Hong-wei Sun, Jun-ping Ren, Jie Wang, A-lei Gu, Xing-yuan Wu, Fu-qing He, Li-o Zuo,Chipilauka Mukofu, Alphet Phaskani Dokowe, Ezekiah Chikamwe, Zi-jiang Liu, Shi Xing

a School of Earth Science and Resources, China University of Geosciences (Beijing), Beijing 100083, China

b Tianjin Center, China Geological Survey, Tianjin 300170, China

c Geological Survey Department of Zambia, Lusaka P.O. Box.50135, Zambia

d No. 5 Geological Brigade, Hebei Bureau of Geology and Mineral Resources Exploration, Tangshan 063000, China

Keywords:

Paleoproterozoic

S-type granite

Bangweulu Block

Geological survey engineering

Lunte area

Northeastern Zambia

A B S T R A C T

The Paleoproterozoic tectonic evolution of the Bangweulu Block has long been controversial.Paleoproterozoic granites consisting of the basement complex of the Bangweulu Block are widely exposed in northeastern Zambia, and they are the critical media for studying the tectonic evolution of the Bangweulu Block. This study systematically investigated the petrography, zircon U-Pb chronology, and petrogeochemistry of the granitoid extensively exposed in the Lunte area, northeastern Zambia. The results show that the granitoid in the area formed during 2051±13-2009±20 Ma as a result of Paleoproterozoic magmatic events. Geochemical data show that the granites in the area mainly include syenogranites and monzogranites of high-K calc-alkaline series and are characterized by high SiO2 content (72.68%-73.78%) and K2O/Na2O ratio (1.82-2.29). The presence of garnets, the high aluminum saturation index(A/CNK is 1.13-1.21), and the 1.27%-1.95% of corundum molecules jointly indicate that granites in the Lunte area are S-type granites. Rare earth elements in all samples show a rightward inclination and noticeably negative Eu-anomalies (δEu = 0.16-0.40) and are relatively rich in light rare earth elements.Furthermore, the granites are rich in large ion lithophile elements such as Rb, Th, U, and K and are depleted in Ba, Sr, and high field strength elements such as Ta and Nb. In addition, they bear low contents of Cr (6.31×10-6-10.8×10-6), Ni (2.87×10-6-4.76×10-6), and Co (2.62×10-6-3.96×10-6). These data lead to the conclusion that the source rocks are meta-sedimentary rocks. Combining the above results and the study of regional tectonic evolution, the authors suggest that granitoid in the Lunte area were formed in a tectonic environment corresponding to the collision between the Tanzania Craton and the Bangweulu Block. The magmatic activities in this period may be related to the assembly of the Columbia supercontinent.

1. Introduction

The geodynamic evolution of the Bangweulu Block in northeastern Zambia is typically attributed to the influence of the Paleoproterozoic Ubendian Orogeny, the Mesoproterozoic Kibaran or Irumide Orogeny, and the Neoproterozoic Lufilian Orogeny. Its unique geographical location allows for better recording information about the magmatic, tectonic, and metamorphic activities related to the above-mentioned orogenic events, which is of great significance for the study of the regional tectonic evolution in Central Africa (De Waele B et al., 2006a, 2006b, 2008; Debruyne D et al., 2014; Kazimoto EO et al. 2015; Ren JP et al., 2017a). Previous studies have shown that the Bangweulu Block is mainly composed of basement complex and sedimentary cover (Brewer MS et al.,1979; De Waele B et al., 2005; Ren JP et al., 2016, 2020,2021a, 2021b). The basement of the block is mainly composed of many Paleoproterozoic granitic rocks. However,its accurate formation age, as well as whether Archean remnants exist in it, is yet to be ascertained. The early understanding of its basement age was obtained from the age data of detrital zircons in the granitic and volcanic rocks near Mansa in the southwest (Brewer MS et al., 1979; Saviaro K,1979; Schandelmeier H, 1980), the southern edge and the contact zones of the Irumide Belt (De Waele B et al., 2005,2006a, 2006b; Debruyne D et al., 2014; Ren JP et al., 2018a,2018b) and the Lufilian Belt (Rainaud CL et al., 2002; Master S et al., 2005; Armstrong RA et al., 2005; Ren JP et al.,2013), from which it was inferred that its basement was formed during 1800-2060 Ma. In recent years, some researchers have successively obtained zircon U-Pb ages of 2012-1913 Ma (Ren JP et al., 2019a, 2019b, 2019c; Gu AL et al., 2020, 2021) in the eastern Kasama area and the Kapatu area. In addition to the formation ages, the Paleoproterozoic tectonic evolution of the Bangweulu Block has also long been controversial. There are two main points of view: (1) Brewer MS et al. (1979) and Kabengele M et al. (1991) suggested that the 1850 Ma high K calc-alkaline felsic igneous rocks in the Bangweulu Block represent an active continental margin arc environment related to the Ubendian Belt. This is similar to the view of Andersen LS and Unrug R (1984) that the metamorphic volcanic rocks related to the high K calcalkaline series exposed in the northwestern part of the Bangweulu Block were formed in a subduction system. (2)Kazimoto EO et al. (2015) believed that the Bangweulu Block was in a passive continental margin environment in the early stage of the Ubendian orogeny (2640-1840 Ma). Therefore, it is necessary to further study the basement granites of the Banwewulu Block to provide new constraints for the regional tectonic evolution of central Africa during the Paleoproterozoic.

