Introduction

Molybdenite, a major source of molybdenum (Mo) and rhenium (Re), is largely recovered from porphyry Mo and porphyry Cu (± Mo–Au) deposits, which collectively constitute > 50% of the global supply of these metals (Carten et al. 1993; Keith et al. 1993; Candela and Piccoli 2005; Cooke et al. 2020). Porphyry-related Mo ± Cu deposits have been broadly classified into three types (Carten et al. 1993; Sun et al. 2015; Audétat and Li 2017). The Dabie-type refers to collision-related molybdenum deposits that are largely confined to China, the largest global supplier of Mo (Mao et al. 2014; Sun et al. 2015; Ouyang et al. 2020). The Endako-type are large-tonnage, low-grade Cu ± Mo deposits associated with arc-related settings where Mo is recovered as a byproduct of Cu ores (Audétat and Li 2017). The Climax-type are low-tonnage, high-grade Mo deposits associated with highly fractionated within-plate silicic rocks (Carten et al. 1993). Largely based on their contrasting geochemical features, Carten et al. (1993) further subdivided the Climax-type into an alkalic subtype (e.g., Henderson, CO, USA) and a transitional subtype (e.g., Questa, NM, USA). Skarn-related Mo-Cu-W-Sn and greisen-related W-Mo-Bi-Sn deposits can also yield Mo in economic quantities (Černý et al., 2005; Kirwin 2012; Jiang et al. 2019).

In recent years, numerous studies based on whole-rock geochemistry, fluid and melt inclusions, and zircon U-Pb and molybdenite Re-Os systematics have significantly advanced our understanding of the genesis of molybdenum deposits, including their tectonic settings and age relationships with associated intermediate and silicic rocks (Lowenstern and Sinclair 1996; Klemm et al. 2008; Pettke et al. 2010; Mao et al. 2014; Mercer et al. 2014; Sun et al. 2015; Audétat and Li 2017; Li et al. 2017; Gaynor et al. 2019;; Ouyang et al. 2020). However, issues regarding the source(s) of Mo and the associated ore components (sulfur), especially in within-plate settings, have not yet been widely studied (Stein and Hannah 1985; Pettke et al. 2010; Mercer et al. 2014; Liu et al. 2019). For instance, given that world-class deposits are confined to a few regions, such as the western USA and the Dabie region of China, Pettke et al. (2010) argued that anomalously large Mo concentrations in the upper mantle and/or lower crust are necessary for formation of giant Mo deposits. Other studies suggest that several factors, such as the volume and redox conditions of mineralizing magmas, as well as the efficiency of metal extraction during late-stage magmatic-hydrothermal processes also play a crucial role in the formation of large Mo deposits (Mercer et al. 2014; Audétat and Li 2017; Ouyang et al. 2020). Consequently, understanding the origin of Mo deposits is therefore linked to determining the sources of metal and to deciphering factors that promoted its concentration in economically significant deposits (Mercer et al. 2014).

In this contribution, we present a combined study of whole-rock geochemistry, Nd-Sr isotopes, Re-Os geochronology, and zircon U-Th-Hf systematics of the Mada Ring Complex, one of the fifty-three Sn, Nb ± (Zn, Cu, Pb, Mo)-rich anorogenic alkaline complexes in north-central Nigeria collectively known as the Nigerian Younger Granite Province (NYG; Bowden et al. 1987; Girei et al. 2019a). Our primary objectives were to (i) determine the sources of magma and Mo, (ii) constrain the timing of magma emplacement and Mo mineralization, and (iii) compare the Mada Mo-mineralized granites with the associated Sn + Nb-mineralized granites in the Nigerian Younger Granite Province.

Regional geology

The Nigerian Younger Granite Province consists of 53 Mesozoic silica-oversaturated anorogenic syenite-granite ring complexes (Bowden et al. 1987; Girei et al. 2019a). The emplacement of these anorogenic ring complexes marked a period of extension that followed a prolonged hiatus (> 350 Ma) after the Neoproterozoic (600 ± 150 Ma) Pan-African Orogeny (Magaji et al. 2011; Martin et al. 2012). These anorogenic syenite-granite ring complexes cover a total area of 75,000 km2 and form part of a larger alkaline province known as the Niger-Nigerian Alkaline Province (Fig. 1a–b; Kinnaird and Bowden 1987). This province, a classic example of within-plate alkaline magmatism, shows an age progression from Ordovician-Devonian (480–400 Ma) in Aїr (northern Niger) to Carboniferous (330–260 Ma) in Damagaram-Mounio (southern Niger) to Triassic-Early Cretaceous (213–141 Ma) in north-central Nigeria (Fig. 1b; Bowden et al. 1987). However, work by Moreau et al. (1994) demonstrated that the Aїr complex is structurally independent of the Damagaram-Mounio complex and the NYG. This necessitates the need for further geochronological studies to confirm the previously reported whole-rock Rr-Sr ages (Moreau et al. 1994; Ngako et al. 2006). In terms of petrological associations, peralkaline rocks and cogenetic anorthosite and leucogabbros predominate in Niger, whereas aluminous granites are dominant in Nigeria, where a few peralkaline granites occur in the northern alkaline ring complexes in association with syenite (Bowden et al. 1987; Demaiffe et al. 1991).

Fig. 1
figure 1

(a) Sketch map showing the distribution of Phanerozoic anorogenic alkaline rocks in Africa (redrawn from Bowden et al. 1987); the black rectangle shows the location of b. (b) Sketch map showing the distribution of Phanerozoic anorogenic ring complexes in Iforas (Mali), Aïr (northern Niger), Damagaram-Mounio (southern Niger), north-central Nigeria, and Cameroon in relation to Pan-African continental structures of Western Gondwana (Aïr, northern Niger). The abbreviation CVL in (b) stands for Cameroon Volcanic Line; black rectangle in (a) shows the location of a

Full size image

In Nigeria, 53 alkaline ring complexes have been identified, with diameters ranging from 2 to 25 km (Fig. 2a; Bowden 1985), each representing the root of an eroded volcano (Kinnaird et al. 1985). The complexes occur as high-level discordant intrusions in sets with overlapping centers (e.g., the Sara Fier, Ningi Bura, and Kila Wargi complexes) or as individual centers (e.g., the Ririwai, Mada, and Kudaru complexes; Fig. 2a). They are circular to elliptical in shape and intrude Precambrian basement consisting of high-grade Eburnean (Paleoproterozoic) gneisses and migmatites with relics of Archean rocks, as well as Neoproterozoic (Pan-African) granites (Bowden et al. 1990; Bute et al. 2019; Girei et al. 2019b). The Neoproterozoic post-collisional granites in Nigeria are associated with Sn-Ta ± Nb-mineralized pegmatites that are largely confined to a NE-trending zone known as the Older Tin Belt (Girei et al. 2019a).

