‘‘Kibaran event’’ in Central Africa: prominent emplacement of bimodal magmatism under extensional regime. Precambr. Res. 180, 63–84. Waleffe, A., 1981. Feuille Cankuzo S4/30 – NE et S3/30 – SE (pars). Geological map at 1:100,000 scale published by Département de Géologie et de Minéralogie du Musée royal de l’Afrique centrale (Tervuren, Belgique) and Ministère des Travaux Publics, de l’Energie et des Mines du Burundi. Wicks, F.J., Plant, A.G., 1979. Electron-microprobe and X-ray-microbeam studies of serpentine textures. Can. Mineral. 17, 785–830. Wicks, F.J., Whittaker, E.J.W., 1977. Serpentine textures and serpentinization. Can. Mineral. 15, 459–488. Wicks, F.J., Whittaker, E.J.W., Zussman, J., 1977. An idealized model for serpentine textures after olivine. Can. Mineral. 15, 446–458. Wilson, A.H., 1982. The geology of the Great ‘Dyke’, Zimbabwe: the ultramafic rocks. J. Petrol. 23, 240–292. Wylie, A.G., Candela, P.A., Burke, T.M., 1987. Compositional zoning in unusual Znrich chromite from the Sykesville district of Maryland and its bearing on the origin of ‘‘ferritchromit’’. Am. Mineral. 72, 413–422.