Introduction

Granitoids comprise significant portions of the continental crust; they display great variability in their sources, their association with tectonics, and their specific magma evolutionary processes; studying the processes involved in their generation may provide important clues to the growth and reworking of continental crust and to regional tectonics and geodynamic processes (Zhao et al. 2016; Goodege & Vervoort 2006; Heilimo et al. 2014; Kemp & Hawkesworth 2003; Ren et al. 2022; Sun et al. 2021). The granitoid can be subdivided into S-, I-, M-, and A-type based on their chemical and mineralogical features (Barbarin 1999; Chappell & White 1992; Eby 1990; Loiselle and Wones 1979). Different magmatic assemblages are closely associated in space and time; these magmatism usually formed during the transition of tectonic regime, such as I- and A-type assemblages in the western margin of the Yangtze Block (Zhao et al. 2008), the Central Tianshan orogen (Dong et al. 2011), the Sanandaj-Sirjan Zone, NW Iran (Sarjoughian et al. 2016), and the Pataz gold-mining district in northern Peru (Witt et al. 2014).

The Lufilian Arc is known for its world-class sediment-host Cu–Co deposits which is located in central Africa (Cailteux et al. 2005; El Desouky et al. 2009; Hitzman et al. 2010; Key et al. 2001; Muchez et al. 2008; Eglinger et al. 2013). It is a large arcuate structure covering eastern Angola, the southern Democratic Republic of Congo, and northwestern Zambia (Katongo et al. 2004; Key et al. 2001; Batumike et al. 2006; Master et al. 2005). Geotectonically, the Lufilian Arc is located between the Congo Craton and the Kalahari Craton, and forms a part of a transcontinental network of Neoproterozoic-early Paleozoic orogenic belts in central-southern Africa together with the Zambezi Belt and the Damara Belt (Fig. 1(b)) (Katongo et al. 2004; Hanson et al. 1993; Dirks & Sithole 1999; Kampunzu & Cailteux 1999; Vinyu et al. 1999; Porada & Berhorst 2000). The Lufilian Arc is divided into five tectonic zones, which are the Plateau Foreland Basin, the External Fold and Thrust Belt, the Domes Region, the Synclinorial Belt, and the Katanga High from north to south (Fig. 1a) (De Swardt et al. 1965; Unrug 1983; Porada 1989; Kampunzu et al. 2000).

Fig. 1
figure 1

Simplified geological map of central-southern Africa showing the distribution of tectonic features discussed in text (modified from Eglinger et al. 2016; Johnson et al. 2005). BB, Bangweulu Block; CFB, Cape Fold Belt; CKB, Choma-Kalomo Block; DB, Damara Belt; GB, Garip Belt; GCB, Ghanzi-Chobe Belt; IB, Irumide Belt; KB, Kaoko Belt; KbB, Kibaran Belt; KhB, Kheis Belt; LB, Lufilian Belt; LmB, Limpopo Belt; MozB, Mozambique Belt; NaqB, Namaqua Belt; NB, Natal Belt; Ri, Richtersveld Terrane; UB, Ubendian Belt; UsB, Usagaran Belt; WCB, West Congo Belt; ZB, Zambezi Belt; ZC, Zimbabwe Craton; I-External Fold and Thrust Belt; II-Domes Region; III-Synclinorial Belt; IV-Kantanga High; V-Foreland Basin

Full size image

Regional investigations show that the Lufilian Arc underwent a phase of magmatic quiescence after the formation of Paleoproterozoic basement, which lasted until Neoproterozoic (Porada & Berhorst 2000; Key et al. 2001; Cosi et al. 1992; Hanson et al. 1994; Rainaud et al. 2005; Katongo et al. 2004). However, there are a substantial quantity of Mesoproterozoic detrital zircons developed in the belt, and the provenance of detrital zircons is controversial; most scholars believe that they are mainly derived from the surrounding Mesoproterozoic belt, such as the Kibaran Belt and the Irumide Belt (Liu et al. 2019; Master et al. 2005). However, detrital zircons are generally sub-angular and exhibit clear fine oscillatory growth zoning, suggesting a magmatic origin and close detrital provenance (Liu et al. 2019; Xu et al. 2021). Furthermore, a large number of ca. 1.10 Ga detrital zircons developed in the Solwezi area have no corresponding magmatism in the Kibaran Belt. Thereby, Xu et al. (2021) speculate that they are mainly derived from the underlying basement. In addition, the absence of Mesoproterozoic magmatism in the Lufilian Arc makes it impossible to comparative study of magmatism with the surrounding terrains, thus limiting the reconstruction of Mesoproterozoic supercontinent and adversely affecting the study of Precambrian geodynamics.

During the geological survey in the north of Zambia, we discovered Mesoproterozoic granitoids for the first time in the Solwezi Dome. Combining zircon U–Pb dating, whole-rock geochemical data, and Sr–Nd-Hf isotope analyses, this paper aims to address the following points: (1) the types and ages of granitoids, (2) petrogenesis and tectonic setting, (3) source of Mesoproterozoic detrital materials and tectonic significance represented by granitoids in the Lufilian Arc.

Geological setting

The Domes Region is characterized by the occurrence of basement inliers defining an arcuate chain parallel to the trend of folds developed during the Lufilian Pan African orogeny (Kampunzu et al. 2000). The study area is mainly located in the Solwezi Dome, where the basement and the Neoproterozoic Katanga Supergroup are developed (Fig. 2).

Fig. 2
figure 2

Simplified map showing the distribution of granitoids in Solwezi area (modified from Johnson et al. 2005; Selley et al. 2005)

Full size image

The basement is composed of the Lufubu schists and gneisses, which are intruded by the Eburnian (ca. 2200 ~ 1800 Ma) granites (Key et al. 2001; Rainaud et al. 2005). The Lufubu schists are usually calc-alkaline and formed as a result of the subduction of oceanic lithosphere between 2.1 and 1.8 Ga (Rainaud et al. 2005). The Mesoproterozoic magmatism had not previously been identified until the present geological survey, and the Nchanga Granite is the youngest intrusion in the pre-Katangan basement with a zircon U–Pb age of 883 ± 10 Ma (Armstrong et al. 2005; Master et al. 2005).

