1. Introduction
Sudoite is characterized by a dioctahedral 2:1 layer and a trioctahedral interlayer sheet and belongs to the di-trioctahedral chlorite mineral group with an ideal composition of (Al3Mg2)(Si3Al)O10(OH)8 [Bailey 1980]. Although this type of Al- and Mg-rich chlorite is far less common than its tri-trioctahedral Mg- and Fe-rich countertype, sudoite usually occurs in low-temperature environments [<300 °C; Beaufort et al. 2015; Lanari et al. 2014], including diagenetic and hydrothermal systems [e.g., Beaufort et al. 2005; Daniels and Altaner 1990; Hayashi and Oinuma 1964; Hillier et al. 2006; Percival and Kodama 1989; Rodríguez-Ruiz et al. 2019; Ruiz Cruz and Sanz de Galdeano 2005; Schultz 1963; Truche et al. 2018]. In such geological settings, sudoite likely forms from modifications of kaolin minerals (i.e., kaolinite, dickite, and nacrite) through a stepwise reaction under the influence of Mg-rich fluids. Sudoite also crystallizes at the expense of diagenetic beidellite when kaolins are lacking [Biernacka 2014]. These mineral transformations are generally controlled by a dissolution–crystallization mechanism since convincing evidence of kaolinite–sudoite or smectite–sudoite mixed-layer minerals (MLMs) are lacking [see review in Beaufort et al. 2015]. A large amount of sudoite has been reported in the Proterozoic uranium (U)-bearing sandstones that unconformably overlie Archean basements [Beaufort et al. 2005, Billault et al. 2002; Kister et al. 2005; Percival and Kodama 1989; Riegler et al. 2014; Truche et al. 2018], where the origin of sudoite was believed to have resulted from interactions of oxidized and acidic fluids with the underlying crystalline rocks within active fault zones. Therefore, sudoite can sometimes be a potential proxy for tracing U mineralization.
Despite numerous studies on the diagenetic mineral phases of the Paleoproterozoic Franceville sub-basin of Gabon [Aubineau et al. 2019; Bankole et al. 2018, 2016, 2015; Bros et al. 1992; Gauthier-Lafaye 1986; Ngombi-Pemba et al. 2014; Ossa Ossa et al. 2013; Weber 1968], the presence of kaolin minerals and di-trioctahedral chlorite have not been well-documented or constrained in the Francevillian rock sequence, especially in sandstones [cf. Ngombi-Pemba et al. 2014]. Synthetically, assemblages of detrital-shaped (including mica and ferromagnesian chlorite) and diagenetic (illite, tri-trioctahedral chlorite, and smectite-rich and illite-rich illite–smectite, I-S MLMs) clay minerals dominantly characterize the Francevillian Group sedimentary rocks. Most of the diagenetic clay minerals in the Francevillian sediments result from the alteration of aluminosilicate minerals [Gauthier-Lafaye 1986]. The surprising survival of smectite-rich (R0) I-S MLMs in localized stratigraphic levels (from the FB1b, to the FB2b units) of the Franceville sub-basin has been attributed to potassium deficiency in the system [Aubineau et al. 2019; Bankole et al. 2015; Ngombi-Pemba et al. 2014]. Indeed the variable proportion of K-feldspar was recognized in other rock units of the Franceville sub-basin [Bankole et al. 2015; Ngombi-Pemba et al. 2014], suggesting mixing of multiple source rocks [Bankole et al. 2020].
The FB2 Member of the Franceville sub-basin is believed to host the earliest potential multicellular life [El Albani et al. 2014, 2010] capable of movement [El Albani et al. 2019] and has provided a considerable interest in the complex life’s early evolution. Thus, deciphering the post-depositional modification processes in these unmetamorphosed sedimentary rocks may provide a glimpse of long-term effects of diagenetic history on the preservation of the Francevillian biota. This paper investigates the di-trioctahedral chlorite and kaolin minerals in sandstone and black shale facies of the Francevillian FB2 Member. Here we integrated X-ray diffraction (XRD), infrared spectroscopy, and optical and electron microscopy to examine the clay mineralogy and petrography of authigenic minerals and their textural relationships to provide additional insights into the diagenetic fluid history and mineral paragenetic sequence in the Franceville sub-basin of the Francevillian Group.
