Neoproterozoic passive margin formation and evolution during the Rodinia–Gondwana supercontinent cycle at the eastern margin of the West African Craton

Abstract

Petrographical and geochemical data from the Togo structural unit (TSU), also referred to as the Atacora structural unit, are presented together with the existing dataset; geochemical and age data from the sedimentary and metasedimentary rocks from the passive margin sequences of the Dahomeyide belt in Ghana to infer their provenance and depositional setting and expand the discussion on the Rodina–Gondwana supercontinent assembly during the Pan-African orogeny. The metasedimentary rocks of the TSU are quartzites and phyllites. The framework grains of the quartzites consisting dominantly of quartz and small amounts of feldspar grains and relict lithic fragments classify them as quartz arenite, subarkose and sublitharenite. Generally, the studied rocks show similar rare-earth element and multi-element patterns, which imply derivation from similar sources. Elemental ratios, including (La/Lu)_N, Th/Sc and La/Sc, suggest sediments sourced from intermediate to felsic rocks. Provenance and depositional setting indicators of the TSU suggest deposition in a passive margin setting, with the West African and Amazonian cratons’ granitoids and granitic gneisses as possible provenance, akin to siliciclastic rocks of the Buem structural unit and the Voltaian Supergroup of the Volta Basin. The deformational history of the TSU is similar to those of the Buem structural unit and the eastern margin of the Voltaian Supergroup, indicating the effect of the Pan-African orogeny on the passive margin of the Dahomeyide belt. We, therefore, propose the formation and evolution of a Neoproterozoic passive margin unit, which was tectonically deformed during the Rodinia–Gondwana supercontinent cycle.

1. Introduction

The tectonic setting, proto-source rock composition, paleoclimatic and paleoweathering conditions of siliciclastic sedimentary rocks are all revealed by the petrographic and geochemical data of these rocks (Bhatia & Taylor, 1981; Dickinson, 1985; Verma & Armstrong-Altrin, 2016). The major and trace element compositions of siliciclastic rocks manifest the composition of the source rocks since certain trace elements, like the rare-earth elements (REEs) and high-field strength elements (HFSE), are usually not mobile during weathering and transportation and stay as solid (particulate) load, i.e., reflecting the source composition (Bhatia & Taylor, 1981; Verma & Armstrong-Altrin, 2016). When eroded, unique siliciclastic sedimentary rocks with distinctive mineralogical compositions and textural features are produced because different tectonic settings are defined by different rock types (Dickinson, 1985). The youngest zircon age population in siliciclastic sedimentary rocks can be used to infer the maximum age of deposition of the sediments into the sedimentary basin (Fedo et al., 2003; Yao et al., 2011). History of crustal growth and evolution, correlations of continental blocks within supercontinents and reconstruction of palaeogeography of geological terranes can be constrained from detrital zircon ages of sedimentary basins (Yao et al., 2011; Cawood et al., 2012; Andersen et al., 2016). Combining petrography and whole-rock geochemical data together with detrital U-Pb zircon ages, the provenance of siliciclastic sedimentary rocks in several basins and their depositional setting have been inferred (Anani et al., 2017, 2019; Baiyegunhi et al., 2017; Jiang et al., 2017; Kwayisi et al., 2022b).

The Dahomeyide belt is a Neoproterozoic orogenic belt that forms a part of the West Gondwana Orogen at the south-eastern margin of the West African Craton (WAC) (Fig. 1a and 1b, e.g., Attoh, 1998; Attoh & Nude, 2008; Kwayisi et al., 2020). The Dahomeyide belt consists of key features of typical collisional belts such as passive margin sequences, oceanic terrane, active margin (i.e. arc crust), post-collisional basins and basement complexes (Fig 1b; Ganade de Araujo et al., 2014a; Kwayisi et al., 2020). This Dahomeyide belt stretches ⁓1000 km N-S and 400 E-W from Ghana, Togo and Benin to Nigeria (Agbossoumonde et al., 2001; Attoh & Nude, 2008; Nude et al., 2015). Three main nappes zones comprise the Dahomeyide belt, and these are (i) internal nappes zone, (ii) suture zone and (iii) external nappes zone (Attoh and Nude, 2008; Nude et al., 2015; Aidoo et al., 2020; Kwayisi et al., 2022a). Occurring in the external nappes zone of the Dahomeyide belt are siliciclastic rocks of the Buem structural unit and Togo structural unit (TSU) (Osae et al., 2006; Attoh & Nude, 2008; Kwayisi et al., 2020, 2022b), also known as the Atacora structural unit (Affaton et al., 1997; Alayi et al., 2023) of foreland and passive margin affinities evident from geochemical and geochronological data (Kalsbeek et al., 2008; Ganade de Araujo et al., 2016; Kwayisi et al., 2022b). The Amazonian Craton, Benin-Nigerian Shield, ± West African Craton have been proposed as the possible sources of sediments for these sequences (Kalsbeek et al., 2008; Ganade de Araujo et al., 2016; Anani et al., 2019; Kwayisi et al., 2022b). Kalsbeek et al. (2008) and Kwayisi et al. (2022b) indicated that passive margin formation may have been deposited c. 1000 – 700 Ma. In contrast, the deposition of foreland sequences occurred between 620 and 570 Ma, according to Ganade de Araujo et al. (2016). The youngest detrital zircon for passive margin sequences is at 900 Ma, whereas that for foreland sequences is at 600 Ma (Ganade de Araujo et al., 2016; Kwayisi et al., 2022b).

Figure 1. (a) A schematic map of Africa and South America showing the various cratons and Pan-African orogenic belts and (b) geological map of West Gondwana orogen (modified after Kwayisi et al., 2022a).