Fig. 2
figure 2

(b) Simplified geological map showing the distribution of anorogenic ring complexes in north-central Nigeria (modified from Kinnaird et al. 1985). Note: the black rectangle in (a) shows the location of b. (b) Geological map of the Mada anorogenic ring complex (modified from Abaa 1985)

Full size image

Emplacement of the alkaline ring complexes started in the Triassic (~ 213 Ma, e.g., Dutse Complex) and continued episodically until the Early Cretaceous (~ 141 Ma, e.g., Afu Complex), with a general progression from north to south (Fig. 2a; Rahaman et al. 1984; Girei et al. 2019a; Kamaunji et al. 2020). The distribution of alkaline ring centers along N-S shear zones and associated deep-seated transcurrent faults indicate that zones of lithospheric weakness exerted significant control on their emplacement (Rahaman et al. 1984).

Primary Sn + Nb (±Zn, Cu, Pb) mineralization occurs in some of the Nigerian ring complexes, including the Ririwai, Tibchi, Afu, Ropp, and Ningi-Bura complexes, as well as in the Rayfield Gona pluton of the Jos-Bukuru Complex (Fig. 2a; Pastor and Turaki 1985; Kinnaird et al. 2016). Weathering and erosion of these epizonal alkaline complexes gave rise to economically significant alluvial Sn-Nb placer deposits that are mostly in proximity to the ring complexes (Girei et al. 2019a). Until this study, Mo mineralization has not been reported in the Mada Complex, but limited Mo mineralization associated with peralkaline granite has been reported from the Kigom Complex (Imeokparia 1985). Both Sn ± (Nb)- and Mo-mineralized granites are spatially associated in the alkaline anorogenic province (Fig. 2a).

Geology of Mada Complex molybdenum mineralization

The Mada Complex is oval in shape, with axes measuring 26.5 km and 15 km, and occupies a total area of 271 km2. It is composed of minor remnants of an extrusive phase, a large core of biotite granite, small outer dykes of granite porphyry, and minor mafic rocks that occur as dykes and cone sheets. The extrusive phase consists largely of comenditic rhyolite with minor breccia, ashfall, and agglomerate (Fig. 2b; Abaa 1985).

Biotite granite is the dominant intrusive phase in the Mada Complex. Largely based on texture, three facies of biotite granite have been distinguished: coarse-grained, medium-grained, and fined-grained. Molybdenum mineralization is confined to the medium-grained biotite granite (hereafter “mineralized granite”; Fig. 2b). The mineralized granite experienced extensive hydrothermal alteration that is spatially associated with Mo mineralization (Fig. 3a–f). In thin section, the main minerals in mineralized granite and coarse-grained granite are quartz (30–32%), alkali feldspar (45–50%), plagioclase (10–13%), and biotite (4–6%), with zircon, apatite, and titanite as accessory minerals (Fig. 4a–i). The feldspar and quartz crystallized first, followed by biotite. Orthoclase constitutes the bulk of the alkali feldspar, and in mineralized granite, it is locally altered to sericite. Quartz grains are typically anhedral and relatively free of inclusions (Fig. 4b). Biotite occurs mainly as an interstitial mineral and ranges in color from light brown to green.

Fig. 3
figure 3

a Coarse-grained biotite granite with barren quartz vein. b–c Alteration zones of Mada Mo mineralization (Stage I). d Contact between the Stage I and Stage II alteration zones. e, f Alteration zones of Mada Mo mineralization (Stage II)

Full size image
Fig. 4
figure 4

a–b Coarse-grained biotite granite. c–d Medium-grained/mineralized biotite granite (samples from Stage I alteration zone). e–f Medium-grained/mineralized biotite granite (samples from Stage II alteration zone). Qz = quartz, Bt = biotite, Kf = K-feldspar, Ch = chlorite, Ser = sericite, Py = pyrite; Mo = molybdenite, W = wolframite

Full size image

Molybdenum mineralization occurs in one of the southernmost anorogenic rings of the Mada Complex (Fig. 2a). Molybdenite constitutes the dominant ore mineral in the ring complex and occurs mainly disseminated in the cupolas of the mineralized granite, where it is associated with traces of wolframite (Fig. 3b–f). Two Mo ore-forming stages, each characterized by a distinctive mineral assemblage, have been distinguished (Fig. 3b–d). During the first stage (Stage I, Fig. 3b), Mo mineralization was accompanied by hydrothermal alteration involving K-rich fluids. The alteration mineral assemblage includes K-feldspar, quartz, and minor sericite (Fig. 4d). This was followed by emplacement of disseminated and minor stockwork Mo mineralization during the second stage (Stage II, Fig. 3c, d). The alteration mineral assemblages that characterize the latter stage include sericite, chlorite, quartz, greenish mica, and minor pyrite (Fig. 4e–i). This mineral assemblage is similar to those observed in greisen-related deposits (Černý et al., 2005), and we therefore consider Stage II to represent greisen-related mineralization.

Sampling and analytical techniques

Sample collection

A total of six samples were collected: WR-02, WR-03, WR-05, and WR-40 are Mo-bearing samples of the mineralized granite collected from mine shafts, and WR-10 and WR-70 are samples collected from outcrops. All samples were analyzed for whole-rock geochemical and Nd-Sr isotopic composition. In addition, four samples (WR-02, WR-03, WR-05, and WR-40) were analyzed for Re-Os geochronology, and zircons were extracted from two samples (WR-02 and WR-10) and analyzed for U-Pb geochronology and Lu-Hf isotope compositions.