In recent years, the authors found a typical “dual structure” phenomenon consisting of basement complexes and sedimentary cover in the working area during the geological mapping in the Lunte area, northeastern Zambia. The basement complexes in the area account for approximately 90% of the working area and mainly include granitoid,providing favorable chances to study the tectonic-magmatic evolution of the Bangweulu Block during the Paleoproterozoic. This paper presents zircon U-Pb age data and major and trace element data of granitic plutons in the central part of the Bangweulu Block to explore the formation age, petrogenesis, magma source, and tectonic setting of the Bangweulu Block and, further, to assess the role of the Bangweulu Block in the assembly of the Columbia supercontinent.

2. Geological background

The Lunte area in northeastern Zambia is located in the central part of the Bangweulu Block (Figs.1a, b), which is named after the Bangweulu Lake and covers an area of approximately 150000 km2. It is adjacent to the Congo-Tanzania Craton and is surrounded by a series of Paleoproterozoic-Neoproterozoic orogenic belts (Drysdall AR et al., 1972; Andersen LS and Unrug R, 1984), including the Paleoproterozoic Ubendian-Usagaran Belt in the northeast,the Mesoproterozoic Kibara Belt, and the Neoproterozoic Lufilian Belt in the west (De Waele B et al., 2008), and the Mesoproterozoic Irumide Belt in the southeast (De Waele B et al., 2006b; Fig.1a).

Fig. 1. Tectonic map (a, modified from De Waele B et al., 2008), location map (b) and regional geological map (c) of the Lunte area in northeastern Zambia.

The Bangweulu Block is mainly composed of the metamorphic crystalline basement and sedimentary cover.Generally, the crystalline basement includes granitic plutons,supracrustal sequences, and other intrusions (Andersen LS and Unrug R, 1984; Debruyne D et al., 2014). Among them,the granitoid is widely outcropped in the block as the most significant component of the basement, accounting for more than 50% of the entire block area (De Waele B et al., 2005,2006a, 2006b; Ren JP et al., 2019a, 2019b; Zuo LB et al.,2020, 2021). The metamorphic supracrustal sequences mainly consist of schists and andesitic-rhyolitic metavolcanic.Banded schists are exposed in the eastern and northern parts of the Bangweulu Block, extending in a NW-SE or nearly EW direction (Saviaro K, 1979; Schandelmeier H, 1980). In terms of lithology, the schists mainly consist of micaceous or chloritic quartzo-feldapathic schists derived in part from acid volcanic or semipelitic - psammitic sediments (Andersen LS and Unrug R, 1984; Debruyne D et al., 2014). The metavolcanics are exposed along the present margins of the Mporokoso Basin, which is located centrally on the Bangweulu Block (Andersen LS and Unrug R, 1984). From bottom to top, the sedimentary units in the Bangweulu Block include the Paleoproterozoic Mporokoso Group, the Mesoproterozoic Kasama and Mansa River groups, the Neoproterozoic Katanga Supergroup, and the Cenozoic fluvial and lacustrine sediments. The Mporokoso Group unconformably overlies the basement and is the sedimentary cover that was deposited in the Bangweulu Block and the Irumide Belt after 2000 Ma (Daly M and Unrug R, 1982; Sun HW et al., 2019, 2021; Xing S et al., 2021).

