Geochemical and Petrographical Characteristics of the Madzaringwe Formation Coal, Mudrocks and Sandstones in the Vele Colliery, Limpopo Province, South Africa: Implications for Tectonic Setting, Provenance and Paleoweathering

: The sedimentary rocks of the Madzaringwe Formation in the Tuli Basin have been investigated using geochemical and petrographic methods to reveal their source area composition, tectonic setting, provenance and paleoweathering conditions. The petrographic studies show that the rocks consist mostly of clay minerals and quartz. The major elements geochemistry indicates that the rocks of the Madzaringwe Formation have the same source area. Based on the discriminant function plots, it can be inferred that the rocks are of quartzose sedimentary provenance, suggesting that they were derived from a cratonic interior or recycled orogen. The binary plots of TiO 2 versus Zr and La/Sr against Th/Co shows that the rocks were derived from silicic or felsic igneous rocks. The tectonic setting discrimination diagrams of SiO 2 against Log (K 2 O/Na 2 O), Th–Sc–Zr/10, and TiO 2 versus (Fe 2 O 3 + MgO) support passive-active continental margin settings of the provenance. The A–CN–K (Al 2 O 3 –CaO + Na 2 O–K 2 O) ternary diagram and binary plot of the index of compositional variability (ICV) against chemical index of alteration (CIA) shows that the rocks have been subjected to moderate to intensive weathering. Geochemical and petrographic characteristics of the rocks point to uplifted basement source areas predominantly composed of sedimentary rocks and/or granite-gneiss rocks. These source areas might have been from adjacent areas near the Tuli coalﬁeld which include the Limpopo Belt (igneous and sedimentary rocks), and basement uplifted rocks of the Beit-Bridge Complex, consisting of the granite, granite-gneisses and schists.


Introduction
The Madzaringwe Formation in the Tuli Basin is one of the coal-bearing Late Palaeozoic units of the Karoo Supergroup [1]. The Permian stratigraphy of the Tuli Coalfield, from the base to the top, comprises of the Tshidzi Formation, Madzaringwe Formation, Mikambeni Formation and Fripp Formation [2]. The Madzaringwe Formation in the former Vele Colliery (now referred to as Vele MC Mining) is the focus of this study and it consists of carbonaceous mudstone and shale with thin layers of coal seams and sandstones. Mudrocks (i.e., shale and mudstone) are fine-grained sedimentary rocks composed of mostly silt and clay size fragments. Owing to their small grain size, mudrocks are difficult to study, even with the petrographic microscope [3]. To date, mudrocks are the least understood, and one of the most understudied sedimentary rocks. Nonetheless, they are vital rocks because they are the most abundant sedimentary rocks on earth, contributing more than 65% of all sedimentary rocks [3].
The geochemistry of clastic sedimentary rocks (i.e., mudrocks and sandstones) is a vital tool used in the study of provenance, paleoweathering conditions, and tectonic setting [3,4]. The mineralogical and chemical compositions of these rocks are the products of numerous variables that include provenance, weathering conditions, transport, diagenesis, climate and tectonism [5]. In geochemical studies, some selected major oxides especially TiO 2 and trace elements like La, Y, Sc, Cr, Th, Zr, Hf, and Nb are sensitive indicators of the source rocks, provenance, paleoweathering, and tectonic setting [6]. This is as a result of their relatively low mobility and insolubility during the aforementioned sedimentary and post-sedimentary processes [7]. Also, comparative distribution of the immobile trace elements with changing concentration in felsic and basic rocks has been employed to deduce the relative contribution from felsic and/or basic sources in clastic sedimentary rocks from different tectonic environments [8].
The Madzaringwe Formation in the Vele Colliery has about 120 m of coal-bearing unit interbedded within mudrocks and occasionally sandstones and it is the lowermost part of the Ecca Group in the Tuli Basin [9]. Geographically, the study area is located between longitudes 29.56 • E and 29.66 • E and latitude 22.15 • S and 22.22 • S ( Figure 1). Previous investigation on the Madzaringwe Formation in the Tuli Basin mainly focused on the sedimentary environment and palaeogeographical analyses [9][10][11]. In this paper, we investigate the geochemistry and petrography of the Madzaringwe Formation coal, mudrocks and sandstones from the Vele Colliery in the Tuli Basin, Limpopo Province, South Africa. The purpose of this research work is to assess the source area composition, tectonic setting, provenance and paleoweathering conditions of the formation. The result from this study could give insights into the geologic/tectonic evolution of the Tuli Basin and will also add to the database on the geochemical and petrographic characterization of sedimentary units in the Tuli Basin, which is almost lacking at present.  [12,13]).