The TSU, which predominantly consisted of sandstones and shales, is now metamorphosed into quartzites, phyllites and schists (Kalsbeek et al., 2008; Ganade de Araujo et al., 2016; Anani et al., 2019). The geochemical characteristics of the phyllites are akin to passive margin sequences of the external nappes zone of the Dahomeyide belt (Anani et al., 2019). A passive margin setting of deposition was proposed for the TSU because of the 2200 – 930 Ma ages (U-Pb detrital zircon obtained for the quartzite (Kalsbeek et al., 2008). In addition, Ganade de Araujo et al. (2016), based on a zircon age population between 900 and 600 Ma (U-Pb detrital zircon) from the schist, inferred a foreland basin. The available zircon U-Pb populations for the TSU are comparable to the siliciclastic rocks of the external nappes zone of the Dahomeyide belt and the Borborema Province in Brazil (Kwayisi et al., 2022b). Nonetheless, no published petrographical and geochemical data are available on the quartzites and schists of the TSU to infer their depositional setting. Thus, petrography, major and trace element data of quartzites, together with new major and trace element data of the phyllites from the TSU, are presented in this study to discuss their provenance and depositional setting. The results of this study are compared to available petrographical, geochemical and geochronological data from the passive margin units of the Dahomeyide belt (i.e., Buem structural unit and the Voltaian Supergroup of the Volta Basin) to provide constraints for the development of passive margin basins and their evolution during the Pan-African orogeny and expand the discussion on the position of West African–Amazonian cratons in supercontinent Rodina break-up and West Gondwana assembly.

1.a. Geological setting and previous work

Occurring at the eastern margin of the WAC is the Neoproterozoic Dahomeyide belt formed during the Pan-African orogeny (Attoh & Nude, 2008; Aidoo et al., 2020; Kwayisi et al., 2022a; 2022b). The WAC consists of a western Archean basement comprising granitoid-greenstone belts of ages between 3500 and 2500 Ma, and an eastern Paleoproterozoic Birimian domain also of granitoid-greenstone belts with ages in the range of 2200 - 1800 Ma (e.g. Kouamelan et al., 1997; Potrel et al., 1998; Egal et al., 2002; Thiéblemont et al., 2004; Sakyi et al., 2014, 2018; 2020; Anum et al., 2015; Grenholm et al., 2019; Nunoo et al., 2022; Kazapoe et al., 2022; Amponsah et al., 2023). Associated with the granitoid-greenstone belts are metasedimentary basins of Paleoproterozoic age (Asiedu et al., 2017; Sakyi et al., 2019; Kazapoe et al., 2023). In most of the Rodinia supercontinent reconstruction (1000 – 1200 Ma), the Amazonian Craton is located near the WAC (e.g. Evans, 2009, 2013). Three main Archean blocks, dated at 3300 – 2600 Ma, were stabilized around 2100 Ma during the Transamazonian orogeny (2180 – 1950 Ma), which is the time equivalent to the Ebunean orogeny in the WAC (Ledru et al., 1994; Tassinari et al., 2001; Lofan et al., 2003; Cordani et al., 2009). Occurring abundantly on the Amazonian Craton in contrast to the WAC are granitoid and granitoid gneisses of late Paleoproterozoic and Mesoproterozoic (1700 – 1000 Ma) ages (Tassinari et al., 2000; Santos et al., 2008; Cordani et al., 2009).

As stated earlier, the Dahomeyide belt consists of three main zones. These are (i) the external nappes zone, (ii) internal nappes zone and (iii) a well-defined suture zone (Fig. 2; Affaton et al., 2000; Attoh & Nude, 2008). The Benino-Nigerian Shield’s Paleoproterozoic (2190 – 2140 Ma, U-Pb zircon ages; Attoh et al., 2013; Kalsbeek et al., 2020) granitoids and gneisses comprise the internal nappes zone and are intruded by juvenile magmatic arcs and post-collisional plutons at ages of 670 – 610 Ma (Attoh et al., 2013) and 580 – 540 Ma (Kalsbeek et al., 2012; Alayi et al., 2023), respectively. The Dahomeyide belt’s upper plate and active margin (i.e. the internal nappes zone) experienced widespread migmatization and granitoid intrusion during mid-Cryogenian to early-Ediacaran (Fig. 2; Kwayisi et al., 2023). A well-defined suture zone of high-pressure metamorphic rocks, including eclogite and granulite, can be found west of the internal nappes zone (Attoh, 1998; Agbossoumonde et al., 2004; Attoh & Morgan, 2004; Duclaux et al., 2006; Guillot et al., 2019). Peak metamorphism was recorded at 610 Ma (U-Pb zircon dates; Attoh et al., 1991; Affaton et al., 2000). The end of collision is marked by the exhumation of the high-pressure rocks at 600 – 580 Ma (U-Pb zircon age; Attoh et al., 1997).

Figure 2. Schematic cross-section of the Dahomeyide belt from west to east showing the relationship between the external nappes, suture and internal nappes zones (modified after Kwayisi et al., 2022b). ¹Clauer (1976); ²Clauer et al. (1982); ³Ganade de Araujo et al. (2014a); ⁴Sakyi et al. (2014); ⁵Amponsah et al. (2023); ⁶Kalsbeek et al. (2012); ⁷Attoh et al. (2013).