3. Samples

In the study area, granite outcrops consist mainly of graywhite porphyritic biotite monzogranites, gray-white mediumfine-grained biotite monzogranites, gray medium-fine-grained biotite monzogranites, and gray-white fine-grained biotite syenogranites. This study conducted the whole-rock analysis and zircon U-Pb isotope dating of ubiquitous granitoid in the study area. The sampling locations are shown in Fig. 1c, and specific petrographic characteristics are described as follows.

Sample D3052 was collected from 30°26′31 "E and 10°18′35" S, and it is a gray-white porphyritic biotite monzogranite with a porphyritic texture and a massive structure (Fig. 2a). The monzogranite mainly contains plagioclase, K-feldspar, quartz, and a small amount of biotite and garnets. The phenocrysts are 1-6.5 mm in size,accounting for 58% of the entire sample. The matrix has a grain size of 0.3-0.9 mm, accounting for 42% of the total composition. The composition of the matrix is essentially the same as that of the phenocrysts. The K-feldspar is mainly composed of microcline with well-developed crossed twinning, with a content of 36%. The plagioclase is semieuhedral to anhedral and tabular in shape, with welldeveloped polysynthetic twinning and a content of 36%. The quartz is anhedral granular, colorless, and transparent, with well-developed intragranular cracks and a content of 22%.The biotite is semi-euhedral, flaky, and scaly, showing bright interference color and with a content of 5%. The garnets are anhedral granular, with well-developed intragranular cracks and a content of 1% (Fig. 2b).

Fig. 2. Hand specimen (a-D3052, c-D4353, e-YPM078, g-YPM075) and photomicrographs (b-D3052, d-D4353, f-YPM078, h-YPM075;crossed nicols) of granites from the Lunte area in northeastern Zambia. Q-quartz, Pl-plagioclase, Bt-biotite, Pth-perthite, Kfs-K-feldspar,Phl-phlogopite, Mc-microcline, Grt-garnet.

Sample D4353 was collected from 30°14′55 "E and 10°22′40" S, and it is a gray-white medium-fine-grained biotite monzogranite with a granular texture and a massive structure (Fig. 2c). The sample is mainly composed of K-feldspar, plagioclase, quartz, and a small amount of biotite and muscovite. The K-feldspar mainly includes perthite. It has a well-developed striped texture and is semi-euhedral to anhedral and platy in shape, with a particle size of 0.5-4.8 mm and a content of 44%. The plagioclase is semi-euhedral to anhedral and platy in shape, with developed polysynthetic twinning. It has a particle size of 0.3-3.5 mm and a content of 28%. The quartz is anhedral granular, with developed intragranular cracks and showing parallel extinction in the whole rock. It has a particle size of 0.5-3.0 mm and a content of 21%. The biotite is light brown - dark brown semi-euhedral fine scales and shows strong pleochroism, with a content of 7% (Fig. 2d).

Sample YPM078 was collected from 30°7′27 "E and 10°12′43" S, and it is a gray medium-fine-grained biotite monzogranite with a granular texture and a massive structure(Fig. 2e). The monzogranite mainly consists of K-feldspar,plagioclase, quartz, and a small amount of biotite, with silicification alteration being observed. The K-feldspar is dominated by semi-euhedral to anhedral and platy perthite with a well-developed striped texture. It has a particle size of 0.4-5.2 mm and a content of 36%. The plagioclase is semieuhedral and platy in shape and shows developed polysynthetic twinning. It has a particle size of 0.3-4.5 mm and a content of 32%. The greasy quartz is distributed in anhedral granular aggregates, with a content of 24%. The light brown-dark brown biotite is semi-euhedral and fine-scale in shape. It shows strong polychromism, with a particle size of 0.25-0.98 mm and a content of 8% (Fig. 2f).