Geological Setting
The Tuli Basin is also known as the Limpopo Basin and it is a small intracratonic east-west trending fault controlled depository with a preserved basin width of about 80 km [2]. The basin forms part of the Limpopo Karoo Supergroup and is deposited at the same time with the Main Karoo Basin, but it is formed as an isolated basin which was fault-controlled from the onset [1]. The Tuli Basin occurs across the borders of South Africa, Botswana and Zimbabwe and is filled with the igneous and sedimentary rocks of the Karoo Supergroup [11]. The sedimentary rocks which are found in the South African part of Tuli Basin consists of different terrigenous clastic and chemical deposits. These rocks are para-breccias, conglo-breccias, conglomerates, sandstone, fine-grained sediments, calcretes and silcretes amongst others [10]. Generally, the word "Karoo" is coined from the Main Karoo Basin of South Africa to describe sedimentary fill of all basins of similar age across Gondwana [14]. As reported by [15], the Karoo Basin occupies about 66% of the southern part of the African continent and it is composed mostly sedimentary rocks of Paleozoic age. The Main Karoo Basin of South Africa is a unique type of basin of all the Karoo basins in southern Africa in terms of its thickness and stratigraphy [14]. According to [16], the deposition of the Karoo Supergroup occurred in two broadly different tectonic settings in southern Africa. The Tuli Basin represents the Limpopo area of Karoo-age basins.
The Tuli Basin lies at the triple junction of the Zimbabwean, Botswanan and South African borders and it is generally known that the basin separated from the Tshipise Basin [1]. It has been proposed that the Limpopo area forms the western arm of a failed rift triple-junction [17], which later extended in a north-south direction and from the Save Basin in Zimbabwe to the Lebombo Formation in South Africa [11]. According to [18], the separation is caused by a highly deformed Messina Block in the southern boundary, which could have acted as a positive area during the accretion of the Karoo strata. The Tuli Basin consists of a northern boundary and the boundary is marked by a major ENE trending fault, which is continuous for about 100 km [17]. To date, the tectonic evolution of the Tuli Basin is still debatable, and several models have been used to unravel the tectonic development of the Tuli Basin. One of the tectonic models suggested or provided by [1] shows an ENE trending divergent wrench zone which was responsible for the formation of the Tuli Basin, Nuanetsi Basin and Tshipise Basin ( Figure 2). The displacement formed in this model occurred along pre-existing ENE and NE trending faults during the sedimentation of the lower part of the Karoo Supergroup [1]. There is a compressive foreland system, which existed north of the Cape Fold Belt and it was developed in response to the Late Palaeozoic to Early Mesozoic subduction of the Pacific plate [1]. Perhaps, as result of the flexural warping of the lithosphere, the occurrence of three different areas across the distorted profile of the system appears, which include a foredeep, forebulge and back-bulge zone. The Tuli Basin may be recognized with this back-bulge setting based on its distance, which is approximately 1500 km from the northern margin of the Cape Fold Belt and has a relatively thin sedimentary sequence [1]. The area investigated in this study is located in the Tuli Basin, in the north-eastern part of South Africa near Tshipise-Pafuri Basin. Generally, the Tuli Basin trends in the East-West direction and links up with the North-South trending Lebombo Basin, which represents the southern termination of the East African Rift System [19]. In the Tuli Basin, the Madzaringwe Formation, Mikambeni Formation and Flipp Formation are the equivalents of the Ecca Group in the Main Karoo Basin [1] (Figure 3). As documented by [10], the Madzaringwe Formation in Tuli Basin consists of an alternating succession of feldspathic, cross-bedded conglomerates, sandstone, siltstone, and shale containing some coal seams. Furthermore, they reported that this formation was deposited on top of the Tshidzi Formation. Coal series in the Madzaringwe Formation are located at the depth of less than 50 m along the southern margin and they can reach a depth of over 300 m close to the Limpopo River [1]. The two major seams which are flat lying coal have thickness of about 1.6 m and 1.2 m [16]. These seams are overlain by mudstones and minor sandstones. The economically important coal in Madzaringwe Formation is interbedded with mudstones [1]. Succeeding the Madzaringwe Formation is the Mikambeni Formation, which is about 60 m thick, consisting of shale, sandstone and coal [16]. The Flipp Formation overlies the Mikambeni Formation and it is composed of about 5-10 m thick successions of sandstone and conglomerates [16].