The Neoproterozoic Katanga Supergroup which unconformably overlies the basement is consisted of a ca. 10,000-m-thick sequence of sedimentary and metasedimentary rocks; it can be divided into the Roan Group, the Nugba Group, the Kundelungu Group, and the Biano Group from bottom to top (Cailteux 1994, Cailtrux et al. 2007; Batumike et al. 2007; Cailteux & Putter 2019). The Roan Group contains important Cu–Co mineralization (Armstrong et al. 2005; Master et al. 2005); it is characterized by basement gravels and detrital materials with overlying carbonates interbedded with shales; the Mwashya Subgroup at the top is composed of black shales and dolomitic siltstones. The stratigraphic succession shows that the sedimentary hydrodynamic environment changed from shallow to deep (Cailteux 1994; Cailteux et al. 2005). The Nguba and Kundelungu groups are two sedimentary sequences having diamictites at the bottom with overlying carbonates, followed by predominantly siliciclastic sedimentary rocks at the top (Batumike et al. 2006, 2007). The Biano Group occurs at the top of the Katanga succession, which is a continental clastic molasse-type deposit consisting of coarse to fine-grained arkoses and intraclastic conglomerates (Cailteux & Putter 2019).

At the late stage of sedimentation, the Katanga Supergroup was deformed by the Lufilian orogeny as a consequence of an interaction between the Congo Craton and the Kalahari Craton; metamorphic assemblages are principally in the greenschist facies, but higher grades up to eclogite facies have been locally recorded (Porada & Berhorst 2000; Naydenov et al. 2014; Rainaud et al. 2005). Eclogite facies metamorphism from Central Zambia yielded a Sm–Nd isochron at 595 ± 10 Ma, and records the timing of the subduction which took place in an oceanic environment (John et al. 2003). While a phase of high-pressure whiteschist metamorphism yielded a U–Pb monazite age of 529 ± 2 Ma, it represents the final stage of the collision between the Congo and the Kalahari cratons (John et al. 2004).

Petrography

The granitic complex is located in the northeastern part of the Solwezi Dome, which is funnel-shaped with an area of about 35km2 (Fig. 2). The complex intruded into the Lufubu schists with the Katanga Supergroup unconformably overlaid in the north. It exhibits typical granitic characteristics with some minerals extending roughly N-S or NW–SE because of later regional tectonics (Fig. 3a, b). The complex is mainly composed of the gneissic K-feldspar granite and the gneissic biotite monzogranite; the latter often occurs copper mineralization (Fig. 3c, d). The contact relationship between them has not been detected owing to the extensive coverage of the study area. The dating samples of the K-feldspar granite (ZS50) and the biotite monzogranites (ZS29, ZS30) are collected from the northern quarry and the southern quarry, respectively. The sample list and corresponding GPS coordinates are given in Table 1.

Fig. 3
figure 3

Photographs of the outcrops and photomicrographs of the K-feldspar granite and the biotite monzogranite in the Solwezi Dome. a The gneissic K-feldspar granite; b the gneissic biotite monzogranite; c, d mineralized biotite monzogranite; e, f photomicrographs showing mineral assemblages and textural characteristics of the K-feldspar granite (ZS45) and the biotite monzogranite (ZS34), respectively. Clp: Chalcopyrite; Mal: Malachile; Kf: K-feldspar; Pl: Plagioclase; Mc: Microcline; Bi: Biotite; Q: Quartz

Full size image
Table 1 Sample locations and corresponding coordinates (WGS 84)
Full size table

The K-feldspar granites are pink with porphyritic texture and massive/gneissic structure. The minerals are mainly composed of the K-feldspar (40%), plagioclase (25%), quartz (25%), and biotite (4%) as well as a small amount of accessory minerals including zircon, apatite, and Fe–Ti oxide, etc. (6%). The K-feldspars are long columnar with Carlsbad twinning, and some appear as microcline with weakly sericitized and kaolinitized. The plagioclases show albite twin with weakly sericiticized. The biotites are flaky and typically along the boundaries between the K-feldspars and plagioclases, suggesting they are late crystallization phases (Fig. 3e).

The biotite monzogranites are gray with porphyritic texture and massive/gneissic structure. The minerals are mainly composed of the K-feldspar (35%), plagioclase (30%), quartz (20%), biotite (8%), and muscovite (2%) and a small amount of accessory minerals including zircon, apatite, and Fe–Ti oxide, etc. (5%). The K-feldspars are long columnar with Carlsbad or crossed twinning. The plagioclases show albite twin and are weakly sericiticized. The quartz and biotites are often filled between the feldspars, indicating that they are also late crystallization phases (Fig. 3f).

Analytical methods

LA-MC-ICP-MS zircon U–Pb dating

Zircons were separated by using standard density and magnetic separation techniques and then random zircon grains were handpicked under a binocular microscope at the Yu’neng Geological and Mineral Separation Survey Centre of Langfang, Hebei Province, China. Zircons were mounted in epoxy in a 1.4-cm diameter circular grain mount and then polished to section the crystals in half for analysis at the GeoAnalysis CO. Ltd., Beijing. Meanwhile, transmitted and reflected light images and cathodoluminescence (CL) images were made at the same laboratory, using a JXA-8800 R electron microprobe and a JXA-8800 R electron microprobe, respectively. Lastly, zircon U–Pb analyses were conducted on an Agilent 7500 a ICP-MS equipped with a New Wave Research laser-ablation system at Institute of Geology and Mineral Resources in Tianjin, China (TIGMR). The laser system delivered a beam of 193-nm UV light from a frequency-quintupled Neptune instrument (Thermo Fisher Company). Analyses were carried out with a beam diameter of 35 μm, a repetition rate of 8–10 Hz, and an energy of 10–11 J/cm2. Data acquisition for each analysis took 20 s for the background and 40 s for the signal. Zircon 91,500 and silica glass NIST610 were used as external standards for optimizing instruments. An Excel-based software ICP-MS-Data-Cal was used to perform offline selection and integration of background and analyzed signals, time-drift correction, and quantitative calibration for trace element analysis and U–Pb dating (Liu et al. 2008, 2010a, b). Concordia diagrams and weighted mean age calculations were made using ISOPLOT 4.15 software (Ludwig 2003).