Whole-rock major- and trace-element analyses

Samples for whole-rock major- and trace-element analyses were crushed using a corundum jaw crusher and pulverized to 200 mesh size using an agate ball mill prior to analysis. Major-element analysis was performed using an energy-dispersive X-ray fluorescence (XRF) spectrometer (Shimadzu EDX 700), and trace-element analysis was performed using an Agilent 7700e inductively coupled plasma-mass spectrometer (ICPMS) with an analytical precision of ~ 5–10%. These analyses were carried out at the Australian Laboratory Services in Guangzhou, China. For XRF analysis, a pressed-powder disc was used following the procedure described in Girei et al. (2019a). Prior to the ICP-MS analysis, samples were dissolved sequentially with lithium borate fusion (by Li2B4O7 flux) followed by aqua regia (HNO3 + 3HCl) digestion. International standard (JG-2) was used as a reference material for both major and trace elements.

Whole-rock Nd-Sr isotopes

Whole-rock Rb-Sr and Sm-Nd isotope analysis was performed on a Micromass Isoprobe Multi-Collector Inductively Coupled Plasma Mass Spectrometer (MC-ICP-MS) at the State Key Laboratory of Geological Processes and Mineral Resources (GPMR). Prior to the analysis, Sr and Nd were separated and purified using conventional ion exchange procedures. The procedure involved digesting ~ 50–100 mg of powdered sample into a distilled HF-HNO3 solution in screw-top PFA beakers at a minimum temperature of 120 °C for 2 weeks. Rubidium and Sr were separated and purified using conventional column chemistry with cation exchange resins. Neodymium was separated from the other REEs using a di-(2-ethylhexyl) phosphoric acid (HDEHP) column and purified with dilute HCl. During the analytical session, samples were analyzed alternately with the NBS-987 (87Sr/86Sr = 0.710275 ± 0.000012; 2σ and n = 11) and JNdi-1 standards (143Nd/144Nd = 0.512109 ± 12; 2σ) for Sr and Nd, respectively. The measured 87Sr/86Sr and 143Nd/144Nd values were corrected to 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219, respectively (cf. Li et al. 2019). A single-stage neodymium model age was calculated assuming linear neodymium isotopic growth from a depleted mantle reservoir with ɛNd = 0 at t = 4.56 Ga to ɛNd = +10 at present, and a two-stage model age was calculated assuming that the protoliths of the granitic magmas had a ƒSm/Nd value or Sm/Nd ratio equal to that of average continental crust (e.g., Jahn et al. 2001; Girei et al. 2019a).

Zircon U-Pb dating and Lu-Hf isotope analysis

Zircons for U-Pb dating and Lu-Hf isotope analysis were separated using conventional heavy liquid and magnetic separation techniques and then handpicked under a binocular microscope. The zircons were then mounted in epoxy resin and polished prior to the LA-ICP-MS U-Pb dating and Lu-Hf isotope analysis. Cathodoluminescence (CL) images of the zircon grains were captured using a scanning electron microscope (SEM) fitted with an energy-dispersive system (EDS). U-Pb dating was performed on an Analytik Jena PQMS Elite ICP-MS coupled with an ESI NWR 193-nm laser ablation system at a 50-Hz frequency on a 32-μm ablation spot and a background acquisition time of 15 s (gas blank). The analysis was conducted at GPMR, China University of Geosciences (Wuhan), using zircon 91500 and glass NIST610 as external standards. Detailed analytical conditions including raw data selection, background integration, and time-drift correction are described by Liu et al. (2010). The analytical results for all standards used in zircon U/Pb dating are given in Electronic Supplementary Material (ESM) 1. Overall, the analyses yielded mean 206Pb/238U ages of 1063 ± 20 Ma (95% conf., MSWD = 1.4) for 91500, 600.3 ± 3.7 Ma (95% conf., MSWD = 0.3) for GJ-1, and 335.6 ± 8.4 Ma (95% conf., MSWD = 2.4) for PLE, which are all consistent with recommended values within error (Jackson et al. 2004; Slama et al. 2008).

Zircon Lu-Hf isotope analysis was carried out on ablation spots close to those previously analyzed for U-Pb ages. The analysis was performed at the GPMR on a Neptune Plus MC-ICP-MS coupled with a Geolas 2005 Excimer ArF laser ablation system with energy density of 5.3 J/cm2. The procedure involves background signal acquisition for 20–50 s on laser mode with a beam diameter of 44-μm using zircon 91500 as a standard. Detailed procedures for correction of isobaric interferences of 175Lu on 176Hf and 176Yb on 176Hf are given in Machado and Simonetti (2001). Epsilon hafnium (ɛHf(t)) calculations were performed using the recommended present-day chondritic ratios of 176Hf/177Hf (0.282772) and 176Lu/177Hf (0.0332; Belousova et al. 2006).

Re-Os dating

Four molybdenum-bearing samples were selected for Re-Os isotopic dating: one sample (WR2) representing Mo mineralization associated with potassic alteration and three samples (WR3, WR4, and WR40) representing greisen-related Mo mineralization. Molybdenite grains were separated from the samples using gravity and magnetic techniques and then handpicked under a binocular microscope. The molybdenite Re–Os dating was performed at the State Key Laboratory of Isotope Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Prior to analysis, chemical separation of Re-Os was performed using the Carius tube digestion technique. The procedure involved digesting 0.05 to 0.10 g of powdered sample and equilibrating it with 185Re- and 190Os-enriched spikes in reverse aqua regia (6 ml concentrated HNO3 + 2 ml concentrated HCl) in sealed Carius tubes for 24 h at 240 °C. The Os fraction was extracted using a CCl4 solvent and concentrated using hydrogen bromide (HBr) followed by microdistillation. In contrast, the Re fraction was separated and purified using anion column chromatography. The extracted osmium was loaded onto Pt filaments and measured as OsO3− ions by negative-thermal ionization mass spectrometry (N-TIMS) using the electron multiplier mode on a Thermo-Finnigan Triton. Repeated analyses of the Os standard solution yielded a mean 187Os/188Os value of 0.12048 ± 0.00028 (2 SD, n = 5), which is in good agreement with the value of 0.12022 ± 0.00020 (2 SD, n = 14) obtained on the same instrument in Faraday cup mode (Li et al. 2010). A Thermo Elemental X2 Series inductively coupled plasma mass spectrometer (ICP-MS) fitted with a conventional low-volume quartz impact bead spray chamber and a 0.4-ml min−1 borosilicate nebulizer was used to analyze the Re isotope ratio. In order to obtain high instrumental sensitivity and low oxide production levels, ion lens’ settings and nebulizer gas flow rate and torch position were optimized daily using a 10-ng ml−1 tuning In–Ce mixture standard solution. The free aspiration of the nebulizer provided better signal stability. All data were blank corrected, and the total procedural blanks were 0.80 ± 0.40 pg (1σ, n = 2) with an 187Os/188Os ratio of 0.68 ± 0.54 (1σ, n = 2) on average for Os and 4.8 ± 0.2 pg (1σ, n = 2) for Re.