The basement Ho-gneisses and passive and foreland sequences of the Buem – TSUs with their lateral equivalence, the Voltaian Supergroup, make up the Dahomeyide belt’s external nappes zone (Kalsbeek et al., 2008; Aidoo et al., 2014; Kwayisi et al., 2022b). Evidence from airborne geophysical and structural interpretations, as well as tectonic windows of Paleoproterozoic gneisses within the TSU, suggest the underthrust deformed WAC crust as the basement rocks of the external nappes zone (Kwayisi et al., 2020; Aidoo et al., 2020). The Voltaian Supergroup, which occupies the Volta Basin, is stratigraphically divided into three groups (Agyei-Doudu et al., 2009). At the base is the Kwahu-Bombouaka, followed upwards by the Oti-Pendjari, with the Tamale-Obosom at the top of the supergroup (Fig. 3; Table 1; Deynoux et al., 2006; Agyei-Doudu et al., 2009; Kwayisi et al., 2022b). Different periods of deposition in two main depositional settings have been proposed for the Voltaian Supergroup from available petrographical, geochemical and geochronological data (Table 1; Kwayisi et al., 2022b). Deposition in a passive margin basin has been proposed for the Kwahu-Bombouaka Group at 959 ± 62 Ma based on Rb-Sr dating on clay minerals (Clauer, 1976; Carney et al., 2010; Anani et al., 2017). The Oti-Pendjari and Tamale-Obosom groups were deposited in a foreland basin setting, although the lower part of the Oti-Pendjari Group shows passive margin signatures (Table 1; Amedjoe et al., 2018). The depositional age of the Oti-Pendjari Group is between 635 and 576 Ma (Barford et al., 2004; Potter et al., 2004); however, that for the Tamale-Obosom Group is unknown.

Figure 3. Geological map of the Dahomeyide belt, showing the lithological distribution of the Buem and TSUs and their relationship to the Voltaian Supergroup of the Volta Basin (Modified after Ganade de Araujo et al., 2016).

Table 1. A summary of data pattern across the external nappes zone of the Dahomeyide belt (modified after Kwayisi et al., 2022b)

¹Kalsbeek et al. (2008), ²Carney et al. (2010),³Ganade de Araujo et al. (2016), ⁴Barfod et al. (2004), ⁵Amedjoe et al. (2018), ⁶Anani (1999), ⁷Anani et al. (2013), ⁸Anani et al. (2017), ⁹Clauer (1976), ¹⁰Osae et al. (2006), ¹¹Kwayisi et al. (2022b) ¹²Clauer et al. (1982), ¹³Anani et al. (2019) and ¹⁴Ganade de Araujo et al. (2014b). WAC = West African Craton, BNS = Benino-Nigerian Shield.

Kwayisi et al. (2020; 2022b) indicated that the lower part of the Buem structural unit consists of siliciclastic sedimentary rocks made up of meta-sandstones and shales that progress upwards to exhumed mantle peridotites, mafic plutonic and volcanic rocks and chert/jasper, with the upper part being siliciclastic sedimentary rocks. Field structural studies and airborne geophysical data interpretation have identified three main deformation events (D1 – D3) in the Buem structural unit (Kwayisi et al., 2020). The most widespread is D2, which affected all the rocks. It is characterized by NNW-SSE-striking S2 foliations and F2 isoclinal and chevron folds with an E-dipping axial plane. It also contains L2 downdip stretching lineation. According to Kwayisi et al. (2020), D2 can be observed in the rocks of the TSU and Kwahu-Bombouaka Group. This was a result of the E-W shortening during the Pan-African subduction-collision events between 620 and 605 Ma. Three groups of siliciclastic sedimentary rocks with U-Pb detrital zircon populations of 2300–1800 Ma, 1700–1100 Ma and 1000–970 Ma exist in the Buem structural unit and are interpreted to suggest passive margin sequences from two potential sources; the Amazonian and West African cratons (Table 1; Kalsbeek et al., 2008; Kwayisi et al., 2022b).

The TSU occur to the east of the Buem structural unit with a tectonic contact and overlies the Ho-gneisses. It stretches from the Ghanaian coast in the southeast, through Togo to Benin, approximately 1000 km long and 10 – 50 km wide (Adjei & Tetteh, 1997; Affaton et al., 1997; Figs. 2 and 3). Quartzites, phylites and schists are the principal units of the TSU (Fig. 3; Table 1; Adjei & Tetteh, 1997). Locally, small occurrences of chert, jasper and serpentinite also occur (Junner, 1935). Rocks typically strike N-S in the northern portion of the TSU, whereas a NE-SW strike is recorded for those in the southern portion (Attoh et al., 1997). Adjei and Tetteh (1997) have identified three different types of folds. These folds are isoclinal, open and recumbent, of which isoclinal is widespread, similar to the Buem structural unit and the eastern margin of the Voltaian Supergroup (Kesse, 1985; Kwayisi et al., 2020). Two metamorphic conditions have been reported for the TSUs, and these are upper greenschist facies at the western portion and lower amphibolite at the eastern portion (Wright et al., 1985; Attoh et al., 2007). Ages obtained for the TSU increase northwards with younger ages of 579.4 ± 0.8 Ma (⁴⁰Ar/⁴⁰Ar studies of muscovite in the quartzite in southern Ghana), while in the north that is, Togo and Benin republics older ages of 608.1 ± 0.2 Ma have been recorded (Attoh et al., 1997). The 608.1 ± 0.2 Ma age corresponds to the formation of the N-S fabrics due to an E-W shortening, whereas the 579.4 ± 0.8 Ma age corresponds to a later NW-SE thrust to produce the NE-SW fabrics (Attoh et al., 1997).