Materials and Methods
Geological core logging was performed on the three existing boreholes (boreholes OV125149, OV125156, and OV125160) at the Vele Colliery core yard. Thickness of the stratigraphic units was measured perpendicular to the strike of the strata using a measuring tape. The stratigraphic data was used to generate the stratigraphic logs of the Madzaringwe Formation in the above-mentioned boreholes. A total of 28 samples, comprising of 7 coal, 12 mudstones, 6 shales and 3 sandstones from the Madzaringwe Formation in the Vele Colliery were collected and subjected to petrographic, X-ray diffraction (XRD) and X-ray fluorescence (XRF) analyses at the Council of Geoscience, South Africa. Fifty-two representative thin sections of the shale, mudstone and sandstone were prepared and studied using a Nikon Eclipse Polarizing microscope to determine their mineral compositions, texture and arrangement of mineral grains. The Bruker XRD D8 Advance (Model: V22.0.28) was used for the XRD analysis. The XRD measurements were done at 25 • C room temperature and the samples were scanned at 2 • θ/min ranging from 2 • to 70 • (wavelength of 1.5406). The XRD results are expressed in wt. % and it only show crystalline phases, and do not recognize amorphous phases such as carbonaceous organic matter and hydrous iron oxide minerals. The XRF analysis of the major and trace elements was performed using Magi-X Fast spectrometer. Bulk sample preparation consists of drying where necessary, crushing to less than 10 mm, and subsequent milling in a tungsten carbide milling machine to less than 50 µm. The major elements were analysed on fused beads, while trace elements were executed on pressed powder pellets.
Different discriminatory plots of the major and trace elements have been used by several researchers [5][6][7]14] to determine the provenance and tectonic setting. These plots were attempted in this study to infer the tectonic provenance of the Madzaringwe Formation. Furthermore, the ternary diagram of Al 2 O 3 -(CaO + Na 2 O)-K 2 O (represented as A-CN-K) was plotted and the chemical index of alteration (CIA), chemical index of weathering (CIW), and plagioclase index of alteration (PIA) were calculated and used to quantify the degree of weathering.

Stratigraphy
The stratigraphic sequence of the Madzaringwe Formation in the Vele colliery consists mostly of mudstone and carbonaceous shale with thin layers of coal seams, siltstone and sandstones. The Madzaringwe Formation has a thickness of about 61 m, 60 m, and 42 m in boreholes OV125149, OV125156, and OV125160, respectively ( Figure 4).

Petrography and Mineralogy
Petrographic studies of the mudstone and shale shows that the rocks are predominantly made up of clay minerals (mostly kaolinite and calcite), quartz, and pyrite ( Figure 5). The shales have lamination or fissility texture, allowing it to break along laminae planes. In cases where lamination on the mudrocks did not occur, thin to massive bedded mudstones were observed. Typically, detritus clay and silt size particles with some chemically precipitated cement and organic materials make up the argillaceous rocks. The sandstones are very fine to coarse grained and mostly poorly sorted ( Figure 5). The grain morphology for the sandstones range from subangular to angular and are mostly made up of clay minerals and quartz. The XRD analysis shows that the dominant clay minerals in the mudrocks and sandstones are kaolinite, calcite and illite, ranging between 8.0-31.0%, 0.4-25.0% and 0.1-26.0%, respectively. The coal samples are dominated by kaolinite and graphite, ranging between 49.2-55.0% and 27.8-36.4%, respectively (Table 1).    Table 3. Rb, Sr, and Nb are more abundant in the coal and shale samples, while Cr and Ni is more abundant in the sandstones. In general, the studied samples have a high concentration of Ba, Zn, Zr, Rb, and Ce.

Interpretation and Discussion
The chemical composition of the studied samples is variable but generally comparable with the average shales documented by [21][22][23][24][25] (Supplementary Tables S3 and S4). The abundance of Al 2 O 3 was used as a normalization factor to make contrasts among the different lithologies due to their immobile nature throughout weathering, diagenesis, and metamorphism. In the shale and mudstone samples, major oxides like TiO 2, and K 2 O show a positive correlation with Al 2 O 3, whereas SiO 2 , MgO, CaO, Na 2 O, and P 2 O 5 show no particular trend ( Figure 6). The positive correlation of these major oxides with Al 2 O 3 indicates that they are associated with micaceous/clay minerals [14]. Also, a positive correlation of K, Ca, and Na oxides with Al 2 O 3 occurs in feldspars. However, the strongest evidence for the presence of mica/clay in the samples is the elevated levels of Al 2 O 3 as revealed by major element analysis. As a general rule, unweathered igneous rocks have  Supplementary Table S1 show shales and mudstones with alumina values in the 17-20% range, which is stronger evidence of clay/mica compositions than the bivariate plots in Figure 6. As Aluminium (Al) concentration is reasonably thought to be a good measure of detrital flux excellent positive correlations of TiO 2 , and K 2 O. Relative to the Upper Continental Crust (UCC) and Post Archean Australian Shale (PAAS), the concentrations of major elements in the studied samples are generally similar, except for MnO and Na 2 O, which consistently yields much lower average relative concentration values than those of UCC and PAAS (Figures 7 and 8). The depletion of Na 2 O (<1%) in the Madzaringwe mudrocks and sandstones can be attributed to a relatively smaller amount of Na-rich plagioclase in them, which is consistent with the petrographic observation and XRD data. The ratio of K 2 O/Na 2 O also shows that K-feldspars dominate over plagioclase (albite) feldspar. The contents of K 2