Results

Whole-rock major- and trace-element compositions

The whole-rock major- and trace-element compositions of the study samples are given in EMS 2 Table S1. The rocks have high SiO2 (74.8 to 76.8 wt%) and K2O (4.7 to 5.9 wt%), but low MgO (0.02 to 0.08 wt%), CaO (0.5 to 0.8 wt%), and TiO2 (0.04 to 0.1 wt%), and moderate Na2O + K2O (7.7 to 8.1 wt%) and Al2O3 (11.7 to 12.4 wt%) concentrations. Compared to the coarse-grained biotite granite, the mineralized medium-grained granites have higher Na2O but lower Al2O3 contents (EMS 2 Table S1). All granites have a metaluminous to weakly peraluminous composition (molar Al2O3/(Na2O + K2O > 1), similar to biotite granite associated with Sn ± Nb mineralization in the Ririwai and Tibchi complexes (Fig. 5a–d). However, granite associated with Mo mineralization in the Kigom Complex and rhyolite from the Mada Complex are peralkaline (Fig. 5c). All samples plot within the fields of (i) granite according to the classification of De la Roche et al. 1980; (Fig. 5a), (ii) high-K to shoshonitic granite in a K2O vs. SiO2 discrimination diagram, (iii) alkali calcic to calc-alkalic granite according to the classification of Frost et al. 2001; (Fig. 5b, d), and (iv) within-plate setting on the discrimination plot of Pearce et al. (1984), a characteristic feature of within-plate A-type granite (Bonin 2007; Fig. 5e–f). However, A-type granites associated with Sn ± Nb mineralization in the Ririwai and Tibchi complexes generally have higher SiO2 and lower K2O than those associated with Mo in the Mada and Kigom complexes (Fig. 5d).

Fig. 5
figure 5

Granite classification diagrams. a SiO2 versus Na2O + K2O plot of Mada granite (after De la Roche et al. 1980). b SiO2 versus (Na2O + K2O − CaO) plot of Mada granite, after Frost and Frost et al. (2001). c A/CNK versus A/NK plot of Mada granite, (after Shand 1943). d SiO2 versus K2O plot of Mada granite, (after Peccerillo and Taylor, 1976). e 10,000 × Ga/Al versus Ce plot of Mada granite, after Whalen et al. (1987). f Yb versus Nb plot of Mada granite, after Pearce et al. (1984). Data of granites plotted for comparisons are from Abaa (1985), Imeokparia (1985), Batchelor and Bowden (1986), and Girei et al. (2019a, b)

Full size image

All samples are characterized by slight enrichment of LREEs relative to HREEs along with a negative Eu anomaly (Fig. 6a–b). On a primitive mantle-normalized plot (Fig. 6a), the samples show trace-element patterns characterized by large enrichments of Zr, Hf, and Rb and depletions of Sr, Ba, and P, consistent with a magma evolution marked by significant fractionation of feldspars and apatite. Samples of the coarse-grained granite generally have higher Nb, Ta, La, Zr, Rb, and Ba than the mineralized granite (EMS 2 Table S1). Calculated zircon saturation thermometry based on the model of Watson and Harrison (1983) yielded temperature estimates ranging from 838.0 to 860.0 °C (EMS 2 Table S1).

Fig. 6
figure 6

a Chondrite-normalized REE plot of Mada granite, with normalization values from Taylor and McLennan (1985). b Primitive-mantle-normalized trace element spider diagrams of Mada granite, with normalization values from Sun and McDonough (1989). Note: green and black lines in (a) represent average composition of bulk continental crust (BCC) and lower continental crust (LCC) as reported by Taylor and McLennan (1985). Black and green line in (b) represent REE pattern of Augite syenite from Kila Warji Complex (Bowden 1985)

Full size image

Whole-rock Sr-Nd isotopes

Whole-rock Sr-Nd isotope compositions are given in Table 1. Initial (87Sr/86Sr) and εNd(t) were calculated using the zircon U-Pb ages of the granites. The samples yielded high 87Rb/86Sr values ranging from 62.0 to 78.7, but low (87Sr/86Sr)i values ranging from 0.6949 to 0.7044, with one value (0.6949) falling below the basaltic achondrite best initial (BABI = 0.69897 ± 3; Villaseca et al. 1998). Such anomalously low initial 87Sr/86Sr values indicate the Rb-Sr system was probably disturbed (Ahmed et al. 2018). The 143Nd/144Nd values range from 0.5122 to 0.5123 and calculated εNd(t) from − 5.87 to − 5.47, corresponding to Mesoproterozoic two-stage models ages ranging from 1243 to 1587 Ma. Calculated εNd(t) values of the Mada molybdenum-related biotite granite are slightly higher than the εNd(t) values (− 3.55 to − 3.26) of the Sn-rich biotite granite from the Ririwai Ring Complex in north-central Nigeria (Girei et al. 2019a).

Table 1 Whole-rock Nd-Sr Isotope compositions of the Mada granites
Full size table

Zircon U-Pb geochronology

Analytical results of zircon U-Pb dating are given in EMS 2 Table S2. The zircon grains are generally large (100–200-μm diam.), euhedral to prismatic in shape, and pale yellow to gray in color. They exhibit concentric oscillatory zoning and display (100) and (101) crystallographic form (ESM 3 Fig. 1). Zircon grains from the coarse-grained biotite granite (n = 14) exhibit Th/U ratios ranging from 0.88 to 1.40, which are typical of magmatic zircon (Belousova et al. 2002; Kirkland et al. 2015). Eight zircon grains yielded 206Pb/238U ages ranging from 152 to 155 Ma, with a weighted mean age of 152.6 ± 1.1 Ma (MSWD = 0.34) (Fig. 7a–b), which is similar to a whole-rock Rb-Sr age of 147 ± 5 Ma (initial strontium = 0.708 ± 0.02) from Rahaman et al. (1984). In addition, four zircon grains yielded younger 206Pb/238U ages ranging from 148 to 149 Ma, suggesting a degree of Pb loss.

