Depositional Setting of Archean BIFs from Congo: New Insight into Under-Investigated Occurrences

Archean banded iron formations (BIF) represent a major contributor to better constraining and assessing the paleogeography and evolution of Archean cratons. In this context, we conducted an exhaustive sampling and analysis campaign of BIF units in the Congo Craton, covering several greenstone belts within the Ivindo, Kelle-Mbomo, and Chaillu blocks. The REE + Y patterns suggest: (1) Interaction of seawater with Fe-oxyhydroxides, as illustrated by strong REE enrichment coupled with La and Y enrichment; (2) contributions from high-temperature (>250 °C) hydrothermal fluids, illustrated by positive Eu anomalies; and (3) detrital input as suggested by relatively consistent REE concentrations and a chondritic Y/Ho ratio. These observations suggest a typical environment of Algoma-type BIF deposition. Moreover, assessment of the Ce anomalies in a combination of HREE enrichment indicates that some basins in the Chaillu and Ivindo blocks may have known potential oxygen-rich episodes in the early Archean during the deposition of these BIFs.


Introduction
Algoma-type banded-iron formations (BIF) are chemically formed sedimentary units consisting of alternating chert and iron-rich layers [1]. They are generally stratigraphically related to bimodal volcanic and sedimentary rocks in deformed and metamorphosed Archean to Early-Proterozoic greenstone belts [2,3]. These Fe-rich sedimentary units have long been considered a significant contributor for constraining and assessing the geochemical evolution of atmosphere and hydrosphere throughout the Precambrian [3][4][5][6]. Consequently, it is crucial to document and assess their geochemical characteristics.
Despite successive chemical transformations during diagenesis, or related to multiple metamorphic recrystallizations, the REE + Y abundance of BIFs reflects a geochemical signature indicative of their primary depositional environment [7][8][9]. This includes: (1) Interaction of seawater with Fe-oxyhydroxide and/or greenalite, as suggested by their strong REE enrichment coupled with positive La and Y anomalies; (2) contribution of high-temperature (>250 • C) hydrothermal fluids as suggested by positive Eu anomalies; and (3) detrital contributions, suggested by relatively flat REE patterns, high REE contents, and chondritic Y/Ho ratios (Y/Ho ≈ [25][26][27]. Here, we investigate the geochemical signatures of several BIFs from the Ivindo, Kelle-Mbomo and Chaillu Archean terrains within the Congo-Gabonian Craton in the Republic of the Congo (Figure 1). We also assess the depositional context of the BIFs within their host-rocks. In detail, the purpose of this paper is two-fold: (1) Document the BIF's

Regional Setting
The northwestern part of the Congo-Gabonian Craton in the Republic of the Congo comprises several exposed Archean nuclei, such as the Ivindo Block, the Kelle-Mbomo Block, and the Chaillu Block (Figures 1 and 2). In these blocks, the Archean rocks consist of three main suites and groups: (1) The 2.8 Ga Belinga Group of supracrustal rocks that are remnants of Mesoarchean greenstone belts [21,22]; (2) a Mesoarchean Gneissic Complex with tonalite-trondhjemite-granodiorite (TTG) affinities; and (3) a Mesoarchean to Neoarchean Granitoid Suite composed of magmatic bodies of sanukitoids and K-rich granitoids.
In the Republic of the Congo, the greenstone belts are older than 2850 Ma, as they were deformed and metamorphosed between 2800 and 2700 Ma (U-Pb dating on zircons) [23][24][25]. At the craton scale, these ages suggest an intense heat transfer from mantle to crust before 2.80 Ga and a major orogenic event around 2.75 Ga [23][24][25].