2. Geological setting
The Francevillian basin is composed of four intracratonic sub-basins—Booué, Lastourville, Okondja, and Franceville—in southeastern Gabon [Boutonet al. 2009]. The basin fill consists of 1.0 to 2.5-km-thick Paleoproterozoic siliciclastic sedimentary succession commonly referred to as the Francevillian Group [Figure 1a; Weber 1968]. Importantly, the Francevillian deposits were not affected by metamorphic processes and have only been subjected to diagenetic and low-temperature hydrothermal processes [Aubineau et al. 2020; Bankole et al. 2016, 2015; Gauthier-Lafaye and Weber 1989]. Specifically, the Franceville sub-basin records a maximum burial depth and temperature of about 2500 m and 80 °C, respectively [Ngombi-Pemba et al. 2014; Weber 1968], which corresponds to burial conditions of a common geothermal gradient.
(a) Geological map of the Paleoproterozoic Franceville sub-basin showing the study area—Moulendé Quarry [modified from Boutonet al. 2009]. (b) Lithostratigraphic column of the FB2 Member in the Moulendé Quarry displaying sampling positions of both microbial mat structures and host sediments (red arrows). Images in (a) and (b) are modified from Aubineau et al. [2020] and Aubineau et al. [2019], respectively.
These sediments rest unconformably on the Archean crystalline basement within the west Congolese craton. The basement rocks are mainly composed of Mesoarchean (3200–2800 Ma), TTG (tonalite–trondhjemite–granodiorite), and greenstone belts [Mouélé et al. 2014; Thiéblemontet al. 2009]. The synthesis of the geodynamic evolution of the Francevillian basin was recently detailed by Weber et al. [2016].
The Francevillian Group has been divided into five lithostratigraphic formations, labeled FA to FE from the oldest to the youngest. The Lower FA formation is dominated by braided fluvial conglomerates and sandstones, while the Upper FA formation is composed of fluvio-deltaic, fine-grained mudstones and sandstones [Bankole et al. 2015; Feybesse et al. 1998; Gauthier-Lafaye and Weber 1989]. This sedimentation occurred during a progressive basin opening. The topmost part of the FA formation is marked by post-depositional formation of U ore deposits in association with bitumen [Bankole et al. 2016; Gauthier-Lafaye and Weber 2003, 1989]. The overlying marine FB formation deposited during a period of tectonic subsidence and basin deepening is subdivided based on the lithostratigraphy into FB1 (including a, b, and c units) and FB2 (a and b units) members based on lithological differences and sedimentary structures [Azziley Azzibrouck 1986; Weber 1968]. The FB1 Member deposited during sea-level rise is mainly characterized by greenish shales (FB1a unit), rhythmic successions of greyish shale and dolomite-cemented siltstones (FB1b unit), and manganese-rich black shales [FB1c unit; Azziley Azzibrouck 1986; Gauthier-Lafaye and Weber 2003; Pombo 2004; Reynaud et al. 2017]. Following an episode of sea-level fall, massive sandstones frequently intercalated by dm-thick black shale layers (FB2a unit) that are capped by laminated black shales with interbedded siltstones (FB2b unit) were deposited [Figure 1b; Reynaud et al. 2017]. In addition, the FB2 Member sandstones and black shales host light-dependent, microbial mat morphotypes [Aubineau et al. 2021, 2019, 2018], previously described as pyritized mat-related structures (MRS) and non-pyritized MRS. These delicate MRS are closely associated with the well-preserved complex forms of life in the FB2b unit [Aubineau et al. 2018; El Albani et al. 2019]. The overlying FC formation consists of shallow marine deposits of massive dolostones and cyanobacteria-hosting stromatolitic cherts with intercalation of black shale beds [Bertrand-Sarfati & Potin 1994; Lekele Baghekema et al. 2017; Préat et al. 2011]. The FD formation is composed of transgressive marine black shales with interbedded volcanic tuffs [Thiéblemont et al. 2014], while the uppermost FE formation contains arkosic sandstones and forms lenticular outcrops in the Okondja sub-basin [Gauthier-Lafaye and Weber 1989; Thiéblemont et al. 2014]. Additionally, extensive descriptions and discussions of petrography, depositional facies, and stratigraphic and sedimentary evolution of the Lower Francevillian Group are provided in Bankole et al. [2016, 2015], Gauthier-Lafaye and Weber [2003, 1989] and Reynaud et al. [2017].