Fig. 7
figure 7

a Zircon U-Pb age concordia plot for the coarse-grained biotite granite. b Mean weighted average age plot for coarse-grained biotite granite. c Zircon U-Pb age concordia plot for mineralized granite. d Mean weighted average age plot for the mineralized granite

Full size image

Thirty spots were dated on the zircon grains extracted from the mineralized granite, but eight spots were rejected due to high common Pb and/or low concordance (< 90%). The remaining 22 zircon grains yielded magmatic Th/U ratios ranging from 0.82 to 1.55 and formed two age clusters. The older one consists of eight 206Pb/238U ages, ranging from 159.0 to 163.0 Ma with a weighted mean age of 160.0 ± 1.1 Ma (MSWD = 1.03), that are interpreted as antecrystic zircon grains that crystallized from an earlier pulse of magma (Miller et al. 2007). The younger population consists of fourteen 206Pb/238U ages, scattered between 149 and 153 Ma with a weighted mean age of 151 ± 0.74 Ma (MSWD = 0.52; Fig. 7c–d), that are interpreted to represent the crystallization age of the rock. Thus, the younger zircons are “autocrysts” that crystallized from their host melt near the time of emplacement (Miller et al. 2007).

Zircon trace-element compositions

The trace-element compositions of zircon from the Mada granite are given in EMS 2 (Table S3). The zircons yield moderate, but variable concentrations of U (143–2133 ppm) and Th (43–859 ppm), with fairly high Th/U ratios (0.88–1.55) that are typical of magmatic zircon (Kirkland et al. 2015). They are also characterized by high concentrations of Hf (8029–17,910 ppb), but low Nb/Ta ratios (0.57–4.18), reflecting growth from a highly fractionated melt. In addition, the zircons exhibit wide variations in elements, such as Y (1102–5491 ppm), Pb (8.7–134 ppm), Ti (3.5–15.1 ppm), and ∑REE (1010–3932 ppm). On a chondrite-normalized REE diagram, they exhibit steeply rising patterns characterized by HREE enrichment and LREE depletion, with a positive Ce anomaly and a negative Eu anomaly (ESM 3 Fig. 2a–c). However, a few zircons, especially those in the mineralized granite, have high LREE contents and low SmN/LaN ratios, which can be attributed to hydrothermal alteration and/or the presence of LREE-rich accessory minerals, such as apatite (ESM 3 Fig. 2d–e).

Zircon Hf isotopes

Analytical results of in situ zircon Hf isotopes are given in EMS 2 (Table S4). Five zircons of the coarse-grained biotite granite yielded 176Hf/177Hf ranging from 0.28243 to 0.2845 and 176Yb/177Hf from 0.0291 to 0.0576. Calculated ɛHf(t) values range from − 8.25 to − 9.85, corresponding to late Paleoproterozoic crustal model (TCDM) ages of 1723–1764 Ma (EMS 2, Table S4). Twelve spots were analyzed on ten zircons (including two antecrystic grains) of the mineralized granite; one of the antecrystic grains (Wr2–14) yielded the highest values of 176Hf/177Hf (0.2426), 176Yb/177Hf (0.0985), and calculated ɛHf(t) (− 4.97), whereas the other antecrystic grain (Wr2–12) yielded values of 176Hf/177Hf (0.2824), 176Yb/177Hf (0.0521) and ɛHf(t) (− 8.77) that overlap the range obtained from the autocrystic grains (176Hf/177Hf = 0.2424–0.2425, 176Yb/177Hf = 0.0339–0.0765, and ɛHf(t) = − 5.12 to − 10.06; EMS 2, Table S4). Calculated crustal model ages (TCDM) range from 1839 to 1526 Ma, with the lowest value (1526 Ma) obtained from the antecrystic grain that yielded the highest ɛHf(t) (EMS 2, Table S4). Overall, calculated crustal model ages are slightly lower than, but overlap the range of, two-stage model ages (1243 to 1587 Ma) obtained from whole-rock Nd isotopes (Table 1).

Molybdenite Re-Os geochronology

The analytical results of molybdenite Re–Os dating for four samples are given in Table 2. Because the common Os contents in the molybdenite are negligible, model ages were calculated using the equation {In (187Os/187Re + 1)/λ} (Suzuki et al. 2001), where λ is the decay constant of 87Re, which is equal to 1.666 (± 0.017) × 10−11 year−1 (Smoliar et al. 1996). The molybdenite-bearing sample associated with potassic alteration (WR2) yielded Re (238 ppb), 187O (260.6 pbb), and a calculated model age of 150.9 ± 1 Ma that, within error, agrees well with the zircon weighted mean 206Pb/238U age (151.3 ± 0.74 Ma) obtained from the same sample. This age (150.9 ± 1 Ma) is therefore interpreted as the approximate age of the Mo ore-forming event in the Mada Ring Complex. In contrast, the three molybdenite-bearing samples from greisen-related Mo mineralization (WR3, WR4, and WR40) exhibit wide variations of Re (95.9 to 269.0 ppb), 187O (123.6 to 217.6), and calculated model ages (144 to 136.9 Ma; mean = 142 ± 11 Ma). For a given aliquot size (i.e., 0.1 g), the sample with the lowest Re content (95.9 ppb) yielded the youngest model age (136.9 Ma), and the mean model age obtained from these samples (142 ± 11 Ma) is significantly younger than their Re-Os isochron age (155.3 ± 7.2 Ma; MSWD = 0.23; initial 187O/188O = − 6.6 ± 9.8; Fig. 8a, b). These ages cannot therefore be interpreted to represent ore formation age, but indicate that the Re-Os system was probably disturbed (Suzuki et al. 2001).

Table 2 Molybdenite Re-Os geochronological data of the Mada mineralized granite
Full size table
Fig. 8
figure 8

a 187O/188O versus 187Re/188O isochrone for the Mada molybdenite sample. b Plot showing distribution of calculated model ages of three molybdenite samples from greisen-related Mo mineralization

Full size image

Discussion

The geochemical data permit evaluation of magma sources and intrusive history related to molybdenum mineralization in the Mada Ring Complex of north-central Nigeria. Evaluation of these data and comparison with studies of Sn ± Nb mineralization in the NYG Province enables us to offer new insights into the petrogenetic factors that favor formation of co-existing Mo and Sn ± Nb Mineralization.