The Ivindo Archean Block
The Ivindo Archean Block (IAB) lies in northern Gabon, northwest Congo, and southern Cameroon (Figure 2A). It consists of a granite-greenstone domain composed of a 3.0-2.7 Ga granitic Archean basement and ca. 2.8 Ga north-and west-trending greenstone belts forming the Belinga Group [20][21][22]26]. Several BIFs occur in this group and their age is estimated from 2866 to 2750 Ma [21]. In the Republic of the Congo, the greenstone belts surrounding BIFs have been essentially dated between 3080 and 2845 Ma (U/Pb dating on zircons) [23][24][25]. In southern Cameroon, however, whole-rock Pb-Pb dating yielded an age of 2679 Ma, interpreted as a metamorphic-reset age, suggesting that BIF deposition predates this age [16]. The basement includes granite, gneiss, and migmatite which underwent amphibolite to granulite facies metamorphism. Greenstone belts occur as disrupted belts in the Badondo, Avima, Nabeba, Egbala Monts, and Elogo areas. They are composed of ultrabasic rocks, amphibole-pyroxene rich rocks, amphibolite, metasedimentary rock, and BIFs, mostly metamorphosed from lower greenschist to amphibolite facies. Deformed and/or migmatized TTG orthogneiss, with greenstone belt xenoliths and sanukitoid-like diorites and granodiorites, and syn-to post-tectonic biotite are common.
Mesocratic dioritic gneiss and felsic meta-volcanic rock within the Badondo greenstone belt yielded ages of 2914 Ma and 3080 Ma (U/Pb dating on zircon), respectively [25], and a meta-volcanic tuff within the Avima greenstone provided an age of 2878 Ma (all U/Pb dates on zircons) [25].

Description of the Sampled BIFs
From west to east, the Badondo and Avima Monts are approximately10-km-long oxide-facies BIF units consisting of alternating Fe-oxide-rich (magnetite and hematite rich) and chert bands affected by regional greenschist metamorphism [32][33][34]. They are mainly surrounded by schist and accessory unsubdivided amphibolite. Two main types of schist were identified in the field: (1) bright yellow mica schist suggesting a felsic metavolcanic affinity; and (2) greenish actinolite-chlorite schist. The latter amphibolite-schist sequence appears to have undergone low-grade greenschist regional retro-metamorphism [25].
In the Nabeba and Letioukabala Monts areas, oxide-facies BIFs at Egbala and Nabeba are surrounded by amphibolite and greenish actinolite-chlorite schist. Iron-carbonate bearing minerals such as siderite occur in the Nabeba BIF [34].
In the Elogo area, oxide-facies BIFs are surrounded by ultramafic rock, amphibolite, talc schist, and greenish actinolite-chlorite schist. Chert and Fe-rich layers, cm-to mm-scale, are dismembered and weakly oxidized.

The Kelle-Mbomo Archean Block
The Kelle-Mbomo Archean Block (KMAB; Figure 2b), from which a schistose komatiitic basalt yielded an age of 2925 Ma for the Youkou greenstone belt [23], lies in the central part of Congo and Gabon. It is composed of deformed and/or migmatized TTG orthogneiss, northwest-to west-trending greenstone belts, mesocratic to melanocratic sanukitoid-like gneiss, and syn-to post-tectonic gray to pink biotite granite.

Description of the Sampled BIFs
From north to south in the Youkou area, BIF consists of cm-to dm-scale chert bands interbedded with minor magnetite to hematite layers. This unit is surrounded by talc and greenish actinolite-chlorite schist, with accessory amphibolite.
In the Obélé, Odia, and Kékélé areas, greenstones are characterized by oxide-facies BIF with amphibolite and accessory ultramafic rock and greenish actinolite-chlorite schist.
In the Ngoyboma area, BIF consists mainly of cm-to mm-scale, magnetite/hematite bands interbedded with minor mm-scale chert layers. They are associated with amphibolite, mesocratic gneiss and greenish actinolite-chlorite schist.

The Chaillu Archean Block
The Chaillu Archean Block (CAB; Figure 2C) is continuous throughout the southern part of Gabon and Congo. It consists of ca. 3.1 to 2.54 Ga [35,36] deformed and/or migmatized TTG orthogneiss, northeast-and north-trending greenstone belts, mesocratic sanukitoid-like granodioritic gneiss, and syn-to post-tectonic, gray to pink, biotite granite.
Here, greenstone belts are disrupted in the Létoufou Mont (Bikélélé and Zanaga areas) and the Mayoko area. These supracrustal rocks consist mainly of oxide-facies BIF (magnetite and/or hematite), paragneiss, unsubdivided amphibolite, and ultramafic rock.