Zircon Lu–Hf isotopic analyses

The in situ Lu–Hf isotopes were analyzed on the same spots or in adjacent domains of zircon grains with similar texture where U–Pb dating was done at TIGMR, following analytical techniques described by Yuan et al. (2004).The energy density of 15 ~ 20 J/cm2 and a spot diameter of 50 μm are used in this study. GJ-1 was used as the reference standard, with a weighted mean 176Hf/177Hf ratio of 0.282004 ± 46 (2σ, n = 31) during our routine analyses. In this study, the decay constant of 176Lu of 1.865 × 10−11/a was adopted (Scherer et al. 2001). Initial 176Hf/177Hf ratios and εHf(t) values were calculated with reference to the chondritic reservoir of Blichert-Toft and Albarède (1997). The depleted-mantle Hf model age (TDM1) was calculated with present-day 176Hf/177Hf (0.28325) and 176Lu/177Lu (0.0384) values (Griffin et al. 2000). Two-stage “crustal” model ages (TDM2) were calculated for the average continental crust with a 176Lu/177Hf ratio of 0.015 (Griffin et al. 2002).

Whole-rock geochemical analyses

For major, trace and rare earth element analyses, the rocks were reduced in a jaw crusher and then powdered to 200 mesh at the Yu’neng Laboratory. Whole-rock major elements were identified by using a PW4400/40 X-ray fluorescence spectrometer at Tianjin Institute of Geology and Mineral Resources (TIGMR), Tianjin. The contents of Fe2O3 and FeO were analyzed through conventional wet chemistry and titration. The analytical precision is generally better than 2% for all the major elements. Trace element and rare earth element (REE) abundances were measured by using an X Series II ICP-MS at TIGMR with analytical precisions better than 5%. The detailed analytical procedure followed that of Gao et al. (2003).

Whole-rock Sr–Nd isotopic analyses

Samples used for Rb–Sr and Sm–Nd isotope analysis were spiked with mixed isotope tracers, dissolved 7 days in teflon capsules using HF, HClO4, and HNO3 acids, and separated by conventional cation-exchange techniques (AG50 × 12). Isotopic measurements were performed using a FinniganTritonTi thermal ionization mass spectrometer at the TIGMR. Procedural blanks yielded concentrations of < 200 pg for Sm and Nd, and < 500 pg for Rb and Sr. Mass fractionation corrections for Sr and Nd isotopic ratios were based on 88Sr/86Sr = 8.375209 and 146Nd/144Nd = 0.7219, respectively, and analysis of the NBS987 and LRIG standards yielded values of 87Sr/86Sr = 0.710245 ± 30 (2σ) and 143Nd/144Nd = 0.512202 ± 30 (2σ), respectively.

The εNd(t) values and Nd model ages were calculated assuming 147Sm/144Nd and 143Nd/144Nd ratios for average chondrite and the depleted mantle at the present day to be 0.1967 and 0.512638, and 0.2137 and 0.51315, respectively (Hamilton et al. 1983; Goldstein et al. 1984). Here, λRb and λSm are 1.42 × 10−11 year−1 and 6.54 × 10−12 year−1, respectively (Lugmair & Harti 1978).

Analytical results

Zircon U–Pb age

The U–Pb isotopic data of the K-feldspar granites (ZS50) and the biotite monzogranites (ZS29 and ZS30) are listed in Table 2.

Table 2 LA-MC-ICP-MS zircon U–Pb dating results of the late-Mesoproterozoic granitoids in the Solwezi Dome
Full size table

The K-feldspar granite

The majority of zircons collected from ZS50 are euhedral-subhedral with lengths and length/width ratios ranging from 80 to 200 μm and 2:1 to 3:1, respectively. Visible oscillatory zoning in cathodoluminescence (CL) images (Fig. 4a) and high Th/U ratios (usually > 0.1) indicate a magmatic origin of these zircons (Hoskin & Schaltegger 2000). However, the disruption of oscillatory zoning and development of irregular domains of homogenous can be observed locally due to later metamorphism; therefore, the least severely affected zircon domains are chosen during the analysis to avoid getting a “mixed” isotopic ages (Corfu & Hanchar 2003).

Fig. 4
figure 4

Cathodoluminescence (CL) images for representative zircons (a, c, e) and U–Pb concordia diagrams for zircons (b, d, f) from the granitoids in the Solwezi Dome

Full size image

Forty analyses were conducted on thirty-nine zircons, one inherited zircon with an old age of 1744 ± 44 Ma, and eight zircons with a low concordant degree (< 90%) were rejected; the remaining thirty-one analyses yield a discordant line with an upper intercept at 1181 ± 18 Ma (MSWD = 1.3), fourteen analyses of which give a weighted mean 207Pb/206Pb age of 1178 ± 15 Ma (MSWD = 0.37) (Fig. 4b), interpreted as crystallization age of the K-feldspar granite.

The biotite monzogranite

Zircons collected from ZS29 and ZS30 are euhedral-subhedral with lengths and length/width ratios ranging from 60 to 210 μm and 2:1 to 3:1, respectively (Fig. 4c, e). The CL images show that zircons mostly have visible oscillatory zoning, and their Th/U ratios are 0.02 ~ 2.12 and 0.01 ~ 1.60, respectively, which support their magmatic origin and later weakly partial metamorphism (Rubatto 2017).