Despite numerous attempts to resolve the age and duration of deposition of the Francevillian Group [Bonhomme et al. 1982; Bros et al. 1992; Sawaki et al. 2017], the depositional age of siliciclastic sequence remains poorly constrained. U–Pb ages of 2083 ± 6 Ma and 2072 ± 29 Ma were reported from tuffs and epiclastic sandstones in the FD formation, respectively [Horie et al. 2005; Thiéblemontet al. 2009]. In addition, the FB and FC formations have recorded extremely positive 𝛿13C values in marine carbonates [El Albani et al. 2010; Préat et al. 2011], which are thought to correspond to the end of the longest-lived positive carbon isotope excursion in Earth’s history [i.e., Lomagundi event, 2.22–2.06 Ga; Karhu and Holland 1996]. Collectively, these age constraints place the age of deposition for the Francevillian Group close to ca. 2.15–2.08 Ga [Canfield et al. 2013].
3. Samples and analytical techniques
3.1. Sampling
Heterogeneous microbial mat morphotypes (both pyritized and non-pyritized MRS, defined from their petrographic and geochemical analyses, Aubineau et al. 2018), and their host sandstone and black shale sediments were collected from outcrops in the Moulendé Quarry (Figure 1b). The weathered outermost surfaces were removed before sampling outcrop sediments. We carefully separated the μm-thick mat laminae from underlying sediments with a stainless-steel razor blade, avoiding as much as possible contamination with the underlying host rocks. The host sediments were directly extracted below the MRS with a hammer.
3.2. X-ray diffraction (XRD)
The <2 μm clay fraction of MRS and host sediment samples were analyzed with a Bruker D8 ADVANCE diffractometer at the University of Poitiers using a CuK𝛼 radiation, operating at 40 kV and 40 mA. The dispersion of gently hand-crushed bulk samples in deionized water with an Elma S60 ultrasonic agitation device without any chemical pre-treatment [Moore and Reynolds Jr 1997] allowed the extraction of the <2 μm clay fraction by sedimentation. Oriented preparations of the <2 μm size fraction were prepared by drying ∼1 mL of suspension on glass slides at room temperature. The oriented mounts were examined at a step size of 0.02° 2𝜃 using a 3 s counting time per step and recorded from a 2–30° 2𝜃 angular range after successive air-dried (AD) and ethylene glycol (EG) saturation. In addition, the octahedral occupancy within chlorite minerals was measured based on the peak position of the 060 reflection on randomly oriented <2 μm clay fraction over a 57–63° 2𝜃 angular range with a step size of 0.025° 2𝜃 per 8 s counting time. Bruker Eva software was used for background stripping, indexing of XRD peaks and mineral identification by comparing with International Centre for Diffraction Data (ICDD) files.
The NEWMOD 2.0 program was used to semi-quantitatively estimate the relative clay proportions in the <2 μm fractions by fitting experimental samples in AD and EG states [Reynolds Jr and Reynolds III 1996]. We introduced the instrumental and experimental factors, including the divergence slit, goniometer radius, soller slits, sample length, and quartz reference intensity, which are specific to the Bruker D8 ADVANCE diffractometer. Sigmastar and the mass adsorption coefficient for Cu radiation were also set between 12 to 14° and 45 cm2⋅g−1, respectively [Moore and Reynolds Jr 1997]. The XRD profile modeling is, however, difficult to apply on the low-angle region with the NEWMOD 2.0 program. The calculation of the fit does not include the non-clay mineral reflections.
3.3. Microscopy
Petrographic observation and documentation of textural relationships were made by optical and scanning electron microscopies at the University of Poitiers. Polished thin sections were first examined under transmitted and reflected light using a Nikon ECLIPSE E600 POL microscope coupled with a Nikon Digital Sight DS-U1 camera and NIS-Elements D software. Carbon-coated rock slabs and thin sections were imaged with a JEOL JSM-IT500 scanning electron microscope (SEM) equipped with a Bruker energy-dispersive X-ray spectrometer (EDX). Investigations of the size and morphology of clay minerals in rock slabs were performed in secondary electron mode, while polished thin sections were studied in backscattered electron mode. Semi-quantitative analysis for mineralogical examination was carried out on C-coated polished rock slabs, using an FEI Quanta 200 SEM equipped with an EDX at the University of Lille. Analytical conditions of both SEM operated at an acceleration voltage of 10–15 kV, 1 nA beam current, and a working distance of 10.5 mm.