Tectonic Setting
Several researchers including [6,9,[27][28][29] have envisaged that tectonic settings control or influence the chemical compositions of clastic sedimentary rocks, indicating that variety of tectonic settings have terrain-specific signatures. Different tectonic setting discrimination diagrams have been used to separate between different tectonic settings, and they all provide consistent results for siliciclastic rocks that have not been strongly affected by post-depositional weathering and metamorphism [6,27,28]. Similarly, the bivariate and ternary plots of major and trace element geochemistry have been widely used by several researchers to determine the tectonic setting of shales [14]. The most used discrimination diagrams are major element-based discrimination diagrams of [27,28]. The geochemical data of the studied samples were plotted on binary plots of SiO 2 against Log (K 2 O/Na 2 O), TiO 2 versus (Fe 2 O 3 + MgO), and ternary plot of Th-Sc-Zr/10. The binary plots of SiO 2 against K 2 O/Na 2 O and TiO 2 versus (Fe 2 O 3 + MgO) show that the sedimentation of the studied samples were related to both active continental and passive continental settings (Figures 9 and 10). Likewise, the ternary plot of Th-Sc-Zr/10 revealed that the samples were mainly deposited in a passive continental setting, with little contribution from active continental margin setting ( Figure 11).

Source Area (Provenance)
Using major oxides as variables, Roser and Korsch [30] established major element discriminant functions to differentiate four major provenance fields, namely mafic, intermediate, felsic and quartzose recycled. Likewise, the binary plots of TiO 2 vs Zr, and La/Sr vs Th/Co have been successfully used to infer provenance of clastic sedimentary rocks [5][6][7]14,31]. The bivariate plot of discriminant 1 against discriminant 2 on the background diagram of [30] shows that the samples are of quartzose sedimentary provenance ( Figure 12). This deduction is in agreement with a probable deposition in the realm of a passive margin (as highlighted in the previous section), with a significant influence of the sedimentary processes (transport from source areas far away from the sedimentary basins) over the mineralogical and geochemical features observed in these rocks, resulting in mature sedimentary rocks with an enrichment of the quartzose fraction. The binary diagrams of the TiO 2 versus Zr ( Figure 13) and La/Sr against Th/Co ( Figure 14) show that the studied rocks are from silicic or felsic igneous rocks. The XRF result also shows the enrichment in Sr and Ba and depletion of transitional elements like Cr, Co, Ni and V, suggesting the trivial or minor influence of mafic provenance as indicated by [32]. The high ratio of Ba to Co (Ba/Co) (Supplementary data;Supplementary Materials Data; Table S2) perhaps point to the fact that the sediments were derived from weathered felsic-granitic sources [33]. In general, the findings of this study agree with the work of [10,16,19] in that the Madzaringwe Formation is of quartzose sedimentary provenance derived from felsic igneous rocks.

