The Precambrian rocks located to the south of the sWAC comprise both an Archean-age nucleus and paleoproterozoic-age volcano-sedimentary bands within the Baoulé-Mossi domain (Figure 1(a)). The present structural architecture of these rocks results from the interplay of tectonics: Leonean (3.3 - 3.2 Ga) and Liberian (2.9 - 2.7 Ga) tectonics in the nucleus and Eburnean orogeny in the Baoulé-Mossi domain. This geological activity led to folding, faulting, and partial melting. Subsequent Eburnean tectonics at 2.1 Ga resulted in the recycling and remobilization of ancient Archean rocks, coupled with the generation of juvenile paleoproterozoic magma [15]-[18]. In Burkina Faso, the paleoproterozoic rocks represent 80% of the surface (Figure 1(b)), are structured into NNE-trending greenstone belts of volcanic flows and epiclastic rocks [19]. These belts span in age from 2.25 to 1.9 Ga [1], with the pinnacle of volcanic activity estimated between 2.2 and 2.16 Ga [20][21]. Sediments within the intrabelt region have been dated from 2170 Ma to 2107 Ma [11][14].

The paleoproterozoic basement underwent NW-SE-oriented crustal shortening between 2150 and 2095 Ma, coinciding with voluminous felsic intrusions [9][10][22]. As a result, two major NNE-oriented crustal faults and other minor plurikilometric offsets in the NE direction were established to accommodate the overall shortening [23] (Figure 1(b)).

Within the study area, the KLB manifests as a narrow belt confined to the southeastern limits of the Houndé volcano-sedimentary trend. The lower sequence consists of basaltic lavas, projections and gabbros of tholeiitic composition at the base, transitioning to an intermediate suite to minor rhyodacite at the top. All are intersected by diorite and granite plutons. Mafic lavas are prominently exposed as mountain ridges in the northern half of the Kampti district (Figure 1(c)). In contrast, andesite pyroclastic products rich in plagioclase megacrysts dominate the southern area [14][24]. The KLB is flanked by the high-strain Boni fault [25][26], serving as a domain boundary between the Tarkwaian detritic sediments [26] and the volcanogenic sediment-dominated Bambela basin (Figure 2).

Figure 1. (a) Simplified geological map of the sWAC [1]; (b) Simplified geology of Burkina Faso outlining main lithologies and tectonic features [9]; and (c) Geological overview of the KLB.

Figure 2. Geological cross-section of the KLB showing the main lithological units and their relationship to faults.

Petrographic examinations were carried out on systematically collected fresh volcanic materials, including flows and projections, as well as granite facies, with the aim of studying their composition. Major and trace elements were measured by ICP-AES/MS. Sm-Nd and Rb-Sr isotope analyses were performed on total-rock powders. The whole-rock geochemistry (seven samples) was performed at the AcmeLabs facility in Canada, and Sm-Nd isotopes (four samples) at the Geological Sciences Laboratory of the University of Cape Town, Rondebosch (South Africa). Laser ablation techniques were employed to obtain U-Pb isotopes from zircon samples originating from two granite plutons at the Earthlab laboratory of the University of the Witwatersrand (South Africa).

The whole-rock Sm-Nd protocol follows the procedure by [27], which involved dissolving rock powder in concentrated HF + HNO₃ within closed Teflon beakers at 140˚C for 48 hours. Sequential separation and purification of Nd were achieved using a cation-exchange column, conditioned and eluted with dilute HCl to minimize fractionation effects. Isotope ratio analyses were conducted using a Nu Instruments HR multicollector ICP-MS (equipped with a DSN-100 desolvating nebulizer) in the MC-ICP-MS facility, ensuring high-precision measurements of Nd isotopic ratios. Reproducibility of Nd standards JNdi-1 returned a mean ¹⁴³Nd/¹⁴⁴Nd ratio of 0.512115. Normalisation of Nd isotope values to a value of 0.512115 (JNdi-1) followed the method described by [28], with corrections for Sm interference using the measured signal of ¹⁴⁷Sm and natural Sm. Instrumental mass fractionation was corrected using the exponential law and a ¹⁴⁶Nd/¹⁴⁴Nd ratio of 0.7219. For zircon separation, standard procedures involved crushing and pulverising rocks, followed by sieving at <300 µm. Zircons were then separated, mounted in epoxy, and examined using backscatter imaging and cathodoluminescence to identify suitable locations for laser ablation. U-Th-Pb isotopic data acquisition employed a 193 nm resolution ArF excimer system coupled to an ICP-MS. Measurements were conducted in low-resolution and electrostatic scanning modes, with single-point analysis using a laser repetition rate of 8 Hz and a fluence of 2.5 Jcm⁻². Laser sampling occurred in a dual-volume ablation cell, with a mixed helium-argon atmosphere and a small volume of nitrogen to enhance signal stability and sensitivity. Calibration utilized zircon GJ1, and certified reference materials (CRMs) like zircon Temora-2 and Plešovice served as unknowns for evaluating measurement accuracy and precision.