Description of the Sampled BIFs
From west to east in the Mayoko area, a felsic meta-volcanic rock and an amphibolite were dated at 2883 Ma and 2864 Ma (U/Pb dating on zircons), respectively [22]. Oxide-facies BIFs generally comprise well developed, cm-to mm-thick chert and magnetite/hematite bands. They are surrounded by amphibolite, talc schist, and greenish actinolite-chlorite schist. Ultramafic rocks are rare.
The Bikélélé BIFs consist of cm-to mm-thick silicate-rich chert bands interbedded with oxide-rich bands. They are interstratified with amphibolite, ultramafic rock, and greenish actinolite-chlorite schist and are surrounded by mesocratic gneiss.
The Zanaga BIF consists of mm-thick chert and magnetite/hematite layers with local amphibolitic material. Locally disrupted chert bands are described as "dirty" layers, suggesting the presence of detrital material. Amphiboles occur at the boundaries of the Fe-rich and Si-rich bands ( Figure 2A).

Whole-Rock Analyses
As this study is a first approach for these new BIF occurrences, we followed suggestions in recent work, to adopt whole-rock analyses for representing the prevailing local to regional paleoenvironmental conditions [6]. It is known that both magnetite and hematite as well as Si-rich layers from BIFs retain the primary REE + Y geochemical signature [7,9,37], though detrital components, which cannot be precisely identified in LA ICP MS analyses, can also influence the REE + Y pattern [7]. However, we consider that this detrital component is part of the primary signature, thus constraining part of the regional paleogeography.
Trace and rare earth elements, such as Ba, Ce, Cr, Cs, Dy, Er, Eu, Ga, Gd, Ce, Hf, Ho, La, Lu, Nb, Nd, Pr, Rb, Sm, Sn, Sr, Ta, Tb, Th, Tm, U, V, W, Y, Yb, and Zr were determined by lithium borate fusion prior to acid (HNO 3 and HCl) dissolution and followed by inductively coupled plasma-mass spectroscopy (ICP-MS) analysis (ALS-ME-MS81 method). As, Bi, Hg, In, Re, Sb, Se, Te, and Tl were determined by aqua regia digestion followed by ICP-MS analysis (ALS-ME-MS42 method). Finally, C and S were analyzed by LECO furnace (C-IR07 and S-IR08 methods respectively).
Various ALS standards were used and data quality was assured by running standards CCU-1d, EMOG-17, GRE-3, GS310-7, GS311-3, and GS910-4 between samples as unknowns. The detection limits varied from 0.01 to 0.03 ppm for REE + Y. The Queensland alluvial shale composite (MUQ) [38], which reflects a bimodal, mafic, and felsic component typical of Archean greenstone belts, was used for normalizing the REE + Y concentrations, in order to minimize the influence of potential terrigenous input.

REE + Y Systematics
Despite a few exceptions, the REE + Y normalized data show relatively uniform REE + Y patterns (Figures 4-6; Table S1).    The EW5055 Obélé BIF sample shows high Zn (55 ppm) and the EW5095 sample from the Odia BIF unit has a high V (126 ppm) content. In

Primary Signatures
Several parameters (primary versus secondary inputs) can influence the REE + Y patterns. First of all, it has been demonstrated that the detrital component can strongly influence the primary signature by flattening the REE + Y spectra and mitigating seawater anomalies [6]. However, this input is part of the syn-depositional sedimentary component of the BIF, contributing to its REE + Y budget, and thus has to be studied. Moreover, normalization to MUQ (terrigenous composite for Archean greenstone belt) aims at minimizing the influence of potential detrital input on the REE + Y signal.
Secondly, secondary weathering can affect major-and trace-element budgets. Regarding the REE + Y content, Ce is considered as one of the most robust redox proxies due to the general insolubility of REE at Earth surface conditions, by its immobility during metamorphism [7], and by the fact that Ce becomes more immobile when oxidized [39]. Some studies on BIFs suggest that negative excursions of Ce/Ce* may reflect atmospheric oxygenation [40]. Recent work has highlighted that a negative Ce anomaly, when combined with extreme LREE enrichment relative to HREE, reflects secondary weathering in BIFs sampled from the Barberton Greenstone Belt in South Africa [39]. Thus, Ce must be studied carefully for potential post-depositional alteration in complement to LREE enrichment, in order to reveal such processes and modifications of the REE spectra.
However, in our case, the consistency of REE patterns suggests that the primary signal is relatively well preserved in most BIF units. Many samples show depletion in LREE relative to HREE, as well as positive La, Y, and Eu anomalies, suggesting BIF deposition of the BIF units in seawater under the influence of high-temperature (>250 • C) hydrothermal fluids [7,41,42]. This type of environment is typical of Algoma-type BIFs [43,44].
Despite these observations, samples from the Ngoyboma BIF unit and a few from the Youkou BIF unit, both of which petrographically show strong hematization, have an erratic or flatter REE-Y pattern that suggests potential REE-remobilization, related to either late alteration, or lateritic weathering. In addition, it appears that HREE are remobilized more easily than LREE (cf. erratic pattern), which was observed elsewhere [45,46].