Forty analyses were conducted on forty zircons of ZS29, yielding concordant 207Pb/206Pb ages of 1009 ~ 1169 Ma, with a weighted mean age of 1105 ± 7 Ma (MSWD = 0.60) (Fig. 4d), interpreted as crystallization age of the biotite monzogranite. Forty analyses were conducted on thirty-nine zircons of ZS30, except for one zircon (23) with an old age of 1435 ± 78 Ma, the remaining thirty-nine analyses yield concordant 207Pb/206Pb ages of 1035 ~ 1254 Ma, with a weighted mean age of 1109 ± 7 Ma (MSWD = 1.09) (Fig. 4f), which is compatible with ZS29 within the error.

Zircon Lu–Hf isotopic data

Zircons from the three samples dated by U–Pb were also analyzed for their Lu–Hf isotopes on the same domains (but a larger range), and the results are listed in Table 3 and illustrated in Fig. 5a. The 176Lu/177Hf ratios of all the zircons are less than 0.002, indicating that there is no significant accumulation of radiogenic Hf following zircon crystallization (Patchett et al. 1982).

Table 3 Zircon in situ Lu–Hf isotope data of the late-Mesoproterozoic granitoids in the Solwezi Dome
Full size table
Fig. 5
figure 5

εHf(t) vs. zircon 207Pb/206Pb age (a) and εNd(t) vs. U–Pb age (b) for the granitoids in the Solwezi Dome

Full size image

Thirty-seven analyses were obtained for ZS50, and the crystallization age of 1178 Ma was used for the calculation of εHf(t) and TDM values, yielding εHf(t) values between − 12.03 and − 7.10, corresponding to TDM1 model ages between 1989 and 2168 Ma and TDM2 model ages between 2423 and 2728 Ma.

Twenty-nine analyses were obtained for ZS29, and the crystallization age of 1105 Ma was used for the calculation of εHf(t) and TDM values, yielding εHf(t) values between − 14.85 and − 9.99, corresponding to TDM1 model ages between 2017 and 2214 Ma and TDM2 model ages between 2547 and 2847 Ma.

Thirty analyses were obtained for ZS30, and the crystallization age of 1109 Ma was used for the calculation of εHf(t) and TDM values, yielding εHf(t) values between − 13.06 and − 9.35, corresponding to TDM1 model ages between 1983 and 2224 Ma and TDM2 model ages between 2511 and 2740 Ma.

Major and trace elements

The data of major and trace elements for the K-feldspar granites and the biotite monzogranites are listed in Table 4.

Table 4 Whole-rock major (wt%) and trace (ppm) element contents of the late-Mesoproterozoic granitoids in the Solwezi Dome
Full size table

Alteration and element mobility

As the Solwezi granitoids had suffered deformation and greenschist facies metamorphism, it is imperative to ascertain if metamorphism may have led to significant post-emplacement element mobility that may preclude use of compositional data to make petrogenetic inferences. All the samples have low loss on ignition values (0.70 ~ 1.22%), and the chemical indices of alteration of 43.6 ~ 52.5 are comparable to those of unaltered granite (50, Nesbitt & Young 1982; Katongo et al. 2004), implying that all the samples have rarely been affected by later metamorphism and alteration. The high field strength elements (HFSEs) and REEs are immobile during secondary processes whereas the large ion lithophile elements (LILEs) are considered mobile (Polat & Hofmann 2003). The Solwezi granitoids were assessed for post-emplacement alteration by plotting Zr (a known sensitive indicator of immobility) against REE (represented by La), HFSE (represented by Nb), and LILE (represented by Sr) as well as some major elements (K2O, Na2O, and Al2O3) (Fig. 6). The results show that all elements have a good correlation with Zr, suggesting that the HFSE, the REE, the LILE, and the major elements were not significantly affected by post-emplacement processes, thus allowing their use for purposes of drawing petrogenetic inferences.

Fig. 6
figure 6

Plot of Zr vs (a) La, (b) Nb, (c) Sr, (d) K2O, (e) Na2O and (f) Al2O3 for the granitoids in the Solwezi Dome

Full size image

The K-feldspar granite

The K-feldspar granites have variable contents of SiO2 (71.5 ~ 73.1 wt%), FeOt (FeO + 0.8998 × Fe2O3) (1.62 ~ 2.08 wt%), CaO (0.92 ~ 1.11 wt%), TiO2 (0.20 ~ 0.27 wt%), and P2O5 (0.066 ~ 0.11 wt%). The MgO contents are 0.43 ~ 0.65 wt% and Mg# values are 30 ~ 39. CIPW-normative compositions indicate that they have compositions similar to monzogranite (Fig. 7a). Their K2O and Na2O contents are 6.04 ~ 6.26 wt% and 2.59 ~ 3.12 wt%, respectively, with K2O/Na2O ratios of 1.31 ~ 2.41, mainly falling in the field of sub-alkaline granites on the TAS diagram (Fig. 7b) and the high-K calc-alkaline series on the K2O vs. SiO2 diagram (Fig. 7c). They are metaluminous to weak peraluminous (Al2O3 = 13.4 ~ 14.4 wt%, A/CNK = 0.84 ~ 1.06) on the A/CNK vs. A/NK diagram (Fig. 7d).

Fig. 7
figure 7

(a) Q–A-P plot (Streckeisen 1976); (b) (K2O + Na2O) vs. SiO2 (Middlemost 1994); (c) K2O vs. SiO2 (Roberts & Clemens 1993); (d) A/NK vs. A/CNK (Maniar & Piccoli 1989) diagrams for the granitoids in the Solwezi Domes

Full size image

They have high REE content of 152 ~ 367 ppm, and strongly fractionated REE patterns [(La/Sm)N = 3.99 ~ 4.63, (Gd/Yb)N = 6.92 ~ 8.38], as well as medium negative Eu anomalies (δEu = 0.42 ~ 0.77) (Fig. 8a). In the primitive mantle-normalized spidergram, they are characterized by enrichment in Rb, Th, K, U, and depletion in Ba, Nb, Ta, Sr, P, and Ti (Fig. 8b). Meanwhile, their Sr/Y ratios range from 17.3 to 46.2.