3.4. Fourier transform infrared (FTIR) spectroscopy
The bulk mineralogy and <2 μm clay fractions of MRS and host sediments were further analyzed by the FTIR spectroscopy at the University of Poitiers. Such analytical method is a powerful tool to distinguish the polymorph minerals of the kaolin group as well as the chlorite minerals by the examination of the fundamental hydroxyl-stretching vibrations in the 3800–3200 cm−1 region (middle infrared—MIR) of spectra. The samples were embedded in potassium bromide (KBr) pellets that consist of a mixture of 1 mg of sample and 149 mg of KBr, pressed for 5 min at 8 kbar, and dried overnight in an oven at 150 °C. MIR spectra were recorded in transmission mode from KBr pellets using a Nicolet iS50 FTIR spectrometer equipped with a KBr beamsplitter and a DTGS KBr detector. Each MIR spectrum corresponds to an accumulation of 100 scans at a resolution of 4 cm−1.
4. Results
4.1. Petrographic observation
Petrographically, sandstones in the FB2a unit dominantly consist of poorly sorted detrital quartz grains with minor clay and organic-rich matrix (Figure 2a, b). Elongated muscovite is sometimes bent at grain-to-grain contacts due to compaction and represents the dominant micaceous minerals. Concavo-convex features and stylolites between quartz grain contacts further characterize the compaction features. Although our observations were based on two-dimensional imaging methods, the primary porosity has been likely destroyed in the coarse-grained facies due to physical and chemical compaction effects. However, few intergranular pore spaces, mostly filled with organic matter and authigenic clay minerals, are sometimes preserved (Figure 2b). Petrographic observations also confirmed the absence of K-feldspars and the rare occurrence of plagioclase in the FB2a sandstones, as previously described by Ngombi-Pemba et al. [2014]. Our petrological and mineralogical observations in the FB2b black shales are similar and consistent with the detailed descriptions of Ngombi-Pemba et al. [2014].
Optical and scanning electron microscope photomicrographs of the FB2a sandstones. (a, b) Pore-filling organic matter and clay minerals in poorly sorted sandstones (PL and CP). The quartz grains experienced moderate to high degrees of compaction. (c) Authigenic pore-filling kaolin associated with altered detrital muscovite (BSE). (d) Detrital tri-trioctahedral chlorite. The presence of ferromagnesian chlorite is confirmed through the EDX spectrum in Figure S1a. (e, f) Elemental mapping showing the mineralogical composition of intragranular pores. EDX spectrum of sudoite is visible in Figure S1b. Composite elemental maps are displayed in Figures S2, S3. Orange arrows denote secondary pore spaces. PL: plane polarized, CP: cross polarized, BSE: back scattered electron.
The SEM–EDX examination reveals that the kaolin minerals commonly occur, filling secondary pores and replacements along the cleavage planes of altered muscovite in the FB2a sandstones (Figure 2c, e, and f; Figures S1–S3). Kaolinization of muscovite is common in the studied coarse-grained samples. The consistent association between altered muscovite and kaolin minerals likely suggests the authigenic nature of the kaolin minerals. Kaolin aggregates typically form 5 to 10 μm-long, vermicular booklets composed of thin pseudohexagonal plates orderly stacked face to face (Figure 3a, b), which is consistent with the morphological features of kaolinite polymorph [Beaufort et al. 1998]. The individual platelets are homogeneous in size and morphology, with crystals of 5 μm wide and 0.1 μm thick. Importantly, kaolinite has undergone secondary dissolution (Figure 2f), as suggested by the partial dissolution of the edge of plates (Figure 3c).
SEM images in secondary mode of pore-filling minerals from the FB2a sandstones. (a, b) Authigenic vermiform crystals of kaolinite. (c) Partial dissolution of pseudohexagonal kaolinite (white arrows). (d) Mixture of boxwork-like texture and parallel pattern of sudoite crystals (white arrows). EDX spectrum of analysis points (arrows) is shown in Figure S1c.
SEM–EDX observations show that the ferromagnesian chlorite is present as either detrital particle or pore filling in the FB2a sandstones (Figure 2d–f; Figures S1–S3), as suggested by the presence of O, Si, Al, Mg, and Fe elements. These clay morphologies are also found in the MRS at the top of the massive sandstones [Aubineau et al. 2019]. Authigenic Al- and Mg-rich chlorites are exclusively pore-filling and widely observed in the FB2a sandstones but absent in the FB2b black shales (Figure 2e, f; Figures S1–S3). At higher magnification, these diagenetic minerals consist of densely packed subhedral, folded, and platy <5 μm crystals (Figure 3d), forming a chaotic arrangement where flakes are perpendicular to the substratum surface (boxwork-like texture) and/or parallel to the wall-rock surface (parallel pattern). Such morphological features resemble those reported for sudoite [Billault et al. 2002].