Assessment of High-Temperature (>250 • C) Hydrothermal Fluids
Most samples show the influence of seawater and high-temperature hydrothermal fluids through their REE + Y patterns and La, Y, and Eu anomalies [41,42], as mentioned above. Assessment of a high-temperature hydrothermal-fluid component during BIF precipitation uses Sm/Yb versus Eu/Sm ratios [7,47]. A binary mixing line was calculated between North Pacific seawater samples [48] and an Archean chert breccia sample from the 2.7 Ga Abitibi greenstone belt [44] shows an extremely strong Eu anomaly [7]. This suggests that the sample location was exceptionally close to the hydrothermal source ( Figure 8).  [44]) and seawater data are from [48].
Most samples are distributed along, or a little above, a mixing line like the Sm/Yb one in Figure 8. This may reflect the presence of high-pressure residual phases, such as garnets, amphiboles, etc., which are relatively common in metamorphic sedimentary rocks from greenstone belts. However, all samples indicate a consistent hydrothermal influence of up to 10%. This is similar to Algoma-type BIFs in Canada or Brazil [7,9] and can be used as a proxy for the locus of BIF-precipitation relative to a hydrothermal-source vent. It also indicates a relative distance for an Algoma-type BIF depositional site.
Interestingly, the distribution of BIFs within each Archean block may indicate their geographical location relative to the hydrothermal plumes from which BIFs are deposited [7]. Thus, in the IAB, the Badondo BIF unit appears to be less affected by hydrothermal-vent processes (farther away) than the Egbala and the Avima BIFs ( Figure 8B).
In the KMAB, the Youkou BIF unit is less affected by hydrothermal fluids than the Kékélé/Obélé and Odia BIF units, suggesting that the Odia unit was located closer to the hydrothermal source ( Figure 8C).
Similarly, in the CAB, the Mayoko BIF unit appears less affected by hydrothermal processes than, first, the Bikélélé and then the Zanaga BIFs. This suggests that the Mayoko BIF unit was probably deposited farther way from a hydrothermal source than the Bikélélé and the Zanaga BIFs ( Figure 8D).
However, estimating a distance to hydrothermal sources for the other BIFs would be a bit adventurous, requiring much more data.

Assessment of Redox Conditions
In modern oxic seawater, a relatively well-developed negative Ce anomaly is generally due to the scavenging of Ce by Fe-Mn oxyhydroxides, leading to the oxidation of Ce 3+ to Ce 4+ [48]. In sub-oxic to anoxic seawater, such REE removal does not occur because of the reductive dissolution of settling Mn-(Fe) rich particles [3,[49][50][51]. However, in an oxygenated closed basin, Bau [51] demonstrated that pH conditions strongly affect Ce anomalies, producing a positive Ce excursion when pH ≤5 and a negative Ce anomaly when pH ≥5.
In the context of Algoma-type BIF precipitation, Fe-oxyhydroxide precipitation causes scavenging of La, Y, (Gd), and LREE, providing a pure seawater geochemical signature. However, in a restricted basin without continuous recharge of the REE-Y budget induced by seawater excursions, negative La and Y anomalies can be observed (Figure 9). Thus, basin dynamics can be assessed using Ce, La, and Y anomalies and Y/Ho values below the chondritic value of about 25. Moreover, several negative Ce anomalies are reported from the IAB and CAB (Mayoko and Zanaga BIF samples). They are neither related to negative La and Y excursions nor to low Y/Ho values, suggesting oxic, open-basin depositional conditions (e.g., MO1160-2 from the Mayoko BIF unit, or ZA6210-1 and ZA6210-3 from the Zanaga unit; Figure 8), as recently observed in several Precambrian metasedimentary units [52][53][54].
These observations indicate a relatively active basin in the IAB and CAB realm, with: closed to open basins that may have been affected by sea-level variations; -pH variations; -anoxic to oxic water conditions.
The KMAB, however, may reflect only variations of anoxic to locally oxic basin conditions (Figure 9b).
Meanwhile, redox conditions may have constrained the relative depth of sample deposition, suggesting that samples indicating oxic conditions would have been deposited in the photic zone, whereas others would be from deeper, anoxic areas [54].