Sample YPM075 was collected from 30°6′44 "E and 10°10′2" S, and it is a gray-white fine-grained syenogranite with a fine-grained texture and a massive structure (Fig. 2g).The syenogranite is mainly composed of K-feldspar,plagioclase, quartz, and a small amount of biotite. The K-feldspar is dominated by perthite and is semi-euhedral to anhedral and platy in shape, with a particle size of 1.02-2.85 mm. The K-feldspar is highly clayey overall, with a common striped texture and a content of 55%. The plagioclase is semieuhedral and platy in shape and shows developed polysynthetic twinning, with a particle size of 1.00-1.78 mm and a content of 22%. The greasy quartz is distributed in anhedral granular aggregates. It shows parallel extinction,with a content of 20%. The biotite is semi-euhedral scaly and light brown-dark brown. It shows strong pleochroism, with a content of 3% (Fig. 2h).

4. Analytical methods

Samples for Zircon U-Pb dating first underwent crushing,followed by panning and gravity, and magnetic sorting.Afterward, zircons were selected under a binocular microscope. These procedures were performed at the laboratory of Langfang Yuneng Rock Mineral Separation Technology Service Co. Ltd., Hebei Province, China. The target preparation and cathodoluminescence (CL) imaging and observation of zircons were performed at Beijing Gaonian Pilot Technology Co., Ltd. LA-ICP-MS Zircon U-Pb dating was performed at the laboratory of Tianjin Center, China Geological Survey. The instruments used include a 193 nm FXArF excimer laser ablation system from the New Wave Research and a Neptune multi-collector inductively-coupled plasma mass spectrometer (MC-ICP-MS) from Thermo Fisher Scientific. During the experiment, helium was used as the carrier gas for laser ablation, and the laser spot diameter was set at approximately 35 μm. Specific operation conditions and detailed analysis process of relevant instruments were stated in Li HK et al. (2009). In the analysis process, isotope fractionation correction was conducted using Australian zircon standard GJ-1 as external reference standards. After the experiment, the ICPMSDataCal software was used to process data offline (Liu YS et al., 2010), and Ludwing KR (2003)was followed to plot U-Pb Concordia diagrams and calculate the weighted average U-Pb age.

The whole-rock analyses of major, trace, and rare earth elements were performed at the Element Analysis Laboratory of Tianjin Center, China Geological Survey. First, the sample powder was melted into a glass sheet, and then major elements were determined using an XRF 1500 X-ray fluorescence spectrometer. The analysis precision was better than 1%. FeO was dissolved in a mixed hydrofluoric and sulfuric acid dissolution and was determined using the potassium dichromate volumetric titration method, and the analysis accuracy was better than 2%. For the analysis of trace elements and rare earth elements (REEs), the sample solution was fully dissolved using high-purity HNO3and high-purity HF and then was diluted with 1% HNO3. Finally, trace elements and REEs in the sample solution were determined using a double-focusing sector field inductively coupled plasma mass spectrometer (ICP-MS) from the Finnigan MAT company, and the analysis precision was better than 5%.

5. Results

5.1. Zircon U-Pb geochronology

In this study, zircon U-Pb dating was conducted on four granite samples from the Lunte area in northeastern Zambia.The CL images of some representative zircons are shown in Fig. 3, the zircon U-Pb concordia diagrams are illustrated in Fig. 4, and the analytical results are listed in Table 1.

Fig. 3. Cathodoluminescence (CL) images and analytical spots of zircons from granites in the Lunte area in northeastern Zambia.

Fig. 4. Zircon LA-MC-ICP-MS U-Pb concordia diagrams for granites from the Lunte area in northeastern Zambia.

Table 1. LA-MC-ICP-MS zircons U-Pb date of granites from the Lunte area in northeastern Zambia,

Table 1. (Continued)

In sample D3052, most zircons are columnar, with a particle size of 100-150 μm and a length/width ratio of 3∶2-3∶1. CL images show that most zircons have developed banded oscillatory zoning. The Th/U ratio of 30 analyzed spots varies from 0.14 to 1.19 and is high (＞ 0.4)mostly, indicating magmatic zircons (Belousova EA et al.,2002). The Th/U ratio of some zircons ranges from 0.1 to 0.4,which may reflect incomplete metamorphism and recrystallization (Wu YB and Zheng YF, 2004) or reformation induced by later geological events. Most of the analyzed spots deviate from the concordia curve and form inconsistent lines,and the upper intercept age is 2051±13 Ma.

In sample D4353, the zircon crystals are short columnar,with a particle size of approximately 60-140 μm and a length/width ratio of approximately 2∶1-3∶1. Zircon crystals exhibit banded oscillatory zoning and rarely exhibit core-rim textures. The Th/U ratio of 23 measuring spots is 0.01-2.01 (＞ 0.4 mostly), which is consistent with typical magmatic zircons. The weighted average207Pb/206Pb age is 2034±14 Ma (MSWD=0.05,n=8).