Timing of magma emplacement and Mo mineralization

Our combined LA-ICP-MS zircon U-Pb and molybdenite Re-Os geochronological data provide insights into the timing of magma emplacement in relation to the Mo mineralization event at Mada. A previous study of the Mada Complex yielded a whole-rock Rb-Sr age of 147.0 ± 5 Ma, interpreted to represent its emplacement age (Rahaman et al., 1985). However, as previously stated, the low initial 87Sr/86Sr values of Mada granite relative to their negative εNd(t) values (Table 1) suggest that their Rb-Sr system was probably disturbed (Villaseca et al. 1998). Thus, the whole-rock Rb-Sr ages of the granite may be lower than its actual emplacement age (Walraven et al. 1990). Nevertheless, the age is within error of the weighted mean zircon 206Pb/238U age of 152.6 ± 1.1 Ma obtained for the coarse-grained biotite granite (Fig. 7a–b). On the other hand, zircons from the mineralized granite yielded 206Pb/238U ages ranging from 163 to 150 Ma, with two distinct clusters: (1) an older mean age of 160 ± 1 Ma (n = 8) and (2) a younger mean age of 151 ± 0.7 Ma (n = 14; Fig. 7c–d). The ~ 160 ± 1 Ma age closely matches the 206Pb/238U age of the biotite granite from the nearby Ropp complex (161 ± 0.8 Ma; Vandi Dlama Kamaunji, pers. comm., 2020), implying that older zircons may be antecrystic (Miller et al. 2007; Gaynor et al. 2019). Thus, assuming Pb-loss can be discounted, the younger age (151 ± 0.7 Ma) represents the approximate crystallization age of the mineralized granite.

LA-ICP-MS molybdenite Re-Os dating yielded a wide range of model ages (136.9 to 150.9 Ma; Table 2). A single molybdenite-bearing sample associated with potassic alteration yielded a model age of 150.9 ± 1 Ma, coeval with the U/Pb age of the mineralized granite (151 ± 0.7 Ma), and is interpreted as the approximate age of commencement of Mo mineralization in the study area. In contrast, three molybdenite-bearing samples from greisen-related Mo mineralization yielded dissimilar model ages ranging from 144.0 to 136.9 Ma with a mean of 142.0 ± 11.0 Ma, which is younger than their isochron age (155.3 ± 7.2 Ma). When the Re-Os system remains closed, the mean model age of molybdenites will agree reasonably well with their Re-Os isochron age, and can be interpreted as the time of ore formation (Selby et al. 2003; Stein 2013; Zhai et al. 2019). However, given the stark disagreement between the mean model age and the isochron age, neither can be interpreted with assurance to represent the age of Stage II mineralization in the Mada Complex. Instead, such ages suggest that the Re-Os system was probably disturbed due to open-system processes (e.g., Re loss; Stein et al. 1998; Suzuki et al. 2001) or decoupling related to heterogeneous distribution of Re in molybdenite caused by variation in the physicochemical properties of the mineralizing fluids (Barra et al. 2017). Thus, the magma emplacement and Mo mineralization event in the Mada Ring Complex probably took place between 152 ± 1 and 150 ± 1 Ma.

Source of magma and molybdenum

A-type granites associated with Mo mineralization in the Mada Ring Complex are highly silicic with low MgO, CaO, TiO2, Sr, and Ba consistent with highly differentiated granite (Yakymchuk 2019). The high degree of fractionation means that their major- and trace-element compositions cannot provide reliable information regarding the possible source(s) of their parental magma(s) (Christiansen and Keith 1996; Girei et al. 2019a).

Paired Sm-Nd and Rb-Sr isotope analyses are widely regarded as a robust tool for constraining the source of a granite magma (Dickin et al. 1991; Li et al. 2018, 2018). However, in highly differentiated granites associated with mineralization, the Rb-Sr system may be disturbed as a result of open-system processes such as loss of radiogenic strontium (87Sr), which usually gives rise to low initial Sr ratios (87Sr/86Sr)I (Walraven et al. 1990). The Mada A-type granites yield low (87Sr/86Sr)i (0.6949 to 0.70440), with one value (0.6949) falling below BABI (0.69897 ± 3), suggesting that their Rb-Sr isotopic systems have been disturbed (Fig. 9a; Walraven et al. 1990; Li et al. 2019). As a result, the initial strontium ratio of such granites cannot be used to constrain their source (Ahmed et al. 2018). In contrast, the Sm-Nd system is more resistant to fluid-induced isotopic disturbance and can provide more reliable constraints on the parental source(s) of the mineralized granites (Jiang 2005; Li et al. 2018, 2018). Thus, assuming that equilibrium conditions existed between the melt and residual crystals during the partial melting event that led to generation of the Mada A-type magmas, the Nd isotope composition of the granites still reflects their sources the isochron age (Poitrasson et al. 1995; Christiansen and Keith 1996). The Mada granites yield εNd(t) values (− 5.87 to − 5.47) lower than those of Sn ± Nb-mineralized granite in the Ririwai complex (− 3.55 to − 3.26) but higher than those of the Pan-African basement in the region (− 8.4 to − 20.9; Fig. 9b; Girei et al. 2019a, b), suggesting magma derivation involving large amounts of crustal material with time-integrated Nd enrichment relative to the chondrite uniform reservoir (CHUR), along with a moderate contribution from the upper mantle (Fig. 9b; e.g., Dickin et al. 1991; Kerr and Fryer 1993). This is consistent with the zircon Lu-Hf results, which exhibit moderate negative yet variable εHf(t) values (Fig. 9c–d).