4.2. Mineralogy and chemical composition in the FB2 Member
The MRS, host sandstones, and black shales were analyzed using the XRD technique. Aubineau et al. [2019] pointed out the large contribution of illite-rich (R3) I-S MLMs in the MRS (both non-pyritized and pyritized), while clay minerals in the underlying host sediments are enriched in R0 I-S MLMs. These assemblages contain both 1Mt and 2M1 illite polytypes in the FB2 Member. Bulk mineralogy data show that the black shale facies have higher proportion of pyrite relative to sandstones, while dolomite is the dominant carbonate mineral in both lithologies [Aubineau et al. 2019].
The studied XRD profiles of oriented preparations of the <2 μm fraction in the AD state show characteristic 00l reflections of chlorite at ∼14.2–14.1 Å, 7.1 Å, 4.72 Å, and 3.54 Å that remain unaffected after glycolation (Figure 4). There is no significant difference in the mineralogical composition of chlorite between the lithologies within the same rock unit, which suggests that the MRS did not affect the chlorite composition. The intensities of the 002 and 004 reflections of chlorites in the FB2b unit are greater than the 003 reflection (Figure 4a), which is typical of tri-trioctahedral chlorite [Brindley and Brown 1980]. In contrast, the high intensity of the 003 reflection of chlorite (Figure 4c) coupled with the 060 reflection at 1.51 Å (Figure 4d) indicate the presence of di-trioctahedral chlorite in the FB2a unit [Billault et al. 2002]. The presence of 060 reflection at 1.55 Å also confirms the occurrence of tri-trioctahedral chlorite in the non-pyritized MRS and sandstones. Kaolinite in the FB2a MRS and sandstones is characterized by the 001, 002, and 060 peaks at 7.15 Å, 3.57 Å, and 1.49 Å, respectively. The 00l kaolinite reflections are sharp and narrow, implying their well-crystallized phase.
Representative XRD patterns of the <2 μm clay fraction of mat-related structures and host sediments in the FB2b Member (a, b) and FB2a Member (c, d). (a, c) Oriented preparations after air-drying (black lines) and glycolation (red lines). (b, d) XRD profiles of 060 reflection of randomly oriented preparations. The blue and green areas correspond to R0 I-S MLMs and R3 I-S MLMs, respectively, as previously described in Aubineau et al. [2019]; chlorite (Ch); illite/mica (I/M); gypsum (Gy); quartz (Q); calcite (Ca); sudoite (Su); barite (Ba); kaolinite (K).
Experimental (crosses) and modeled (lines) XRD profiles of representative samples after air-dried (AD) preparation and ethylene glycol (EG) saturation. (a) Black shale. (b) Sandstone. (Chlorite (Ch); R0 I-S MLM (R0 I-S); R3 I-S MLM (R3 I-S); illite/mica (I/M); illite polytype (1Mt); sudoite (Su); kaolinite (K)).
The XRD profile modeling (NEWMOD program) confirm that sudoite is completely absent in the black shales but occasionally present in the sandstones and siltstones, reaching about 28% in the <2 μm clay fraction (Figure 5; Table S1). In contrast, variable abundances of chlorite without relationships with any clay minerals are observed in the FB2 Member. Although there is no clear trend between sudoite and I-S MLMs, XRD modeled data show that the increasing amount of sudoite tends to be accompanied by decreasing R0 I-S MLMs in sandstones and siltstones. Moreover, sudoite contents seem to be high with increasing kaolinite contents. Nonetheless, these correlations might be biased as little XRD profile simulations have been conducted and absence of quantitative analyses.