Mineralogical and geochemical studies of the phyllites indicate a passive margin setting of deposition as compared to the siliciclastic rocks of the Buem structural unit and lower Voltaian Supergroup (i.e., Kwahu-Bombouaka Group) (Table 1; Anani et al., 2019; Kwayisi et al., 2022b). Besides, a passive margin setting of deposition was inferred by Kalsbeek et al. (2008) because of U-Pb detrital zircon ages population between 2200 and 930 Ma of the quartzite in the TSU which accumulated sediments mainly from the West African–Amazonian cratons (Kalsbeek et al., 2008). Also, detrital zircon grains in sediments related to the active margin of the Dahomeyide belt vary from 781 to 617 Ma (Ganade de Araujo et al., 2016). Given that, Kwayisi et al. (2022b) correlated the TSU, Buem structural unit and the Kwahu-Bombouaka Group to have deposited in a similar depositional (passive margin) setting with the same provenance. Petrographical and geochemical data, which can better constrain the provenance and depositional setting of the quartzite of the TSU, are generally lacking. Thus, there is a need to carry out comprehensive petrographical and geochemical studies of the quartzites of the TSU to evaluate their provenance and depositional setting and compare the result to the siliciclastic rocks of the Buem structural unit and Kwahu-Bombouaka Group.

1.b. Field relations, petrography and structures

Rocks of the TSU encountered in the field are quartzites and phyllites, which form several broadly NNE-SSW to North-South-trending hills and valleys. The quartzites dominate over the phyllites in terms of volume. Two types of quartzite occur in the field: foliated and massive quartzites (Fig. 4a and 4b). The foliated quartzite shows both thick and thin foliations, with the foliation striking generally N-S in the northern section and NE-SW in the southern section with moderate to steep dips to the East or SE.

Figure 4. Field photos of (a) foliated quartzite and (b) massive quartzite, photomicrographs of (c) foliated quartzite showing elongated quartz grains in a preferred orientation, (d) foliated quartzite showing proto-mylonitic texture, (e) massive quartzite and (f) massive quartzite showing cataclastic texture. All photomicrographs were taken in crossed polars. Qz = quartz, Ms = muscovite and Lt = lithic fragment. Mineral abbreviations are from Dickinson et al. (1983) and Whitney and Evans (2010).

Medium to coarse-grained characterized the foliated quartzites. They are dominantly composed of quartz with a minor amount of mica and feldspar (Fig. 4c). The grains are sub-rounded to rounded, sutured, with few being slightly elongated and preferred oriented. Few are highly strained, showing mylonitic texture (Fig. 4d). The quartz grains are stained with undulose extinction. Trace amounts of relict lithic fragments, mostly metamorphic materials and micas, also occur. The massive quartzites are composed almost entirely of course-grained quartz grains with small amounts of feldspar, mica grains and lithic fragments (Fig. 4e) and sometimes show cataclastic texture (Fig. 4f). The phyllites observed are of two varieties: the dark grey and whitish phyllites (Fig. 5a). In general, the phyllites are fine-grained with a silky sheen lustre. They contain clast of quartz, mica, chlorite and carbonaceous matter (mostly in the dark grey variety) minerals (Fig. 5b). Dominant structures observed in the phyllites include crenulations, cleavages and joints. Occasionally, the phyllites and foliated quartzite are intercalated (Fig. 5a). Table 2 presents the modal composition of the quartzites. The framework grains of the studied quartzites, which are characterized by dominant quartz grains with a minor amount of feldspars and relict lithic fragments, are classified as quartz arenite and sublitharenite with one sample each plotting as subarkose and arkose (Fig. 6).

Figure 5. (a) Field photo of phyllite interbedded in foliated quartzite and (b) photomicrographs of phyllite showing quartz–seritcite–chlorite mineral assemblage. The photomicrographs were taken in cross-polars. Qz = quartz, Pl = plagioclase, Chl = chlorite and Lt = lithic fragment. Mineral abbreviations are from Dickinson et al. (1983) and Whitney and Evans (2010).

Table 2. Mineralogical compositions of the quartzites of the TSU

Figure 6. QFL diagram (Dickinson et al., 1983) for the quartzites of the TSU compared with the sandstones of the Buem structural unit (Osae et al., 2006; Kwayisi et al., 2022b) and the Kwahu-Bombouaka Group of the Voltaian Supergroup (Anani et al., 2017). Q = quartz, F = feldspar and L = lithic fragment (excluding polycrystalline quartz).

The TSU rocks are folded with the folding expressed as recumbent, isoclinal and crenulations (Fig. 7a and 7b). Foliation, folds, lineation, faults, joints and thrust contacts define three main deformational events the TSU’s rocks have undergone. D1, which is expressed by S1 tectonic foliation, is not well preserved and has a partially or wholly transposed S0 bedding plane. S1 is axial planar to F1 recumbent folds, which strikes 030-040° NE and dips between 04-05° SE (Fig. 7a). F1 have an east plunging B1 fold axis with near horizontal plunges.

Figure 7. (a) F1 recumbent fold with near horizontal axial plane, (b) partially transposed S1 foliation by D2 event and the development of S2 foliation, (c) down-dip stretching L2 lineation on S2 foliation plane and (d) photomicrograph showing the relationship between S1 and S2 foliation plane in microscopic view (crossed polars).

The dominant deformational event is the D2, which is widespread throughout the entire TSU. D2 is characterized by S2 tectonic foliations, F2 isoclinal and crenulation folds (that are cylindrical at the mesoscale; Fig. 7b). The crenulation is observed mainly in the phyllites. S2 strikes 340–358° NNW-SSE and dips between 45° and 70° to the east, and this is akin to rocks of the Buem structural unit (Kwayisi et al., 2020). F2 has a west-verging axial plane and an N-plunging B2 fold axis. Contained in the S2 foliation are downdip L2 stretching lineations (Fig. 7c). In thin-section, the S2 is expressed as the elongation and preferred orientation of quartz, mica and chlorite in the quartzite and phyllites (Fig. 7d). Late joints and faults characterized the D3 deformation (Figs. 4a, 4b, and 7c).