Fig. 8
figure 8

Chondrite-normalized REE patterns and Primitive Mantle (PM) normalized trace element diagrams for the K-feldspar granites (a, c) and the biotite monzogranites (b, d) in the Solwezi Dome. The values of chondrite and PM are from Sun & McDonough (1989)

Full size image

The biotite monzogranite

Compared with the K-feldspar granite, the biotite monzogranites have higher FeOt (2.58 ~ 4.95 wt%), CaO (1.35 ~ 2.16 wt%), TiO2 (0.44 ~ 0.91 wt%), P2O5 (0.23 ~ 0.36 wt%), MgO (0.80 ~ 1.52 wt%), Na2O (2.68 wt% ~ 3.44 wt%), Mg# (32 ~ 40), and lower SiO2 (66.1 ~ 71.4 wt%), K2O (3.84 ~ 5.68 wt%), and K2O/Na2O ratios (1.12 ~ 2.22). CIPW-normative compositions are also similar to monzogranite (Fig. 7a). The majority of samples fall in the field of sub-alkaline granites (Fig. 7b) and the high-K calc-alkaline series (Fig. 7c). They are weak peraluminous (Al2O3 = 13.5 ~ 15.7 wt%, A/CNK = 1.01 ~ 1.11) (Fig. 7d).

Their REE contents are also high (371 ~ 667 ppm) and they are characterized by strongly fractionated LREE patterns [(La/Sm)N = 3.99 ~ 4.63] and relatively flat HREE patterns [(Gd/Yb)N = 3.38 ~ 5.76], as well as obviously negative Eu anomalies (δEu = 0.28 ~ 0.45) (Fig. 8c). In the primitive mantle-normalized spidergram, their geochemical characteristics are similar to the K-feldspar granites (Fig. 8d). The Sr/Y ratios (4.87 ~ 13.0) are lower.

Whole-rock Sr–Nd isotope

The data of Sr–Nd isotope for the granitoids are listed in Table 5 and illustrated in Fig. 5b. The whole-rock Sr–Nd isotope characteristics of all the granitoids are comparable, indicating a similar source.

Table 5 Whole-rock Rb–Sr and Sm–Nd isotopic composition of the late-Mesoproterozoic granitoids in the Solwezi Dome
Full size table

Initial 87Sr/86Sr ratios and εNd(t) values of the K-feldspar granites were calculated using an age of 1178 Ma. Their initial 87Sr/86Sr ratios are between 0.7162 and 0.7217, and the age-corrected εNd(t) values vary from − 9.17 to − 9.24, corresponding to TDM2 ages of 2523 ~ 2529 Ma.

Initial 87Sr/86Sr ratios and εNd(t) values of the biotite monzogranites were calculated using an age of 1105 Ma. Their initial 87Sr/86Sr ratios are between 0.7246 and 0.7305, and the age-corrected εNd(t) values vary from − 8.67 to − 9.98, corresponding to TDM2 ages of 2424 ~ 2529 Ma.

Discussion

Petrogenetic type: S-type, I-type, or A-type?

All the granitoids in the Solwezi Dome have high initial 87Sr/86Sr ratios and negative εNd(t) values, distinguishing them from those of M-type granite. Furthermore, they are metaluminous to weak peraluminous, contain no aluminum-rich minerals (garnet or cordierite) and have low CIPW normative corundum (0 ~ 0.74% and 0.51 ~ 1.50%, respectively), in contrast to strongly peraluminous S-type granites formed by partial melting of metapelitic protolith (Fig. 5d) (Clemens 2003; Zhao et al. 2018). As a result, the granitoids in the Solwezi Dome should be I-type or A-type.

The K-feldspar granite

The A/CNK values (0.84 ~ 1.08) and P2O5 contents (0.07 ~ 0.11 ppm) of the K-feldspar granites are low, and there is an increase in Y and Th as Rb increases (Fig. 9n, o), which are typical characteristics of I-type granite (Maniar & Piccoli 1989; Li et al 2007; Liu et al. 2009). The calculated zircon saturation temperatures (Tzr) for the K-feldspar granites using the equation of Watson & Harrison (1983) are 771 ~ 806℃, which are similar to the formation temperature of I-type granite (781℃, King et al. 1997) (Fig. 10). Meanwhile, all of the samples fall in the transition field between fractionated and unfractionated I-, S-, and M-type granite on the petrogenetic discrimination diagrams (Fig. 11). Therefore, we believe that the K-feldspar granites are I-type.

Fig. 9
figure 9

Harker plots of major and selected trace elements of the granitoids in the Solwezi Dome. Solid line range is the experimentally produced melts of crustal origin (Skjerlie & Johnston 1993)

Full size image
Fig. 10
figure 10

SiO2 vs. the calculated zircon saturation temperature for the granitoids in the Solwezi Dome. The calculated zircon saturation temperatures are after Watson & Harrison (1983), the temperatures of I- and A-type granite are from King et al. (1997)

Full size image
Fig. 11
figure 11

FeOt vs. (Zr + Nb + Ce + Y) (a), (K2O + Na2O)/CaO vs. (Zr + Nb + Ce + Y) (b), Nb vs. 10,000*Ga/Al (c), and Zr vs. 10,000*Ga/Al (d) discrimination diagrams of Whalen et al. (1987), showing that the K-feldspar granites are I-type whereas the biotite monzogranite are A-type. FG: Fractionated felsic granites; OGT: unfractionated I-, S-, and M-type granites

Full size image

The biotite monzogranite

The biotite monzogranites have relatively low Nb (17.6 ~ 45.9 ppm), Ce (156 ~ 310 ppm), Zr (271 ~ 546 ppm), Y (25.7 ~ 45.4 ppm), Zr + Nb + Ce + Y (479 ~ 823 ppm), and 10,000 × Ga/Al ratios (2.85 ~ 3.42), which are similar to the geochemical characteristics of A-type granite (Whalen et al. 1987; Collins et al. 1982). Their Tzr are 836 ~ 902℃ (Fig. 10), which are obviously higher than I- and S-type granite but similar to A-type granite (839℃, King et al. 1997). All the samples fall in the A-type granite field (Fig. 11). Therefore, we believe that the biotite monzogranites are A-type.