The chemical composition, measured by EDX, and calculated structural formulas (based on 14 oxygen equivalents and with all Fe assumed as Fe2+ for chlorite group minerals) are shown in Table S2. The Mg–Fe–AlT ternary plot [Velde 1985] shows that chlorite in the FB2b black shales have a near-uniform compositions and fall between (Mg-rich chlorite) and chamosite (Fe-rich chlorite) endmembers (Figure 6a). When plotted in the trioctahedral half-vector representation of chlorite compositions [Wiewióra and Weiss 1990], the FB2b chlorites display heterogeneous composition due to their wide range of octahedral occupancy, varying from 5.25 to 5.95 atoms per half unit cells (aphuc), with a mean value of 5.58 aphuc (Figure 6b). In contrast, two distinctive chloritic minerals characterize the FB2a sandstones with a compositional gap between di-trioctahedral and tri-trioctahedral chlorites (Figure 6). While one group of FB2a chlorite falls in the tri-trioctahedral domain (mean octahedral occupancy near 5.65 aphuc) similar to that of the black shale facies, the octahedral occupancy of some FB2a chlorite is restricted to 5 aphuc, which is consistent with that of di-trioctahedral chlorite (i.e., sudoite; Figure 6b). In addition, we observed an increase in the octahedral Al content from 2.83 to 3.25 aphuc in sudoite, which is the expected range of ideal sudoite [Bailey & Lister 1989]. Nonetheless, the absence of quantitative data of the ferric iron content in chlorite structure may limit interpretations of the observed chemical trends (Figure 6). The data also dispersed along with the kaolinite–sudoite endmember mixing line.
Projection of the structural formulas of chlorite minerals from the FB2 Member. (a) EDX analyses of chlorite plotted in the triangular diagram Mg–Fe–AlT [Velde 1985]. (b) Chemical compositions of chlorite projected on the Si versus R2+ diagram, as designed by Wiewióra and Weiss [1990]. Data are expressed in atoms per half unit cell. The octahedral occupancy and total number of Al atoms per half unit cell are shown by contours.
4.3. FTIR spectroscopy
The MIR spectra of the bulk and <2 μm clay fraction in both MRS and host sediments from the FB2a unit were analyzed to distinguish the kaolin polymorphs using the hydroxyl-stretching band region between 3800–3200 cm−1 [e.g., Balan et al. 2005]. For both lithologies (Figures 7a and 7b), the bulk spectra exhibit similar vibrational features; the broad band with complicated lineshapes in the 3550–3750 cm−1 region is assigned to 2:1 dioctahedral phyllosilicates including mica, illite, and I/S MLMs [Farmer 1974; Madejová et al. 2011; Russell and Fraser 1994] and the broad doublet at lower frequencies (3509 and 3528 cm−1) diagnose the concomitant presence of sudoite [Billault et al. 2002; Madejová et al. 2011; Russell and Fraser 1994]. The <2 μm clay fraction for both MRS and sandstone also reveals the presence of kaolinite (3620, 3561, 3668, 3696 cm−1), corresponding to the four fundamental stretching vibrations of inner and inner-surface OH groups [Farmer 1974; Madejová et al. 2011; Russell and Fraser 1994].
Typical FTIR spectra of kaolin- and sudoite-bearing samples (bulk and <2 μm clay fractions) from the FB2a unit. The interval from 3800 to 3200 cm−1 characterizes the hydroxyl-stretching band region. (a) Mat-related structures. (b) Sandstones.
5. Discussion
5.1. Origin of sudoite
The study of MRS and host sediment samples has highlighted that the kaolinite and sudoite are one of the dominant authigenic clayey constituents in the non-pyritized MRS and sandstones from the FB2a unit. Lesser amounts of such clay minerals were observed in the siltstones, while they are almost absent in the pyritized MRS and black shales of the FB2b unit. Although the illitization process was microbially enhanced in the FB2 Member [Aubineau et al. 2019], this study reveals that the biological activity of MRS did not influence kaolinite and sudoite formations. However, the change in clay mineralogy is accompanied by a decreasing degree of fluid interaction with primary minerals, which is consistent with the reducing permeability from sandstones to black shales.
Kaolinite and/or smectite have been identified as the main precursor for the formation of sudoite due to their Al-bearing structural composition [Biernacka 2014; Hillier et al. 2006]. Our examinations point to direct evidence of dissolution features of vermiform kaolinite crystals, leading to the assumption that kaolinite could have locally supplied most of Al for sudoite neoformation. The occurrence of mixing lines between kaolinite and sudoite supports the transformation of kaolinite to sudoite. Formation of sudoite has been previously reported to occur in Precambrian sandstones when oxidized and relatively acidic fluids from basement rocks interact with overlying porous sediments [Beaufort et al. 2005; Billault et al. 2002; Riegler et al. 2014]. In addition, kaolinite–illite–sudoite paragenesis thermodynamically forms under relatively acidic and oxidizing conditions [Kister et al. 2005]. In such scenarios, circulating basinal brines along faults and fractures become saturated with respect to ions such as Fe2+, Mg2+, and K+. The Franceville sub-basin has undergone important circulations of oxidizing basinal fluids [Bankole et al. 2016], which may be suitable for sudoite formation. Accordingly, we infer that sudoite crystallization in the FB2a sandstones was promoted by fault-controlled Mg-rich brines that originated from interaction with primary minerals of the underlying Archean crystalline basement rocks. Nonetheless, we cannot completely rule out the possibility that Mg ion was supplied during the dissolution of dolomite since acidic conditions could have destabilized Mg-bearing carbonates present in the MRS and host sediment sample.