Fig. 9
figure 9

a Whole-rock (Sr87/Sr86)i vesus Rb87/Sr86. b Plot of whole-rock ɛNd(t) versus age for Mada granite. c Plot of zircon ɛHf(t) versus age for Mada granite. d Zircon ɛHf(t) versus whole-rock ɛNd(t) isotopic plots for the Mada granites. Whole-rock ɛNd(t) data of Ririwaai Sn ± Nb-mineralized granite are from Girei et al. (2019a, b)

Full size image

A growing body of literature has shown that the rhenium (Re) content of molybdenite can provide insight into possible sources of the ore metals (Mao et al. 1999; Stein et al. 2001; Berzina et al. 2005; Mao et al. 2006; Stein 2006; Wanhainen et al. 2014). In this regard, high Re contents (> 100 ppm) in molybdenite suggest a mantle source for Mo, whereas low Re contents (< 10 ppm) are consistent with a crustal source (Mao et al. 1999; Berzina et al. 2005). Although molybdenite at Mada generally has low Re contents (< 1 ppm; Table 2), favoring a crustal source for the metal, it should be noted that the Re content of molybdenite is also a function of the redox conditions of the melt from which it crystallizes, as highly oxidizing melts are known to partition large amounts of Re into molybdenite (Berzina et al. 2013; Barton 2019). Thus, molybdenite associated with porphyry Cu ± Mo deposits, which usually forms from highly oxidizing magmas, generally has higher Re contents relative to those that form from less oxidizing magmas (Barton 2019). In this regard, until further evidence becomes available, the possibility of mantle contribution to the ore materials (Mo and S) of the Mada Mo mineralization cannot be completely ruled out.

Zircon constraints on temperature and redox conditions

Various elements substitute readily into the zircon structure, providing information about the physicochemical conditions of its growth (Ferry and Watson 2007; Claiborne et al. 2010; Trail et al. 2012; Smythe and Brenan 2016). In particular, Ti-in-zircon thermometry can provide reliable constraints on magma temperature at the time of zircon crystallization (Ferry and Watson 2007; Siégel et al. 2018; Schiller and Finger 2019). However, given that Ti is preferentially incorporated into rutile and can substitute for Si in the zircon structure, Ti-in-zircon temperatures are partly dependent on the activity of silica (aSiO2) and titanium oxides (aTiO2) in the melt (Ferry and Watson 2007). Given that the Mada granites are highly silicic and contain no rutile, we adopted activity values of 1.0 and 0.7 for silica and titanium oxides, respectively (e.g., Li et al. 2014). For the Mada granites, calculated temperatures using the modified thermometric equation of Ferry and Watson (2007) range from 705 to 794 °C, which are slightly lower than temperature estimates based on the Zr-saturation thermometry model (838 to 860 °C; Watson and Harrison, 1983; EMS 2, Table S5). This difference may reflect the presence of antecrystic zircons in the Mada granite, leading to overestimation of zircon saturation temperatures computed using the model of Watson and Harrison (1983), as noted for A-type granites from the Ririwai Complex in north-central Nigeria (Girei et al. 2019a).

Numerous studies have demonstrated that magmatic redox conditions significantly affect partitioning of REEs, especially Eu and Ce, into zircon structure (Ballard et al. 2002; Trail et al. 2012; Burnham et al. 2015; Smythe and Brenan 2016). Unlike other REEs which are exclusively trivalent, Eu and Ce can undergo changes in valence, i.e., between (II) and (III) for Eu and between (III) and (IV) for Ce, depending on the redox conditions of the melt (Trail et al. 2012). Under oxidizing conditions, the tetravalent form of Ce (Ce4+) has a similar radius and the same charge as Zr4+ and hence is more compatible in the zircon structure than Ce3+. Under the same (oxidizing) conditions, Eu occurs as Eu3+, which can be readily accommodated in the zircon structure giving rise to a small (less negative) Eu anomaly (Ballard et al. 2002; Wu et al. 2018). High Ce4+/Ce3+ and low Eu/Eu* in zircon therefore indicate oxidizing conditions at the time of crystallization (Ballard et al. 2002). However, Eu/Eu* in zircon is sensitive to melt Eu/Eu∗, which is chiefly controlled by plagioclase fractionation and/or accumulation (Burnham et al. 2015). Thus, given the presence of plagioclase in the Mada granite, Eu/Eu* cannot provide a reliable estimate of oxygen fugacity (Smythe and Brenan 2016). Zircons from the Mada granite yield moderate, yet variable, Ce4+/Ce3+ values with higher Ce4+/Ce3+ in the mineralized granite than in the coarse-grained biotite granite, indicating intermediate magmatic redox conditions at the time of zircon crystallization (Fig. 10a–b; EMS 2, Table S5). Moreover, calculated oxygen fugacity using the lattice strain model (Smythe and Brenan 2016) all plot above the field of the fayalite-magnetite-quartz (FMQ) buffer on a –logƒO2 vs T/°C diagram, with higher average oxygen fugacity values (log FMQ = +1.16; Fig. 10a) in the mineralized granite than in the coarse-grained biotite granite (log FMQ = +0.45). Thus, it can be inferred that the A-type granite magmas associated with Mo mineralization in the Mada Complex were moderately oxidizing but less so than typical porphyry Mo-related silicic rocks (log FMQ = + 2 to + 3; Keith et al. 1993).

Fig. 10
figure 10

a logƒO2 versus T/°C for the Mada mineralized and coarse-grained biotite granites. b Ce4+/Ce3+ versus Eu/Eu* for the Mada mineralized and coarse-grained biotite granites. All parameters were computed from zircon composition. Note: NNO = nickel-nickel oxide buffer, FQM = fayalite- quartz-magnetite buffer, Hm = hematite buffer, and IW = iron-wüstite buffer

Full size image

Comparison with Sn ± Nb-mineralized granites in the Nigerian alkaline province

Magma composition, redox state, and degree of fractionation exert fundamental controls on the propensity of granitic magmas to cause Mo or Sn mineralization (Blevin et al. 1996; Černý et al., 2005). The occurrence, in the Nigerian Mesozoic alkaline granite complexes, of Sn ± (Nb) and Mo mineralization implies that petrogenetic factors suitable for the formation of both styles of mineralization were involved in the generation of the alkaline granites during the Mesozoic (Kinnaird and Bowden 1987; Girei et al. 2019a, b). The Mo-mineralized granite from the Kigom Complex yielded a U/Pb age of 166 ± 4 Ma (Hafizullah Abba Ahmed, pers. comm., 2020), close to the whole-rock Rb-Sr age of the nearby Jos-Bukuru Sn ± (Nb)-mineralized ring complex (Rahaman et al. 1984), suggesting that Sn ± (Nb)- and Mo-mineralized granites are spatially and temporally related. However, unlike in the Mo-mineralized ring complexes, the intrusive sequence in the Sn ± (Nb)-mineralized ring complexes almost invariably commenced with the emplacement of fayalite-bearing granite porphyry, implying that their parental magmas were probably more reduced than those of the Mo-mineralized granite (Imeokparia 1985; Girei et al. 2019a, b). This is corroborated by the calculated ƒO2 of the granite magma, which shows that Mo mineralization in the Mada Complex occurred under more oxidized conditions than Sn ± Nb mineralization in the Ririwai Complex (Fig. 10a–b).