Common Pb corrections are applied using the difference between measured and expected ²⁰⁷Pb/²³⁵U and ²⁰⁶Pb/²³⁸U ratios. The corrected ratios were plotted on Wetherill Concordia diagrams [29], allowing the effects of common Pb, radiogenic Pb loss, inheritance, and mixing of zircon domains of different ages to be assessed. For the calculation of crystallization ages, analyses with less than 85% concordance were excluded. Discordant analyses defining a linear trend were evaluated using a Discordia line. Its intersections with the Concordia curve can constrain the crystallization age and the age of a subsequent Pb-loss event or reopening of the isotopic system [29]. Concordia ages were calculated from the retained analyses by combining the two U-Pb systems and accounting for analytical uncertainties. The results are reported with their uncertainties, the number of analyses, the MSWD, and the associated probability. An MSWD below five was considered indicative of an acceptable fit, provided that the result was consistent with the petrographic and geological observations [30].

The predominant geological feature in the study area is the volcanic sequence, characterized by green to greyish basaltic lava flows and andesitic projections exhibiting common pseudo-pillow textures (Figure 3(a)). The microlitic, glomeroporphyric, and microporphyric textures consist of spherules of actinote and green hornblende. Feldspar microlites, often zoned, display anorthoclase cores rimmed by albite. The feldspars predominantly form a cloudy mass, undergoing metasomatic kaolinization. Porphyritic and vesicular andesite flows are commonly associated with volcaniclastic rocks. The micro-phenocrysts, poorly preserved phenocrysts, and microlitic matrix exhibit zoned plagioclase that has recrystallized to aggregates of secondary epidote, chlorite, carbonate, and quartz. Vacuoles are filled with aggregates of those secondary minerals. The volcaniclastic rocks mainly consist of pyroclastic units with abundant plagioclase and hornblende as lithic fragments and/or minerals embedded in a matrix of andesitic composition (Figure 3(b)).

The KLB has a suite of intrusive bodies, notably elliptic plutons and granitoids. Elliptic plutons like Yelonhira and Galgouli are characterized as mafic to ultramafic cumulate complexes [31] and are distinguished by their high magnetic susceptibility, setting them apart from other intrusive bodies in the region. In contrast, the Eburnean granitoid such as the monzodiorite, amphibole and mica-rich granite are NW elongated boudinaged plutons. It is grey to pinkish due to the abundance of coarse k-feldspar (Figure 3(c)). The heterogranular texture contains flakes of muscovite minerals (Figure 3(d)), and the main minerals include zoned plagioclase, k-feldspar, perthitic microcline, interstitial quartz, muscovite (derived from ancient biotite associated with chlorite), and accessory minerals such as sericite, epidote, and ilmenite.

Figure 3.(a) Exposed pillowed basalt showing the internal structure; (b) Volcanoclastic breccia; (c) Outcrop of monzodiorite; and (d) Coarse-grained muscovite granite.

The whole-rock geochemistry of the most represented formations discussed in the text is presented in Table 1. The KLB rocks expressed a bimodal character (Figure 4(a)). The volcanic rocks predominantly manifest as tholeiitic basalts (Mg# = 40.5 - 45.2), with rare magnesian basalts (Mg# = 64.2 - 72.8), and a subordinate calc-alkaline andesite (Mg# = 35.2) suite (Figure 4(b)). The tholeiitic volcanics exhibit a significant variation in multi-elements, marked by negative anomalies in P, Ti, and V, alongside positive anomalies in Ni and Cr. These variations are attributed to the fractionation of titanium minerals, spinel, pyroxene, and amphibole. The rare-earth profiles of these rocks exhibit broadly flat characteristics (Figure 4(d)), indicative of N-MORB island arc features (ΣTR = 17.6 - 54.55 ppm and (La/Yb)_N = 1.5). Within the KLB, the granitoids (Figure 4(c)) include tonalite and various types of potassic granite characterized by diverse compositions (metaluminous and peraluminous granite (A/CNK = 0.95 - 1.03) to hyperaluminous granite (A/CNK = 1.16 - 1.45)) derived from both igneous and sedimentary sources. These granitoids also exhibit anomalies in specific elements such as Nb, P, Ti, Sr, and Zr, influenced by the fractionation of specific minerals like leucoxene, apatite, and feldspar (Eu/Eu*: 0.92 - 1.1) (Figure 4(d)).