Petrogenesis

The K-feldspar granite

The K-feldspar granites have high initial 87Sr/86Sr values (0.7162 ~ 0.7217), negative whole-rock εNd(t) values (− 9.17 ~  − 9.24) and zircon εHf(t) values (− 12.03 ~  − 7.10), corresponding to TDM2 ages of 2523 ~ 2529 Ma and 2423 ~ 2728 Ma, respectively, implying that the granitic magma was primarily derived from anatexis or remelting of Neoarchean crust (Allègre & Othman 1980). The Mg# is a useful index of discriminating melts of purely crustal origin from those that have interacted with mantle-derived magma. Melts from the basaltic lower continental crust are characterized by low Mg# values (< 40) regardless of the degree of melting, whereas those with higher Mg# values (> 40) can only be generated by involvement with a mantle component (Rapp & Watson 1995; Liu et al. 2014). The Mg# values of the K-feldspar granites are 30 ~ 39, which fall in the crustal melts and the lower crustal melts (TTG) fields but deviate from the pure crustal partial melts on the SiO2–Mg# diagram (Fig. 12), implying that the magma might be primarily crustal-derived with a tiny component of mantle-derived magma. The zircon εHf(t) values are heterogeneous (up to 4.93 units), further implying a hybrid source of mixing between crust-melts and mantle-melts (Chiu et al. 2009). A large number of Neoarchean detrital zircons (2710 ~ 2460 Ma) (Liu et al. 2019; Master et al. 2005) and a suite of metamorphic and igneous rocks (2560 ~ 2540 Ma) (Key et al. 2001) have been identified in the Lufilian Arc, implying that Neoarchean was an important period of crustal accretion in this belt. Combined with the Neoarchean TDM2 ages; we can further suggest that the K-feldspar granites might be derived from partial melting of Neoarchean basement. Such a genetic relationship between K-rich granite and TTG protolith has also been widely documented in many Archean cratons (Mark 1999; Shang et al. 2007; Zhao et al. 2008). In addition, the K-feldspar granites have strongly fractionated LREE patterns [(La/Sm)N = 3.99 ~ 4.63] and HREE patterns [(Gd/Yb)N = 6.92 ~ 8.38], as well as medium negative Eu anomalies (δEu = 0.42 ~ 0.77) (Fig. 8a) and obviously depleted in Sr (Fig. 8b), but their Sr/Y ratios are high (17.32 ~ 46.19), implying that the source melting occurred at pressures between the garnet and plagioclase stability field, corresponding to a slightly thickened crust of 0.8 ~ 1.4 Gpa (Zhang et al. 2011; Zhao et al. 2018).

Fig. 12
figure 12

Mg.# vs. SiO2 diagram suggesting that the parental magma sources of the granitoids in the Solwezi Dome are not pure crustal partial melts but are mainly derived from crustal-melts (Dokuz 2011) and mixed with a small amount of mantle-melts (Kinzler, 1977). Data sources: the fields of pure crustal partial melts determined in experimental studies on the dehydration melting of low-K basaltic rocks at 8 ~ 16 kbar and 1000 ~ 1050 °C and medium- to high-K basaltic rocks at 7 kbar and 825 ~ 950 °C are from Rapp & Watson (1995) and Sisson et al. (2005), respectively. The crustal AFC is from Stern & Kilian (1996), and the TTG (lower-crustal melts) and adakite (slab melts) are taken from Condie (2005)

Full size image

The contents of MgO, FeOt, TiO2, Al2O3, Ni, V, Rb, and Zr stay essentially constant as SiO2 increases (Fig. 9), indicating a low-degree fractional crystallization of the K-feldspar granite; this result is compatible with Fig. 11. Rb/Sr and Sr, Ba and Sr, Dy and Er, (La/Yb)N, and La suggest that the magma underwent low-degree fractional crystallization of plagioclase, K-feldspar, amphibole, and some accessory minerals (e.g., monazite, epidote) (Fig. 13). Meanwhile, the depletion of Nb–Ta-Ti and P might be related to the fractionation of the Ti-bearing phases (e.g., ilmenite and titanite) and apatite, respectively. In summary, we can speculate that the K-feldspar granites were most likely formed by partial melting of Neoarchean basement in a slightly thickened lower crust, followed by low-degree fractional crystallization of plagioclase, K-feldspar, amphibole, and some accessory minerals.

Fig. 13
figure 13

Magma evolution discrimination diagrams for the granitoids in the Solwezi Dome. Rb/Sr vs. Sr (a), Ba vs. Sr (b), Dy vs. Er (c), and (La/Yb)N vs. La (d) showing the trending of fractional crystallization

Full size image

The biotite monzogranite

The biotite monzogranites have high initial 87Sr/86Sr values (0.7246 ~ 0.7305), negative whole-rock εNd(t) values (− 8.67 ~  − 9.98) and zircon εHf(t) values (− 14.85 ~  − 9.35), corresponding to TDM2 ages of 2424 ~ 2529 Ma and 2511 ~ 2847 Ma, respectively, implying that the magma was also derived from partial melting of Neoarchean crust. Melting experiments demonstrate that dehydration melting of fluid-bearing tonalitic and granodioritic rocks can generate A-type granitic melts (Skjerlie & Johnston. 1993; Patiño Douce 1997). The biotite monzograntes and the K-feldspar granites have similar geochemical components; both of them are within or close to the experimentally produced melts of crustal origin (Fig. 9a–i), indicating that the magma probably also derived from partial melting of Neoarchean basement. The Mg# values are 33 ~ 40, which fall in the crustal melts and the lower crustal melts (TTG) fields; however, they also have a clear evolution trend to the mantle-melts (Fig. 12). Combined with the higher MgO, TiO2, Cr, Ni, and the heterogeneous zircon εHf(t) values (up to 5.50 units), we speculate that some mantle melts might be involved during the crustal-dominated melting process. Meanwhile, the strongly fractionated LREE patterns [(La/Sm)N = 3.99 ~ 4.63] and relatively flat HREE patterns [(Gd/Yb)N = 3.38 ~ 5.76], obviously negative Eu anomalies (δEu = 0.28 ~ 0.45) (Fig. 8c), depleted in Sr and lower Sr/Y ratios (4.87 ~ 13.0), implying that the source might be at relatively shallow depths where plagioclase was present and garnet was absent, corresponding to the middle-lower crust of 0.8 ~ 1.0 Gpa (Patiño Douce & Beard 1995; Watkins et al. 2007; Zhao et al. 2008; Sarjoughian et al. 2016).