1.c. Analytical procedure

A total of 51 rock samples were analysed for their major and trace element compositions at the ALS laboratory in Vancouver, Canada. The samples included 42 quartzites and 9 phyllites. The major elements were analysed using Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES), while the trace elements were analysed using multi-element fusion Inductively Coupled Plasma Mass-Spectrometry (ICP-MS). The analytical procedure for the major and trace elements analyses is described in Nude et al. (2015) and Kwayisi et al. (2017). Fusing in a furnace of approximately 0.200 g of the prepared sample with lithium at 1025°C was done for the purpose of major element analysis. The resulting melt was then dissolved in an acidic solution consisting of Nitric (NH₃) + hydrochloric (HCl) and hydrofluoric (HF) acids after being cooled. The solution was analysed for the concentrations of major elements using ICP-AES. With a temperature of 1000°C, loss on ignition was determined. Similar procedures were used for the trace element analysis. However, in this instance, 0.100 g of the prepared sample was weighed, and ICP-MS was used for the analyses. ICP-AES was used to analyse the base metals after the 0.25 g prepared sample was digested with perchloric, NH₃, HCl and HF acids. The residue was then treated with diluted HCl, and the ICP-AES was used to analyse the solution. The results were corrected for spectral inter-element interferences (ALS laboratory). Precision and accuracy are better than 3% for the major elements and 10% for the trace elements.

2. Results

2.a. Major elements

The studied quartzites have SiO₂ content in the range of 82.07 – 99.0 wt % (Table 3). These SiO₂ values are higher than those of the phyllites, which have values between 70.50 and 89.20 wt %. Al₂O₃ is higher in the phyllites (Al₂O₃ = 6.68 – 16.37 wt %, avg. 12.23 wt %) and lower in the quartzites (Al₂O₃ = 0.22 – 7.98, avg. 1.91 wt %). The ferromagnesian elements are relatively higher in the phyllites (Fe₂O₃t = 1.24 – 5.94, avg. 2.59 wt % and MgO = 0.37 – 1.81, avg. 0.74 wt %) than the quartzites (Fe₂O₃t = 0.31 – 2.90, avg. 0.96 wt % and MgO = 0.01 – 0.32, avg. 0.09 wt %). K₂O content varies with the higher content observed in the phyllites (2.14 – 5.22, avg. 3.46 wt %). K₂O contents of the quartzites are low in the range of 0.04 – 3.24, avg. 0.47 wt %. Major elements ratios of Al₂O₃/TiO₂ are nearly similar for all the studied rocks of the TSU and akin to Upper Continental Crust (UCC) but slightly higher than Post Archean Australian Shale (PAAS) (Table 3).

Table 3. Whole-rock major and trace elements compositions of the rocks of the TSU

2.b. Trace elements

The trace element concentrations of the quartzites and phyllites of the TSU are presented in Table 3. The studied rocks of the TSU exhibit a similar REE pattern comparable to UCC and PAAS on the REE-diagrams normalized to chondrite (Fig. 8), characterized by Light Rare Earth Elements (LREE) enrichment relative to Heavy Rare Earth Element (HREE) and also negative EU anomaly. The quartzites in general, have lower overall REE concentrations than the phyllites (Fig. 8). Figure 9 shows the UCC-normalized multi-element diagrams for the studied quartzites and phyllites of the TSU. On these diagrams, the studied rocks display strong Sr depletion akin to PAAS. In general, the patterns depicted by the studied rocks appear to resemble that of PAAS, however, with significant variations (Fig. 9). The studied quartzites and phyllites display negative Ba, P, Nb-Ta and Ti peaks compared to PAAS.

Figure 8. Chondrite-normalized REE plot for the rocks of the TSU (a) quartzites, (b) phyllites. The TSU data are compared to published data from the sandstones and shales of the Buem structural unit, Kwahu-Bombouaka Group of the Voltaian Supergroup and phyllites from the TSU (Osae et al., 2006; Kalsbeek and Frei, 2010; Abu and Zongo, 2017; Anani et al., 2017; 2019; Amedjoe et al., 2018; Abu et al., 2020). Normalizing values for UCC from Rudnick and Gao (2003) and Chondrite from Palme and O’Niel (2014). PAAS is from McLennan (1989).

Figure 9. Multi-elements plot normalized to UCC for the rocks of the TSU (a) quartzites, (b) phyllites. The TSU data are compared to published data from the sandstones and shales of the Buem structural unit, Kwahu-Bombouaka Group of the Voltaian Supergroup and phyllites from the TSU (Osae et al., 2006; Kalsbeek and Frei, 2010; Abu and Zongo, 2017; Anani et al., 2017; 2019; Amedjoe et al., 2018; Abu et al., 2020). Normalizing values for UCC from Rudnick and Gao (2003) and Chondrite from Palme and O’Niel (2014). PAAS is from McLennan (1989).

3. Discussion

3.a. Influence of metamorphism on elemental mobility

The presence of chlorite, sericite and quartz observed from the petrographic investigation in this study points to greenschist-facies metamorphism. It is thus crucial to assess how metamorphism affects the mobile elements in the TSU. High Field Strength Element (HFSE) and REEs are generally not affected when siliciclastic rocks are metamorphosed. The rocks in the TSU exhibit an REE pattern similar to PAAS (Fig. 9), which is unexpected if the elements were remobilized during metamorphism (McLennan & Taylor, 1991; Girty et al., 1994). Therefore, the trace elements display chemical uniformity and data homogeneousness, implying that there was no significant remobilization at a large scale.