Sources of the Detrital Components
Detrital components constitute a relatively important feature in basins as they may be used as a proxy for assessing proximal to distal depositional settings in terms of the basin margins, the latter normally being more influenced by such components than the center of a basin far from the coast.
In the IAB, the main detrital sources in these BIFs are felsic rock types ( Figure 10). Based on regional geological data, the bright yellow mica schist might represent a possible source for this felsic signature, as already suggested in the literature [25]. Within the KMAB, most samples show a basic affinity (Figure 10), which is again supported by geological mapping at the block scale where mafic rocks are prominent.
In the CAB, the Zanaga BIF unit shows high concentrations and very strong positive correlations (r > 0.8) between Al 2 O 3 , Na 2 O, TiO 2 and K 2 O, Cr 2 O 3 , Sc, V, Cr, Co, Ni, Ga, Sr, and Zr (Figure 7; Table A1). Strong correlation in Al 2 O 3 , Na 2 O, TiO 2 , and K 2 O is mainly related to weathering of silicate minerals, reflecting river input during sedimentation [42].
The Zr/Hf ratio of 38 confirms the influence of river sediment in this BIF unit as it is relatively close to the chondritic value [56]. These observations suggest the relatively important input from variable detrital sources for this BIF. Figure 10 shows the existence of at least three distinct types of detrital contamination (felsic-intermediate, intermediate-basic, and basic sources) in this BIF unit. Moreover, the ZNDD182-2, ZA6119, and ZNDD209-6 samples, which show variable REE-Y spectra, may indicate a stronger detrital input relative to the other samples, as REE-Y spectra can show negative La and Y and flatter patterns under a detrital influence [7,46,47]. Thus, sample ZNNDD182-2, with a metabasic affinity, is described petrographically by alternating chert layers and greenish material, such as unsubdivided amphibolite, and chlorite is consistent with a detrital source.
Regarding major-and trace element concentrations, high Sc, Mn, Ti, Cr, Co, Ni, and V contents may indicate the leaching of underlying Archean komatiite, tholeiite, and basalt by the vent fluids during Fe-oxyhydroxide precipitation [57,58].
In addition to the fact that Zn may be sourced from mafic volcanic units, it has been suggested that relatively high Zn [59] contents in Archean submarine sedimentary units may be the result of biological activity. Thus, despite ferric-oxyhydroxide particle adsorption, Zn may have occurred in a bio-available form. Biological activity may have occurred at some point in the Zanaga (ZNDD182-2), Obélé (EW5055), and Egbala (SE4471-1 and ME4516) basins based on their relatively high Zn concentrations.

Conclusions
We present the first relatively exhaustive geochemical datasets for BIFs from the Congo Craton. We sampled and analyzed the Ivindo, Kelle-Mbomo, and Chaillu Archean blocks and BIF units. Their study provided insights into their typical stratigraphic and depositional setting in greenstone belts and provided information concerning the Craton's paleo-geomorphology (distribution of Archean basins and relative distances from basin borders and hydrothermal sources).
Despite the strong tropical weathering observed in Congo, most of the sampled BIF units show LREE depletion relative to HREE and positive La, Y, and Eu anomalies, suggesting deposition in a marine basin under the influence of high-temperature (>250 • C) hydrothermal fluids. This is the typical of the environment of Algoma-type BIFs.
We assessed redox and pH conditions of these paleo-basins based on Ce anomalies and REE + Y patterns. This suggested oxic, open-basin conditions for the Ivindo and Chaillu blocks (Mayoko and Zanaga BIF samples). This observation is important as it suggests that free oxygen was present in early Archean basins.
Algoma-type BIFs act as a favorable host rock for iron mineralization (intrinsic geochemical composition) and orogenic gold within several Archean cratons. Moreover, they may be used as a proxy for hydrothermal-related occurrences regarding their corresponding distance to the hydrothermal vent sources.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/min12020114/s1, Table S1: Chemical composition of the sampled BIF samples-major element (wt.%) trace elements (ppm).  We further acknowledge the contributions of two anonymous reviewers, whose constructive comments significantly improved the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A
Appendix A: Linear inter-element correlations for the Zanaga BIF unit. Bold characters highlight significant correlations between elements.