The biotite monzograntes display a higher degree fractional crystallization of plagioclase, K-feldspar, amphibole, apatite, titanite, and some accessory minerals than the K-feldspar granites (Fig. 13). In summary, the biotite monzograntes were also likely formed by partial melting of Neoarchean basement, but at relatively shallow depths with a minor quantity of mantle melts, followed by fractional crystallization of plagioclase, K-feldspar, amphibole, and some accessory minerals.

Tectonic setting and implications

Tectonic setting

Granitoid types were potentially proposed to be related to tectonic setting. A granitoid can well be used as a geodynamic indicator when it is correctly classified and also precisely dated (Pearce et al. 1984; Foerster et al. 1997; Barbarin 1999; Dong et al. 2011). It has been demonstrated that the granitoids formed in various tectonic settings usually have different geochemical features, which can be used to infer the formation environment (Pearce et al., 1984).

Previous researches have revealed that continental collision is an important factor that contributes to crustal thickening in the orogenic belt (Zhang et al. 2006; Wu et al. 2005). The petrogenetic analysis mentioned above reveals that the K-feldspar granites were the products of partial melting of the slightly thickened lower crust; therefore, they were most likely generated in a syn-collisional environment. On the tectonic discrimination diagrams of Rb-(Y + Nb), Nb-Y, Nb-SiO2, and R1-R2, the K-feldspar granites fall in the field of the syn-collisional granites (Fig. 14), further suggesting that they were generated in a syn-collisional environment.

Fig. 14
figure 14

Tectonic discrimination diagrams of the granitoids in the Solwezi Dome. (a) Nb versus Y, (b) Rb versus (Y + Nb) and (c) Nb versus SiO2 (Pearce et al. 1984), (d) R1-R2 diagram of Batchelor & Bowden (1985). Data source: the data of the Munali Hill Granite and the Mpande Gneiss are from Katongo et al. (2004)

Full size image

A-type granite is usually associated with extensional tectonic setting (Bonin 2007; Karsli et al. 2012; Martin et al. 1994; Zhao et al. 2008). The biotite monzogranites have characteristics of A-type granites, and they mainly fall in the fields of the post-collisional or the within-plate granites (Fig. 14), combined with the geochronological data and tectonic environment of the K-feldspar granite; we believe that the biotite monzograntes were likely generated in a post-collisional environment.

The provenance of detrital material

A substantial quantity of Mesoproterozoic detrital zircons were found in different areas of the Lufilian Arc, such as in the Chambishi Basin (1.50 ~ 1.05 Ga, Liu et al. 2019), Konkola and Kipushi (1231 ~ 1025 Ma, Master et al. 2005), Solwezi (1220 ~ 931 Ma, Xu et al. 2021), and Likasi (1.1 ~ 1.6 Ga, Rainaud et al. 2003). The U–Pb ages of the granitoids in the Solwezi Dome are between 1178 ± 15 and 1105 ± 7 Ma, which are consistent with the age peak of detrital zircons in the garnet mica schist near Solwezi (Xu et al. 2021), implying that the late-Mesoproterozoic granitoids were important sources for detrital materials of the Katanga Supergroup.

Buried Mesoproterozoic orogen in the Lufilian Arc

Tectonic discrimination diagrams and geochronological data show that the Mesoproterozoic granitoids in the Solwezi Dome are related with orogeny (Fig. 14). As we know, the Mesoproterozoic orogenies developed in Central Africa are the Kibaran event (1.42 ~ 1.0 Ga, Kokonyangi et al. 2004; Debruyne et al. 2015; Tack et al. 2010) and the Irumide event (1.05 ~ 0.95 Ma, De Waele et al. 2006, 2009). In the Irumide Belt, voluminous syn- to post-kinematic Irumide granitoids emplaced between 1.05 and 0.95 Ga, peak metamorphism is diachronous across the belt and bracketed between 1.05 Ga in the southeast and 1.02 Ga in the northwest. The zircon U–Pb ages of the Solwezi granitoids are 1178 ± 15 Ma ~ 1105 ± 7 Ma, which are obviously earlier than the emplaced ages of the Irumide granitoids, and the tectonic environment of syn- to post-collision is also inconsistent with the tectonic evolution stage of the Irumide Belt. Therefore, we speculate that the granitoids may be related to the Kibaran orogeny.

The explanation for the ca. 400 Ma magmatic and metamorphic history of the Kibaran Belt remains enigmatic (Tack et al. 2010; Debruyne et al. 2015). In the southern part of this belt, Kampunzu et al. (1998) consider that the Kibaran deformation events mark the development of an active continental collision (ca. 1.4 ~ 1.25 Ga), followed by a continental collision (ca. 1.25 ~ 1.0 Ga), but there is no consensus on this interpretation (Klerkx et al. 1987). The latest research on the arc-type signature of the mafic intrusions (Kokonyangi et al. 2006) and the depositional or erosional hiati (Fernandez-Alonso et al. 2012) indicate the proposed transition from accretionary tectonics and sedimentation toward a collisional setting during ca. 1420 ~ 1375 Ma (Debruyne et al. 2015); the following medium-pressure and medium-temperature metamorphism at 1079 ± 14 Ma (Kokonyangi et al. 2001) is mainly a far-field tectonic linkage between the Kibaran Belt and the Irumide Belt (Johnson et al. 2005).