3.b. Maturity and recycling of the sediments

The modal composition of the quartzites varies from 88 to 99% quartz that is rounded to sub-rounded (Fig. 4c and 4e). These imply that the quartzites are texturally and mineralogical mature with high recycling due to either long-distance travel or long periods at the depositional basin. However, they have been metamorphosed and deformed during the Pan-African orogeny, evident in the sutured grain contacts and elongation (Fig. 4). Maturity in siliciclastic rocks can be tested using the SiO₂/Al₂O₃ and Na₂O/K₂O ratios (Dabard, 1990; Armstrong-Altrin, 2015). High SiO₂/Al₂O₃ (> 10) and low Na₂O/K₂O (<1) ratios are characteristics of mature siliciclastic rocks (Dabard, 1990; Armstrong-Altrin, 2009, 2015). The quartzites have high values of SiO₂/Al₂O₃ (>10; Table 3) and low Na₂O/K₂O (<1) values, and these, coupled with their high quartz content, suggest they are composed of mature sediments. The phyllites, on the other hand, have low SiO₂/Al₂O₃ (<10) and low Na₂O/K₂O (<0.1, respectively), which, except for two samples with high SiO₂/Al₂O₃, suggest immature sediment. Recycling of sediments within the sedimentary basin can be inferred from the Zr/Sc ratio because, during sorting and recycling, the ratio of Zr/Sc increases due to the addition of zircon (McLennan et al., 1993). Usually, variation in the geochemical composition of siliciclastic rocks due to variation in the source rock (i.e., mafic and felsic rocks) may result in a strong positive correlation on the Th/Sc versus Zr/Sc diagram, whereas sedimentary recycling will result in higher ratios than UCC and PAAS. The studied rocks of the TSU have a wide range of Zr/Sc and Th/Sc ratios (Fig. 10), which are higher than UCC and PAAS, indicating a significant recycling process, and thus, the rocks are composed of mature sediments. Compared to the shales, siltstones, sandstones and phyllites of the Buem and TSUs and Kwahu-Bombuaka Group, the studied rocks show similar Zr/Sc and Th/Sc ratios. This may suggest similar sediment maturity and recycling.

Figure 10. (a) Th/Sc versus Zr/Sc after (McLennan et al., 1993). Note that data points from TTG, granite, felsic volcanic rocks, andesite and basalt are from Condie (1993) The TSU data are compared to published data from the sandstones and shales of the Buem structural unit, Kwahu-Bombouaka Group of the Voltaian Supergroup and phyllites from the TSU (Osae et al., 2006; Kalsbeek and Frei, 2010; Abu and Zongo, 2017; Anani et al., 2017, 2019; Amedjoe et al., 2018; Abu et al., 2020).

3.c. Provenance

Medium to coarse sand-size quartz is the dominant grain in the quartzites, which may suggest sediment from a felsic proto-source such as granitoid, gneiss or pre-existing quartz-rich sandstone. The phyllites, conversely, contain micas and chlorite, which suggest pelitic proto-source rocks such as shale with significant mafic input. Generally, the ratio of LREE/HREE is high in felsic igneous rocks and low in mafic igneous rocks (Cullers, 1994). Mafic and felsic igneous rocks usually have Eu anomaly of 0.8 – 1 and 0.5 – 0.8, respectively (McLennan et al., 1993). Accordingly, the ratio of LREE/HREE and the Eu anomaly have been used to infer the proto-source rock of sedimentary rocks (e.g., McLennan, 1989). The rocks of the TSU all show a higher LREE/HREE ratio (5.72 – 13.32; Table 3) and significant negative Eu anomalies (0.6 – 0.8; 3; Fig. 9). In view of that, the provenance of the rocks of the TSU can be explained as sedimentary deposits dominated by detritus derived from intermediate and felsic proto-source rocks. Ratios of La/Sc, Th/Sc, Eu/Eu* and (La/Lu)_N provide information about the composition of the source rock (Cullers et al., 1988; Cullers, 1994, 2000). The elemental ratios of the rocks of the TSU suggest their sediments were derived from a felsic source rock (Table 4). Since some of the ratios overlap between felsic and mafic rocks, especially for the phyllites, it would suggest a significant contribution from a mafic rock. These elemental ratios are akin to siliciclastic rocks of the Buem structural unit and the Kwahu-Bombuaka Group of the Voltaian Supergroup, which might suggest similar source rocks. Overall, both felsic and intermediate proto sources contributed significant sediments to the rocks of the TSU. Available U-Pb detrital zircon age population of the quartzites from the TSU range between 2200 and 930 Ma, with abundant Mesoproterozoic (1600 – 1000 Ma) zircon grains, which are yet to be found on the West African Craton but abundant on the Amazonian Craton (Kaslbeek et al., 2008; Ganade de Araujo et al., 2016). Therefore, the most likely sources for this sediment are the basement granitoid, granitic gneisses and/or sedimentary rocks of the West African Craton and/or Amazonian Craton.