5.2. Mineral paragenetic sequence in the FB2 Member
Mineral paragenesis allows to unravel the relative timing of distinct diagenetic events in any sedimentary basin. The occurrence of stylolites, concavo-convex contacts between the detrital quartz grains, and closure of primary pore spaces indicate that the FB2a sandstones have undergone moderate to strong degrees of compaction (Figure 2a, b). The presence of kaolinite within the secondary pore spaces and along cleavages of altered muscovite suggest their authigenic nature (Figure 2c). Kaolinization of the muscovite may have occurred at shallow burial depth during early diagenesis when the circulation of meteoritic fluids was still promoted [De Ros 1998]. Consequently, the altered muscovite released potassium ions (K+) into the pore system, which may have resulted in precipitation of illite within the pore space, implying that kaolinite likely predates the authigenic illite. Moreover, the occurrence of 1Mt illite polytype suggests environmental conditions marked by high fluid/rock ratios, while the 2M1 illite polytype is likely due to high burial conditions near the transition between diagenesis and metamorphism or remnants of detrital input [Meunier and Velde 2004]. The presence of trace amounts of R0 I-S MLMs in the sandstones indicates low temperature conditions [e.g., Velde et al. 1986], supporting dominant detrital source of 2M1 illite phase. The progressive conversion of smectite into randomly interstratified and ordered I-S MLMs in sedimentary basins is mostly kinetic controlled, and depends on several factors including burial depth, time, temperature, pressure, primary composition, porosity, permeability, and K availability during diagenesis [Cuadros and Linares 1996; Eberl and Hower 1976; Howard and Roy 1985; Hower et al. 1976; Huang et al. 1993; Inoue 1983; Li et al. 2016; Velde et al. 1986; Whitney 1990, among others]. In light of these considerations, the occurrence of the 2.1 Ga old R0 I-S MLMs suggests that the incomplete illitization process was not only related to insufficient K [Bankole et al. 2018; Ngombi-Pemba et al. 2014] but also to the low porosity and permeability of the FB sandstones.
The diagenetic tri-trioctahedral chlorite could have crystallized at the expense of the detrital ferromagnesian chlorite in the FB2a sandstones. This is, however, unlikely due to the unaltered nature of the detrital chlorites (Figure 2d), suggesting direct precipitation of the diagenetic tri-trioctahedral chlorite from pore fluids within the pore space. This would therefore imply a higher iron content in the structural composition of authigenic chlorite relative to the detrital component [Beaufort et al. 2015]. On the other hand, our samples contain only one population of tri-trioctahedral chlorite (Figure 6), favoring the formation of the diagenetic chlorite from dissolving detrital biotite in the FB2 Member. This inference is supported by the presence of trace amounts of titanium, released from the chloritization of biotite [Ferrow and Baginski 1998], in the structural formulas of FB2 tri-trioctahedral chlorite (Table S2).
The moderate to high degrees of rock compaction and occlusion of pore spaces may have limited the growth of sudoite. In addition, the small amount of octahedral iron in the structural composition of sudoite is likely responsible for the small particle size and subhedral to anhedral morphology of plates [Billault et al. 2002]. This probably explains the overall chaotic arrangement of sudoite crystals in the FB2 Member. The presence of illite between quartz grain contacts indicates its formation or precursor before significant compaction or during compaction (syn-compaction), while the absence of chloritic minerals between quartz grain contacts suggests their formation postdates compaction. However, it is difficult to constrain the sequence of formation of the diagenetic illite and chlorite within pore space based on their textural relationships. Sudoite and illite were observed cross-cutting each other, suggesting that both minerals may have precipitated concurrently (Figure 2e–f).