Paleoweathering Conditions
Chemical weathering intensity in source rocks is mostly controlled by the composition of the source rock, period of weathering, climatic conditions and rates of tectonic uplift of the source region [35]. Calcium (Ca), sodium (Na), and potassium (K) are largely removed from source rocks during chemical weathering and the amount of these elements remaining in the sediments derived from the rocks aided as an indicator of the intensity of chemical weathering [36]. If siliciclastic sedimentary rocks are free from alkali related post-depositional modifications, their alkali contents (K 2 O + Na 2 O) and K 2 O/Na 2 O ratios should be considered as reliable indicators of the intensity of the source material weathering [37]. A few indices of weathering have been proposed based on the molecular proportions of mobile and immobile element oxides (Na 2 O, CaO, K 2 O, and Al 2 O 3 ) to determine the degree of source rock weathering. Therefore, chemical compositions of weathering products in a sedimentary basin are anticipated to disclose the mobility of various elements during weathering [35]. The indices of weathering/alteration include chemical index alteration (CIA), chemical index weathering (CIW), and plagioclase index weathering (PIA). To determine the paleoweathering conditions for the studied samples, the formula for the weathering indices were used as follows: where CaO * is the content of CaO incorporated in silicate fraction.
The calculated values for the chemical index alteration (CIA), chemical index weathering (CIW) and plagioclase index weathering (PIA) are presented in the supplementary data (Supplementary Materials Data; Table S5). The CIA is in a simplified way the ratio of primary minerals to secondary products (i.e., clay minerals). CIA values, usually ranging from about 50 for unweathered rocks up to 100 in highly weathered rocks. The CIA values for the studied samples range from 45.64% to 87.00%, suggesting weak to intense weathering conditions. Also, the binary plot of CIA against ICV shows that the samples are matured and have undergone weak to intense weathering conditions, mostly intense weathering ( Figure 15).
In addition, the weathering of feldspars to clay minerals can be monitored using the plagioclase index of alteration (PIA) [38,39]. The PIA values of the studied samples range from 45.16% to 96.81% (averaging 88.58%). The PIA values indicate moderate to intense destruction of feldspars during transportation, sedimentation and diagenesis. The ternary plot of A-CN-K proposed by [3] is another way of evaluating the composition of the source rock as well as the mobility of elements during the process of chemical weathering of source materials and post-depositional chemical modification. The ternary plot of Al 2 O 3 -(CaO + Na 2 O)-K 2 O (exemplified as A-CN-K) shows that the studied samples plotted above the line joining plagioclase and K-feldspar ( Figure 16). The weathering trendline of the samples tends towards the A-K boundary, suggesting the silicates (i.e., feldspar) have experienced moderate to high weathering, perhaps causing the leaching of calcium (Ca) and sodium (Na) out of plagioclase [38]. Furthermore, the trendline (arrows in Figure 16) parallels with the A-CN boundary and slightly extend towards the illite, signifying leaching of potassium (K) and enrichment of aluminium (Al) resulting in the decomposition of Kbearing minerals (biotite, kaolinite, and potassium feldspar). Thus, illite dominates the secondary clay minerals. On the other hand, the weathering trendline of the samples is relatively closer to the A-CN boundary, indicating that plagioclase is the first to be weathered, out of which Ca and Na leached rapidly, whereas K-feldspar is relatively stable. In the sandstones, illite is the main weathering product [38].

Conclusions
Geochemical compositions of the coal, mudrocks and sandstones were analysed to identify provenance, paleoweathering conditions and tectonic setting of the Madzaringwe Formation. The geochemical data of major and trace elements show that the studied rocks have the same source. Based on the discriminant function plots, it can be inferred that the rocks are of quartzose sedimentary provenance, suggesting that they were derived from a cratonic interior or recycled orogen. The binary diagram of the TiO 2 versus Zr and La/Sr against Th/Co as well as the discriminant function plots revealed that the mudrocks and sandstones are mostly of silicic or felsic igneous provenance. These source areas might have been from adjacent areas which include the Limpopo Belt (igneous and sedimentary rocks), and basement uplifted rocks of the Beit-Bridge Complex, consisting of the granite, granite-gneisses and schists near Tuli coalfield. The binary plots of TiO 2 versus (Fe 2 O 3 + MgO) and SiO 2 against K 2 O/Na 2 O, and ternary plots of Na 2 O-CaO-K 2 O and Th-Sc-Zr/10 revealed that the studied samples are of passive-active continental margin settings. The A-CN-K ternary diagram and indices of weathering revealed that the source rocks underwent a moderate to high degree of chemical weathering. Also, the indices of weathering/alteration and the binary plot of the index of compositional variability (ICV) against the chemical index of alteration shows that the studied samples are geochemically matured and have been subjected to moderate to intensive weathering.
Supplementary Materials: The following are available online at https://www.mdpi.com/2076-3 417/11/6/2782/s1, Table S1: Results of oxides (wt%) analysed by X-ray fluorescence spectrometry, Table S2: XRF data for trace element (in ppm) distribution in the rock samples of the Madzaringwe Formation, Table S3: Comparison of the average major oxide composition of the shales and coal from Madzaringwe Formation with published average shales, Table S4: Comparison of the average concentration of trace elements in the shales and coal from Madzaringwe Formation with published average shales, Table S5: Indices of weathering (CIA, CIW and PIA) calculated from the major elements.