In sample YPM075, the zircon crystals are long columnar,with a particle size of 100-210 μm and an aspect ratio of approximately 2∶1-4∶1. CL images show that most zircons have zoning textures and a few zircons have core-rim textures. The Th/U ratio of 20 zircon core measurement spots is 0.11-2.16 and is high (＞ 0.4) mostly, reflecting the characteristics of magmatic zircons. The Th/U ratio of a small amount of zircon ranges from 0.1 to 0.4, which may reflect incomplete metamorphism and recrystallization (Wu YB and Zheng YF, 2004) or reformation induced by later geological events. Most of the analyzed spots deviate from the concordia curve to form inconsistent lines, and the upper intercept age is 2009±20 Ma.

In sample YPM078, the zircon crystals are short columnar with core-rim textures. The grain size and length/width ratio of the zircons are 60-110 μm and approximately 1∶1-2∶1,respectively, and zoning textures can be observed. The Th/U ratio of 27 core measuring spots is 0.10-2.71, which is consistent with that of typical magmatic zircons. Most of the analyzed spots deviate from the concordia curve to form inconsistent lines, and the upper intercept age is 2036±19 Ma.

5.2. Whole-rock geochemistry

The geochemical analysis results of granite samples from the Lunte area are given in Table 2. According to this table and the geochemical discrimination diagrams, granitoid do not show notable differences in geochemical characteristics,and different terranes show relatively consistent geochemical distribution characteristics. This indicates that they may have similar source areas and evolution mechanisms of magmas.

Table 2. Major elements (%) and trace elements (10-6) compositions of granites from the Lunte area in northeastern Zambia.

5.2.1. Major elements

The major element analysis results show that the SiO2content varies from 72.68% to 73.78% in different granitoid,indicating the characteristics of typical acidic intrusive rocks.The samples from the Lunte area are relatively rich in potassium (K2O = 5.14%-5.60%) and have high ALK [total alkali content (K2O+Na2O): 7.97%-8.27%], and their K2O/Na2O ratio is 1.82-2.29. In the TAS diagram of rock classification, all the sample spots fall into the granite zone(Fig. 5a). However, in the SiO2-K2O diagram, the sample spots fall into the high-K calc-alkaline to shoshonite zone,reflecting the high-K characteristics of rocks (Fig. 5b). The total iron content is 1.55%-2.50% in TFeO and 0.44%-0.66% in MgO. The Al2O3content is 13.45%-14.76%, the Al saturation index A/CNK is 1.13-1.21, and the CIPW standard mineral calculation yielded 1.27%-1.95% of corundum molecules, exhibiting the characteristics of strongly peraluminous granites. In the A/NK-A/CNK diagram, all samples fall into the peraluminous zone representing S-type granites (Fig. 6). The results are slightly different from the samples (monzogranite and syenogranite) from the Kapatu area (Fig. 6).

Fig. 5. TAS diagram (a, after Irvine TH and Baragar WR, 1971) and K2O-SiO2 diagram (b, after Rollinson HR, 1993) of granites from the Lunte area in northeastern Zambia. 1-foidolite, 2-foid syenite, 3-foid monzosyenite, 4-syenite, 5-foid monzodiorite, 6-monzonite, 7-foid gabbro, 8-monzogabbro, 9-monzodiorite, 10-quartz monzonite, 11-granite, 12-peridotgabbro, 13-gabbro, 14-gabbroic diorite, 15-diorite,16-granodiorite.

Fig. 6. A/NK vs. A/CNK diagram of granites from the Lunte area in northeastern Zambia (after Maniar PD and Piccoli PM, 1989; data of the Kapatu area from Gu AL et al., 2021).

5.2.2. Trace elements

The total REE content in all samples ranges from 212.05×10-6to 611.23×10-6. The fractionation of light and heavy REEs is obvious, showing a rightward inclination [Fig.7a; (La/Yb) N = 5.95-33.57]. The samples are relatively rich in light rare earth elements (LREEs) and relatively depleted in heavy rare earth elements (HREEs). Meanwhile, they show noticeable negative Eu anomalies (Eu = 0.16-0.40),suggesting that strong plagioclase segregation and crystallization may have occurred during diagenesis or that plagioclase existed in the residual phase of protolith melting.The primitive mantle-normalized trace element spidergram(Fig. 7b) shows that the samples are relatively enriched in large ion lithophilic elements (LILEs) such as Rb, Th, U, and K, and are relatively depleted in Ba, Sr, and high field strength elements (HFSE) such as Ta, Nb, and Ti.