5.3. Implications for diagenetic history of the Franceville sub-basin
While putative detrital kaolinite particles have been invoked in the overlying Francevillian sediments [Hao et al. 2021], our study reveals the presence of authigenic kaolinite in the Franceville sub-basin through a combination of petrographic, microscopic, and mineralogical analyses. Authigenic kaolin polytypes promote the investigation of burial history of sedimentary basins since a continuous sequence of kaolinite-to-dickite transformation occurs during diagenesis [Beaufort et al. 1998; Lanson et al. 2002]. Through a dissolution–precipitation mechanism, vermicular kaolinite crystals can be progressively replaced by blocky dickite, which is the stable kaolin polymorph during late diagenesis [Zotov et al. 1998]. The main parameters controlling the reaction are the burial depth and temperature, which establishes that the first step of kaolinite–dickite transition usually takes place at burial depth about 2500 m in conditions of normal geothermal gradient. The lack of petrographic and mineralogical evidence of dickite in the FB2 Member, therefore, suggests that the Francevillian sediments have not experienced deep burial diagenesis.
The kaolinite–illite–sudoite paragenetic sequence in the FB2 Member can be explained by an increase of the K+∕H+ and Mg2+∕(H+)2 activity ratios with a higher Mg2+∕(H+)2 ratio than K+∕H+ ratio [Kister et al. 2005]. This kinetic fluid–rock reaction path predicts that kaolinite does not completely dissolve and sudoite crystallizes concomitantly with 1Mt illite, which is consistent with our petrographic observations. Such reaction occurs under relatively acidic and oxidizing conditions with the involvement of basinal fluids circulating along faults and fractures [Kister et al. 2005]. Through upward circulations, similar diagenetic hydrothermal brines are known to have promoted the U mobilization and its enrichment at the reduction front of the FA–FB transition in the Franceville sub-basin [Bankole et al. 2016; Gauthier-Lafaye and Weber 1989]. These fluid circulations could have been favored by major tectonic structures during basin inversion [Gauthier-Lafaye and Weber 1989]. Accordingly, the timing of sudoite formation in the overlying FB2 sandstone facies could have coincided with upward migrating fluids that resulted in U mineralization during the late burial stage of the Francevillian Group. One possible assumption for the lack of sudoite in the FA sandstones would be the complete absence of kaolin minerals [Bankole et al. 2015], which contrasts with the mineralogical composition of other sandstone-hosted U deposits in the Paleoproterozoic Athabasca basin (Canada) and McArthur basin (Australia) [Beaufort et al. 2005; Billault et al. 2002; Kister et al. 2005; Percival and Kodama 1989; Riegler et al. 2014; Truche et al. 2018].
6. Conclusion
The Paleoproterozoic FB2 Member in the Franceville sub-basin consists of sandstones and black shales intercalated with siltstone beds that were deposited in relatively shallow environments, which triggered the development of photosynthesizing microbial mat communities [Aubineau et al. 2018; Reynaud et al. 2017].
This study documents the clay mineralogy in remnants of microbial mats and host sediments to better constrain the mineral paragenesis in the FB2 Member. The textural relationships and mineralogical compositions of the MRS and host rocks indicate that multiple diagenetic processes affected the FB2 Member. In addition to the FB2 detrital clay minerals and 2M1 illite, as well as authigenic minerals (1Mt illite, I-S MLMs, and ferromagnesian chlorite), the FB2a MRS and sandstones and also FB2b siltstones to a lesser extent are further characterized by vermiform kaolinite and sudoite. Specifically, the medium- to coarse-grained sediments reveal the co-occurrence of kaolinite, 1Mt illite, and sudoite, which supports a fault-controlled environment with a high fluid/rock ratio under acidic and oxidized condition. The absence of kaolinite and sudoite in the pyritized MRS and black shale facies highlights, thus, the role of porosity and permeability of sedimentary rocks in the clay authigenesis in the FB2 Member.
Acknowledgments
We deeply thank the Gabonese Government, CENAREST, the General Direction of Mines and Geology, and the Agence Nationale des Parcs Nationaux of Gabon for the logistic supports. La Région Nouvelle Aquitaine, France, the European Union (ERDF), and the French Embassy in Libreville, Gabon, supported this work. We would like to acknowledge Professors P. Mouguiama Daouda, J. C. Balloche, L. White, and R. Oslisly for their support. The authors also thank C. Boissard, C. Fontaine, C. Laforest, and P. Recourt for laboratory support at the universities of Poitiers and Lille. The authors are particularly thankful to Professors A. Meunier and D. Beaufort for scientific discussions and C. Lebailly for administrative support.