Table 4. Elemental ratios of the rocks of the TSU compared with sediments derived from felsic and mafic rocks

3.d. Depositional setting of the rocks of the TSU

The mineralogical and geochemical compositions of siliciclastic rocks offer important information regarding the depositional setting of the source rocks (Armstrong-Alrin et al., 2015). The main assumption is that different tectonic settings impact different mineralogical compositions and unique geochemical features on siliciclastic rocks that are usually retained in the sediments via diverse sedimentary processes (Dickinson et al., 1983; Armstrong-Alrin et al., 2015; Verma & Armstrong-Altrin, 2016). Consequently, Anani et al. (2019), on the basis of mineralogical and geochemical studies of the phyllites of the TSU, proposed a passive margin depositional setting (Table 1). Nevertheless, because of a U-Pb zircon age population of 2200 – 930 Ma, Kalsbeek et al. (2008) inferred a passive margin setting for the quartzites of the TSU. This is attributed to the passive margin sequence in the Dahomeyide belt and the West Gondwana orogen having the youngest detrital U-Pb ages >900 Ma (Kalsbeek et al., 2008; Ganade de Araujo et al., 2016; Caxito et al., 2020; Kwayisi et al., 2022b). In this study, the framework grains are used together with the geochemistry to determine the depositional setting of the rocks of the TSU. Figure 11a is a plot of the Q–F–L proposed by Dickinson et al. (1983) for depositional setting discrimination. The studied quartzites of the TSU plot mainly in the craton interior, with two samples plotting in the recycled orogen and one sample in the transitional continent (Fig. 11a). Continental blocks represent sediments that originated from stable craton interiors or the uplift of basement rocks. Sedimentary rocks from this setting are generally quartz arenites with subordinate arkosic sandstone, as in the case of the rocks of the TSU (Dickinson et al., 1983). The two samples that plot in the recycled orogen are those that originated from sedimentary strata and volcanic/plutonic rocks or metamorphic equivalent that are eroded as a result of uplift during an orogenic event of fold-and-thrust belts (Dickinson et al., 1983). Sedimentary rocks formed here are feldspar-poor and lithic fragment-rich; in the case of the TSU, they are sublitharenite (Fig. 6; Dickinson et al., 1983).

Figure 11. Tectonic setting discrimination diagrams for the rocks of the TSU (a) Q–F–L diagram (Dickinson et al., 1983), (b) DF(A-P)M diagram (after Verma and Armstrong-Altrin, 2016) and (c) Th–Sc–Zr/10 plot (Bhatia and Crook, 1986). The TSU data are compared to published data from the sandstones and shales of the Buem structural unit, Kwahu-Bombouaka Group of the Voltaian Supergroup and phyllites from the TSU (Osae et al., 2006; Kalsbeek and Frei, 2010; Abu and Zongo, 2017; Anani et al., 2017, 2019; Amedjoe et al., 2018; Abu et al., 2020). DF = Discriminant function, A and P are active and passive margin, respectively, and M = major element composition.

The studied rocks of the TSU again plot dominantly in the passive margin field with few plotting in the active margin field on the DF(A-P)m tectonic discrimination diagram (Fig. 11b) after Verma and Armstrong-Altrin (2016). In order to better understand the tectonic environment of the TSU, a Th–Sc–Zr/10 immobile trace elements discrimination diagram by Bhatia & Crook (1986) was plotted. This diagram is helpful in distinguishing between four distinct tectonic settings: Passive Margins, Active Continental Margins, Continental Island Arcs and Oceanic Island Arcs. The TSU samples dominantly plot in and around the passive margin setting, with some samples falling in and around the active continental margin and continental arc settings (Fig. 11c). Overall, the mineralogical, major and trace element compositions of the rocks of the TSU suggest dominantly passive margin affinity with subordinate active margin signature.

Prior to the Pan-African convergence, there was a rifting formation and evolution of a passive margin basin at c. 1000 and 700 Ma (Ganade de Araujo et al., 2016; Caxito et al., 2020; Kwayisi et al., 2022b). The depositional age of the rocks of the TSU is poorly constrained at c. 703±8 Ma (U-Pb zircon age metabasaltic rock), and this age falls within the age of passive margin formation (1000 – 700 Ma; Fig. 12; Ganade de Araujo et al., 2016). Again, Kalsbeek et al. (2008) inferred a passive margin setting for the quartzites of the TSU because of the presence of the youngest detrital zircon ages of >900 Ma. This interpretation is, however, inconsistent with the active margin signature exhibited by some of the rocks of the TSU from geochemistry. Thus, the TSU might have received sediments during the pre-convergence (i.e. rifting and passive margin) phase and later during the convergence (subduction and collisional) phase of the Pan-African orogeny. This inference needs to be tested from detailed geochronological studies, as the convergence phase would introduce detrital zircons of ages between 700 and 550 Ma (Fig. 12).

Figure 12. Space–time plot showing the age range of principal rock units and major events of the Dahomeyide belt. Published age data are from Clauer (1976), Clauer et al. (1982), Ganade de Araujo et al. (2014b) and Ganade de Araujo et al. (2016).

3.e. Formation and evolution of passive margin sequences of the Pan-African Dahomeyide belt

Many researchers have proposed the formation of TSU rocks in the same depositional setting and similar provenance as the other siliciclastic rocks of the Dahomeyide belt external nappes zone (Buem structural unit and the Kwahu-Bombuaka Group (e.g., Affaton, 1990; Anani et al., 2019; Kwayisi et al., 2022b). This interpretation is based on similar U-Pb zircon age populations obtained for these units (Kalsbeek et al., 2008; Kwayisi et al., 2022b). The TSU, Kwahu-Bombuaka Group and the Buem structural unit have zircon age populations in the range 930 – 2200 Ma, 1130 – 2200 Ma and 970 – 2200 Ma, respectively, with significant Archean zircon grains (Kwayisi et al., 2022b). The presence of the youngest detrital zircon ages older than 900 Ma is an indication of the deposition of these units in a passive margin depositional setting (Kwayisi et al., 2022b).