The tholeiitic volcanic rocks span 0.51218 to 0.51253 for ¹⁴³Nd/¹⁴⁴Nd and 0.16813 to 0.19481 for ¹⁴⁷Sm/¹⁴⁴Nd (Table 2). These closely clustered isotope ratio values potentially indicate a cogenetic source for these rocks. At 2.2 Ga, the εNd(t) values range from −1.20 to −0.22, indicating that all analysed samples were slightly less radiogenic than contemporaneous CHUR. This signature is consistent with derivation from, or interaction with, older continental crust and/or a previously differentiated mantle component. Two samples have εNd(t) values close to zero, suggesting compositions broadly similar to the CHUR reference at the time of magmatic crystallization. NK272 has the most negative value (εNd(t) = −1.20) and, together with its clastic texture, records the most evolved isotopic signature in the dataset. Negative εNd(t) values are uncommon in the Precambrian rocks of the eastern side of sWAC, with exceptions such as the Dabakala tonalite in Côte d’Ivoire, displaying an εNd(t) value of −2.9 [4], and the Winneba pluton in Ghana, with an εNd (2.173 Ga) value of −5.3 [32].

The TDM model ages are consistently older than the corresponding TCHUR ages (Table 2). They range from 3.22 to 3.86 Ga and represent approximate model ages for extraction of the crustal source from a depleted-mantle reservoir. Because these are model ages, they should be interpreted as estimates of source residence or extraction rather than direct crystallization ages. Nevertheless, all three TDM values predate the 2264 ± 0.1 Ma Sm-Nd isochron age, indicating that the source material was extracted from the mantle well before the magmatic event recorded by the volcanic rocks. The 2264 ± 0.1 Ma Sm-Nd isochron obtained from the selected tholeiitic volcanic rocks (Figure 5) therefore dates the isotopic closure or emplacement-related event of the analysed suite, whereas the older TDM ages indicate an inherited source history. Taken together with the slightly negative εNd(t) values, the data support derivation from, or assimilation of, an older and mildly evolved crustal component during emplacement of the tholeiitic volcanism.

The zircons extracted from granite samples exhibit zoning, taking on prismatic, bipyramidal, stubby, or acicular forms, and occasionally xenomorphic, with sizes ranging from 50 μm to 600 μm (Figure 6(a), Figure 6(c)). SEM images reveal pronounced resorption pockets in certain magmatic grains, mineral inclusions, and evidence of recrystallization (Figure 6(b), Figure 6(d)). The population of U-Pb laser data exhibiting isotopic homogeneity (>85% agreement) selected for this study comprises 35 measurements were integrated to calculate the crystallization age, as detailed in Table 3.

In sample NK222, laser-ablation analyses of zircon cores yielded mean ages of 2119 Ma based on the ²⁰⁷Pb/²³⁵U system and 2124 Ma based on the ²⁰⁶Pb/²³⁸U system. Analyses of the zircon rims yielded mean ages of 2114 Ma and 2104 Ma, respectively. In sample NK229, analyses of zircon cores produced mean ages of 2164 Ma (²⁰⁷Pb/²³⁵U) and 2169 Ma (²⁰⁶Pb/²³⁸U). Because concordant U-Pb ages provide the most reliable estimates of zircon crystallization, the concordant analyses were used for the geological interpretation [29].

The monzodiorite sample NK222 yielded a concordia age of 2120 ± 1.8 Ma (MSWD = 1.4; N = 25). The 5 - 20 Ma age difference between zircon cores and rims is interpreted to reflect a period of zircon overgrowth during the later stages of pluton development. The muscovite-granite sample NK229, collected from the central part of the KLB greenstone belt, yielded a concordia age of 2161 ± 2.4 Ma (Figure 7).

(a)

(b)

(c)

(d)

Figure 4.Classification diagrams using major and minor elements of the KLB suite. (a) AFM plot indicating the affinities of the studied rocks; (b) TAS diagram [33] discriminating volcanic rocks; (c) R1-R2 diagram for plutonic rocks [34]; and (d) Chondrite-normalized REE diagram for studied rocks. The volcanic rock patterns are highlighted in orange (normalized to Boynton).

Figure 5. Mean age of the Sm-Nd isotopes of three tholeiitic KLB volcanics in the diagram [35].

Figure 6. Zircon textures in the NK222 monzodiorite and NK229 muscovite granite plutons. (a) Bipyramidal zircon grain from NK222 with weak internal zoning and a xenocrystic core (M1C6). (b) Pyramidal NK222 zircon grain showing oscillatory zoning and a metamict core (M1R5C5). (c) Fractured, acicular to bipyramidal NK229 zircon grain with truncated zoning and a Zr-Th-enriched outer rim (M1C2). (d) Prismatic NK229 zircon grain exhibiting planar resorption surfaces (M1R1C3).