Geological constraints on the margins of the Congo-Tanzania-Bangweulu lack evidence for collisional orogenesis at the late stage of the Kibaran orogeny; however, in the southern part of Africa, a substantial quantity of late-Mesoproterozoic magmatism has been identified, and one possible interpretation is that they are related to a buried southwestern continuation of the Kibaran Belt (Singletary et al. 2003; Bulambo et al. 2004). In the Choma-Kalomo Block, ages of the two main magmatic pulses have recently been confirmed by SHRIMP dating of zircon, which have yielded a crystallization age of 1368 ± 10 Ma (Singletary et al. 2003) for the syn-tectonic granitoid and ages between 1188 ± 11 and 1174 ± 27 Ma (Bulambo et al. 2004) for the post-tectonic granites (Hanson et al. 1988). Hanson et al. (1988) argued that the Choma-Kalomo Block represents part of a larger Mesoproterozoic orogen that continues in the subsurface into Botswana to the southwest along strike. Bulambo et al. (2004) described that the Choma-Kalomo Block may have been detached from the Kibaran Belt during a rift-drift stage of the Katangan basin. In the Ghanzi-Chobe Belt, U–Pb geochronological data constrain granitoid plutonism and amphibolite-facies ductile deformation to have occurred between 1.20 ~ 1.15 and 1.1 Ga (crystallization age of post-tectonic intrusion) (Singletary et al. 2003). In the Namaqua Belt exposed still farther southwest along strike, the rocks record ductile deformation, regional metamorphism at variable grades, and pre- to synorogenic magmatism at 1.38 to 1.20 Ga (Cornell et al. 1992; Hoal & Heaman 1995; Robb et al. 1999; Gutzmer et al. 2000). Singletary et al. (2003) inferred that the Choma-Kalomo Block, coeval parts of the Ghanzi-Chobe Belt, and the Namaqua Belt belong to a single, northeast-trending Mesoproterozoic orogen.

Because of thick Neoproterozoic and younger cover, it is impossible to directly trace this Mesoproterozoic orogen farther to the southwest (Key & Mapeo 1999). However, the K-feldspar granite, which is not only consistent with the contemporaneous granites in the Choma-Kalomo Block, the Ghanzi-Chobe Belt, and the Namaqua Belt in terms of petrogenesis and tectonic environment, but also allows these belts to be geographically connected to the Kibaran Belt, thus further proving that there is a buried southwestern continuation of the Kibaran Belt.

Extension-related magmatism in the Lufilian Arc

Bimodal volcanism is generally considered to be the dominant igneous assemblages in the continental rift zone (Tacket al., 2010; Kampunzu et al., 1998). A substantial quantity of late-Mesoproterozoic mafic and felsic volcanic rocks and small granitic intrusions have been identified in the buried Kibaran Belt (Singletary et al. 2003).

In the Ghanzi-Chobe Belt, all dated rhyolites have a U–Pb zircon age of 1106 ± 2 Ma (Schwartz et al. 1996); the related dolerite sills have yielded a conventional U–Pb zircon age of 1105 ± 2 Ma (Hanson et al. 1998) and a SHRIMP zircon and baddeleyite age of 1099 ± 9 Ma (Wingate 2001). In the Damara belt, rhyolites within bimodal assemblages have a U–Pb zircon age of 1094 ± 20 Ma (Hegenberger & Burger 1985). Still farther southwest, metarhyolite has a U–Pb zircon age of 1107 ± 2 Ma (Pfurr et al. 1991). In addition, the Munali Hill Granite (ca. 1090 Ma) and the Mpande Gneiss (ca. 1100 Ma) in the Zambezi Belt, which are considered to be related to the Irumide orogeny (Katongo et al. 2004), are also mainly in the field of the within-plate granites (Fig. 12).

All these rocks appear to represent a single episode of rift-related magmatism. Based on the studies of magmatism in the Ghanzi-Chobe Belt, which have geochemical signatures of post-orogenic, within-plate rhyolites, Kampunzu et al. (1998) pointed out that superimposition of the rift on an older Mesoproterozoic orogen suggests that rifting and bimodal magmatism were triggered by extensional collapse of previously thickened crust. Although rift-related basic magmatism have not been found in the Lufilian Arc, the geochronology, petrogenesis, and tectonic environment of the biotite monzogranite are consistent with the magmatism mentioned above; therefore, it can be inferred that the Lufilian Arc was also in a extensional environment of post-collision at ca. 1.10 Ga.

Conclusions

Based on the field investigations, geochemical analyses and Sr–Nd-Hf isotopic studies of the late-Mesoproterozoic granitoids in the Solwezi Dome, the following conclusions are drawn:

  1. (1)

    The Solwezi Dome complex is composed of the gneissic K-feldspar granite and the gneissic biotite monzogranite; LA-MC-ICP-MS zircon dating results show that the crystallization ages of the K-feldspar granites and the biotite monzogranites were 1178 ± 15 Ma and 1109 ± 7 Ma ~ 1105 ± 7 Ma, respectively, suggesting that they are late-Mesoproterozoic magmatism.

  2. (2)

    The petrological and geochemical data show that the K-feldspar granites are I-type, which formed by partial melting of Neoarchean basement in a slightly thickened lower crust; the biotite monzograntes are A-type, which also formed by partial melting of Neoarchean basement but at a relatively shallow depths.

  3. (3)

    The Lufilian Arc, the Zambezi Belt, the Choma-Kalomo Block, and the Ghanzi-Chobe Belt, and so on, may represent a buried southwestern continuation of the Kibaran Belt. The late-Mesoproterozoic granitoids in the Lufilian Arc are probably related to the Kibaran orogeny, which are the productions of syn-collisional and post-collisional stages, respectively.