Two major passive margin sequences have been identified from the detrital zircon age populations. (i) lower passive margin sequences with ages dominantly between 1800 and 2200 Ma, which lack Neoproterozoic detrital zircons and (ii) upper passive margin sequences with significant Neoproterozoic detrital zircons between 900 and 1000 Ma (Kwayisi et al., 2022b). Nonetheless, similarities in zircon age population among units could mean one of the following:

1. i. the units have been deposited at the same time, sampling the same sources (basement rock or older sediments), or

2. ii. the units have been deposited at different times but have sampled the same sources or unrelated sources with similar zircon age distributions, or

3. iii. the units are of different ages, the younger formed by recycling of the older (Andersen et al., 2019).

In this study, petrographic investigations have revealed that the quartzites of the TSU formed probably in a passive margin depositional setting akin to the sandstones of the Buem structural unit and the Kwahu-Bombouaka Group (Fig. 11a). Major and trace element composition suggests that the sediments of the TSU were sourced from the same source area and deposited in the same depositional setting as the Buem structural unit and Kwahu-Bombouaka Group. The three units depict similar REE and trace element patterns on the Chondrite- and UCC-normalized diagrams (Figs. 8 and 9). Provenance and depositional setting indicators suggest dominantly passive margin deposition for the TSU’s rocks formed with the sediments derived from intermediate and felsic sources. These provenance and depositional setting patterns have also been identified in the Buem structural unit and the Kwahu-Bombouaka Group (Figs. 11b and 11c; Table 1). The Kwahu-Bombouaka Group was deposited earlier (deposition age = 959±65 Ma, Rb-Sr isochron age) than the TSU (depositional age = 703 ± 8 Ma, U-Pb zircon age) and the Buem structural unit (Rb-Sr of 650 Ma; Fig. 12; Table 1). The large variation in depositional ages indicates a long period of passive margin formation, from 1000 to 700 Ma. This age corresponds to the break-up of Rodinia supercontinent (Evans, 2009, 2013).

The structural architecture and deformational history of the TSU are similar to the Buem structural unit and the eastern margin of the Voltaian Supergroup. In this study, three deformational (D1 – D3) events have been identified in the rocks of the TSUs. The main deformational event (D2) expressed as NNW-striking S2 foliation, isoclinal folds with an N-plunging fold axis and downdip stretching lineation is similar to the main D2 event in the Buem structural unit (Kwayisi et al., 2020). This D2 event corresponds to an E-W shortening at 640 – 620 Ma during the convergence phase of the Pan-African orogeny (Kwayisi et al., 2020). Thus, the similarities in deformation style, petrographic and geochemical signatures could mean that the three were deposited within the same depositional basin, sampling the same source area lithology and later deformed by the Pan-African orogeny. Hence, during the break-up Rodinia and prior to the assembly of Gondwana supercontinents, there was the formation of a passive margin basin that was later tectonically deformed and metamorphosed during the Pan-African orogeny at the eastern margin of the West African Craton. Available detrital zircon age data suggest that abundant Mesoproterozoic detrital zircon (1600 – 1000 Ma; Kalsbeek et al., 2008; Ganade de Araujo et al., 2016; Kwayisi et al., 2022b) grains are recorded for the passive margin sequences of the Dahomeyide belt. No record of Mesoproterozoic granitoid and granitic gneisses on the WAC suggests sediment provenance from the Amazonia Craton. This finding thus supports the fact that the West African–Amazonian cratons coexisted without any major oceans separating them during the Rodinia Period (Fig. 13; Kalsbeek et al., 2008; Kwayisi et al., 2022b).

Figure 13. The palaeogeography reconstruction of Rodinia (modified after Antonia et al. 2021), which suggests that during Rodinia time, the WAC and the Amazonian Craton were likely not separated by any major seas. It is postulated that these cratons may have been connected from the Paleoproterozoic era until the break-up of Pangea.

4. Conclusion

The framework grains of the quartzites of the TSU, which consist of dominant quartz, minor feldspar and lithic fragments, suggest they are quartz arenite, subarkose and sublitharenite. This classification is comparable to the siliciclastic rocks of the Kwahu-Bombouaka Group of the Voltaian Supergroup and Buem structural unit. On the whole, the studied rocks show similar REE and multi-element patterns on the chondrite and UCC normalized diagram that indicate the derivation of the sediments from similar sources. The presence of abundant quartz grains coupled with the (La/Lu)_N, Th/Sc, La/Sc, La/Co and Th/Co values suggest mature sediments sourced from intermediate to felsic source area lithology. Provenance and depositional indicators of the studied rocks of the TSU resemble those of the siliciclastic rocks of the Kwahu-Bombouaka Group and Buem structural unit. Available petrographic, geochemical and geochronological data of the Buem structural unit and Voltaian Supergroup indicate deposition in a passive margin setting with sediment sourced from the Amazonian and West African cratons, and this is akin to the TSU. Therefore, the TSU was deposited in a passive margin setting with the Amazonian and West African cratons as potential provenance. Passive margin formation is dated between 1000 and 700 Ma, corresponding to the break-up of the supercontinent Rodinia. The structural architecture and deformational history of the TSU are similar to those of the Buem structural unit and the eastern margin of the Voltaian Supergroup. Thus, a Neoproterozoic passive margin basin formed and was deformed and metamorphosed during the break-up of Rodinia and prior to the assembly of Gondwana supercontinents. Hence, the West African–Amazonian cratons coexisted without any major oceans separating them during the Rodinia Period.