Figure 7. Wetherill diagrams showing uncorrected (common Pb) U-Pb isotope ratios of zircon and Concordia age from the tonalite and muscovite granite.

Table 1. Whole rock geochemical data of selected rocks on the KLB greenstone.

Table 2. Whole-rock Sm-Nd isotopic data calculated from the selected volcanic suite.

Table 3.U-Pb isotope data for zircons from the KLB granitoids.

The mafic rocks exhibiting tholeiitic affinity in the KLB of various units (basalt, pyroclastic, andesite, and associated gabbro) have nearly flat rare-earth spectra (ΣTR = 17.6 - 54.55 ppm and (La/Yb)N = 1.5), resembling N-MORB island arc signatures. Comparable birimian mafic suites show tholeiitic compositions together with LILE and LREE enrichment, HFSE depletion, and negative Nb-Ta anomalies, features widely interpreted as evidence for subduction-modified mantle and arc magmatism [18]. The ¹⁴³Nd/¹⁴⁴Nd isotope ratios of these rocks (0.51218 to 0.51250) indicate a common magmatic source, but the negative εNd (2.2 Ga) values (−1.2 to −0.22) show that the rocks were slightly less radiogenic than CHUR at the time of crystallization. This is consistent with the primary magma being extracted from a metasomatized mantle source during the Archean period, subsequently enriched through interaction with crustal materials, similar to modern subduction settings. This contradicts the hypothesis of a juvenile source for the sWAC magma [4][12][13][19][36][37].

The granitoids in the belt exhibit a diverse composition, including calc-alkaline and adakitic types, with both igneous (A/CNK = 0.95 - 1.03) and sedimentary (A/CNK = 1.16 - 1.45) protoliths showing meta-aluminous to peraluminous characteristics. Normalized rare earth ratios (La/Yb)_N = 2.2 - 52.3 combined with trace elements discrimination (Rb < 170 and Y + Nb < 50) are compatible with a volcanic-arc granite field. These data support generation from more than one parental reservoir rather than a single homogeneous magma source. Regional studies of birimian granitoids have addressed calc-alkaline I-type and TTG-like compositions, enrichment in Ba and Th, and depletion in Nb, Ta, Zr, Hf and Ti [9][10][12][13].

Combined geological and geochemical features are consistent with modern subduction-related plutonism and support a model of magma generation through partial melting of the metasomatized mantle wedge, fractional crystallization and magma mixing along with crustal assimilation during ascent, similar to contemporary oceanic arc systems [18].

The Sm-Nd model ages for the lower stratigraphy sequence (tholeiitic suites) yield an average age of 2264 ± 0.1 Ma (Ndi = 0.50967), overlain by calc-alkaline units comprising andesite, dacite, and rhyodacite with associated plutonic rocks. This bimodal volcanism pattern, with a tholeiitic lower sequence succeeded by calc-alkaline upper stratigraphy, is well documented across birimian greenstone belts of the sWAC [38] is recognized broadly as a signature of subduction initiation during early paleoproterozoic plate tectonics and arc growth. Down south in Côte d’Ivoire, calc-alkaline rhyodacite yields 2104 ± 2 Ma, contemporaneous to the plutonic rocks [19]. [9] previously dated a rhyolite tuff from the Goren belt in the Kaya region to 2238 ± 5 Ma by Pb/Pb in zircon, which was considered the oldest age in Burkina Faso. The KLB granitoids yield two distinct ages (2161 ± 2.4 Ma and 2120 ± 1 Ma) align to the first generation of Eburnean plutonism with geochemical similarities to the calc-alkaline rocks of recent subduction zones [1][9]. The paleoproterozoic crust of Kampti may have originated from a depleted ancient mantle that was enriched by contamination or reworking of a subducted pre-birimian crust. The second hypothesis is more plausible, as suggested by [32] and [18] on the Dabakala and Winneba terrains in Ghana. The geological, petrological, geochemical, and isotopic data support the model of arc magmatism generated in an oceanic subduction environment (Figure 8).

Figure 8. (a) Diagram (Th/Nb)_N vs (Tb/Nb)_N of characterization of geotectonic sites of volcanic rocks (modified from [39]). N-MORB—depleted type mid-oceanic ridge basalts (N-type MORB); E-MORB—enriched type mid-oceanic ridge basalt (E-type MORB); WPAB—intraplate alkaline basalts; CFB—continental tholeiite; BAB—tholeiite of back-arc basin; IAT—island arc tholeiite; CAB—calc-alkaline basalts of active margin. (b) Diagram Zr versus (Nb/Zr)_N [39][40] for the granitoids with distinction of the different geochemical groups. A—Mantle calc-alkaline magmatism (subduction zones); B—crustal calc-alkaline magmatism; C—intraplate alkaline magmatism; D—peraluminous leucogranites linked to collision zones.