A compilation of previous whole-rock data of aluminous granites associated with primary Sn ± Nb mineralization in the Ririwai and Tibchi ring complexes (Batchelor and Bowden 1986; Girei et al. 2019a, b), and of peralkaline granites associated with Mo mineralization in the Kigom Complex (Imeokparia 1985), together with our data, permits further comparisons between Mo- and Sn ± Nb-mineralized granites in Nigeria. Compared to the Mo-mineralized granites, the Sn± Nb-mineralized granites are more silicic and have higher Rb, higher Rb/Sr, larger Eu anomalies, and a higher differentiation (Thornton and Tuttle 1960) but lower Sr and Zr/Hf ratios, indicating that they are more fractionated than the Mo-mineralized granites (Fig. 11a–d). This is in broad agreement with the observation made by Blevin and Chappell (1995) in eastern Australia, where mineralization by chalcophile elements (Mo and Cu) was associated with less fractionated granites than mineralization by lithophile elements (Sn and Nb).

Fig. 11
figure 11

a Plot of Zr/Hf versus Eu/Eu*. b Plot of differentiation index (DI) versus SiO2. c Plot of Rb/Sr versus Sr. d Plot of Rb/Sr versus Eu/Eu*. Data plotted include those for Mo-mineralized granites from the Mada Complex (this study), and Kigom Complex (Imeokparia 1985), and Sn ± Nb-mineralized granites from the Ririwai Complex (Girei et al. 2019a, b) and Tibschi Complex (Batchelor and Bowden 1986). Blue box indicates field of “normal” (unfractionated) granites that exhibit charge and radius controlled (CHARAC) behavior of HFSE (Zr, Hf) as defined by Bau (1996)

Full size image

Comparison with other Mo-mineralized within-plate granites

Economically significant Mo mineralization associated with highly differentiated within-plate silicic alkaline rocks (i.e., A-type granites) has also been reported from many parts of the world, especially the western USA and eastern Qinling Mountains, China (Carten et al. 1993; Blevin 2004; Audétat and Li 2017). Compared to typical within-plate A-type suites containing porphyry Mo, such as the Climax-type Henderson deposits (CO, USA), the Mada and Kigom A-type granites have lower Rb concentrations (< 500 ppm) and Rb/Sr (< 30) and Nb/Zr (< 1) ratios (Carten et al. 1993; Keith et al. 1997; EMS 3 Fig. S3a-b). This suggests that the Mada and Kigom A-type granites are less differentiated than the A-type suite associated with the Henderson Mo deposits (Carten et al. 1993; Mercer et al. 2014). On the other hand, except for their higher Zr and Nb contents, the Mada and Kigom A-type granites share similar geochemical features with the Climax-type Questa deposits (NM, USA), which are regarded as a “transitional-type” Mo deposit (Carten et al. 1993; Rosera et al. 2013).

Genetic model

The whole-rock and zircon elemental data indicate that the Mada A-type granites are highly fractionated and have relatively high ƒO2, characteristic features that collectively favored the formation of Mo mineralization (Černý et al., 2005; Audétat and Li 2017). In an oxidized aluminous melt, the concentration of Mo increases with increasing degree of magmatic differentiation prior to its final removal into exsolved ore-forming fluids at a late magmatic stage (Audétat and Li 2017). Previous studies have suggested that mafic magmas in a Climax-type mineralization system probably contribute significant amounts of volatiles (S and CO2) and Mo to the mineralizing magmas through volatile fluxes (Keith et al. 1997; Audétat and Li 2017). Mafic rocks are present in the Mada Complex, and it is possible that their parental magmas, apart from contributing volatiles to the mineralizing magmas, may have also contributed heat that served to extend the lifetime of the magma chamber and enhanced the degree of fractional crystallization (Audétat and Li 2017; Zhang et al. 2018).

Molybdenum mineralization in the Mada Complex probably occurred in two separate stages (Figs. 3a–f, 12a–b). During Stage I, high-temperature potassic alteration was accompanied by deposition of disseminated Mo minerals (Fig. 12a). During Stage II, emplacement of greisen-related disseminated Mo mineralization with minor stockwork veins occurred, a process that was accompanied by near-complete replacement of biotite by chlorite and greenish mica due to fluid-rock interactions (Fig. 12b).

Fig. 12
figure 12

Schematic diagram showing the two-staged molybdenum ore-forming event in the Mada Ring Complex. a Stage I. bStage II. See text for detail explanation

Full size image

Conclusions

  1. 1.

    Integrated zircon U-Pb dating and molybdenite Re-Os systematics indicate that granite emplacement and Mo mineralization in the Mada Ring Complex of north-central Nigeria took place between 152 ± 1 and 150 ± 1 Ma, corresponding to the Late Jurassic.

  2. 2.

    Granites associated with Mo mineralization at Mada show elevated whole-rock concentrations of SiO2, Nb, Zr, Y, and REEs (except Eu) and high 87Rb/86Sr ratios, but low concentrations of Sr, Ba, CaO, MgO, and TiO2, typical of highly differentiated A-type granites. They yielded negative whole-rock εNd(t) (− 5.87 to − 5.47) and zircon εHf(t) (− 4.97 to − 10.06), indicating that both the A-type magma and Mo ore components were largely sourced from the lower crust with a contribution from the upper mantle.

  3. 3.

    Compared to their associated Sn ± (Nb)-mineralized granites, the Mada A-type granites are less differentiated but have higher ƒO2. These differences in redox state and degree of fractionation suggest that, despite their close temporal relationship, co-existing Mo- and Sn ± (Nb)-mineralized A-type granites in the Nigerian alkaline province were probably derived from different source(s) and/or may have experienced contrasting evolutionary histories.