Fig. 7. Chondrite-normalized REE element patterns (a) and primitive mantle-normalized trace element spidergram (b) of granites from the Lunte area in northeastern Zambia (chondrite and primitive mantle normalizing values from Sun SS and McDonough WF, 1989, crust contents after Taylor SR and McLennan SM, 1985).

Overall, different granite samples have the same distribution patterns of REEs and trace elements, indicating that they may have the same magma source areas. The comparison and analysis with the standard values of REEs and trace elements in the crust show that the granite samples are more similar to the upper crust, suggesting that its magma source areas may mainly include the upper crust.

6. Discussion

6.1. Emplacement age

To date, there are only a few geochronological data of the Bangweulu Block, and those early age data suffered low accuracy. The earliest age results obtained are the whole rock Rb-Sr isochron ages of granites and volcanic rocks in the Mansa area in the western part of the Bangweulu Block,which is 1833±18 Ma and 1812±22 Ma, respectively (Brewer MS et al., 1979). Subsequently, the whole rock Rb-Sr isochron age of Kate granite (1839±80 Ma; Schandelmeier H,1980) was reported from the northern part of the Bangweulu Block. In recent years, Ren JP et al. (2019a, b) and Zuo LB et al. (2020) have reported new ages of the western Kasama granitoid (1964±8 Ma-2011±20 Ma) obtained using the LAMC-ICP-MS zircon U-Pb dating method, and they are similar to the formation age of the granites in the Kapatu area(2012±11 Ma -1970±20 Ma; Gu AL et al., 2021). However,there is nearly no geochronological data of basement granitoid in the Lunte area. The geochronological data from this study show that the emplacement age of the granitoid in the Lunte area varies between 2009±20 Ma and 2051±13 Ma, which is also consistent with the basement formation age obtained by previous studies and belongs to the Early Paleoproterozoic.

Based on the regional tectonic evolution, De Waele B et al. (2006a) divided the magmatic-tectonic events in the Bangweulu Block into six stages, namely the Pre-Usagaran period (＞ 2100 Ma), Usagaran period (2050-1950 Ma),Ubendian period (1950-1880 Ma), Post-Ubendian period(1880-1650 Ma), Lukamfwa period (1650-1550 Ma), and Irumide period (1080-850 Ma). The zircon ages of four samples range from 2051±13 Ma to 2009±20 Ma, belonging to the Usagaran period. Therefore, the authors believe that the most robust magmatism in the Lunte area occurred during the Usagaran period, which is consistent with the previous conclusion drawn from the statistical data on the distribution patterns of the detrital zircon U-Pb ages of the sedimentary cover (the Mporokso Group; De Waele B et al., 2006a,2006b; Sun HW et al., 2019, 2021). In addition, the presence of some zircon ages of the Usagaran period implies that the crustal growth of the Bangweulu Block must have occurred before ca. 2000 Ma since it was followed by the deposition of the sedimentary sequences—the Mporokoso Group (2000 Ma). This is consistent with the conclusion of this study that the crystalline basement age is older than 2000 Ma.

6.2. Petrogenesis

Granites can be divided into types I, S, A, and M according to their genetic types (Chappell BW and White AJR, 1974). The granitoid in the Lunte area are characterized by high silicon (SiO2= 72.68%-73.78%), high aluminum(Al2O3= 13.45%-14.76%), rich potassium (K2O/Na2O = 1.82-2.29), low phosphorus (P2O5= 0.11%-0.12%), and depleted magnesium (MgO=0.44%-0.66%), with an aluminum saturation index (A/CNK) of 1.13-1.21 and a CIPW corundum standard molecular index of greater than 1% (1.27%-1.95%). The trace element analysis shows that there are relatively low contents of Cr (6.31×10-6-10.8×10-6), Ni (2.87×10-6-4.76×10-6), and Co (2.62×10-6-3.96×10-6), exhibiting the geochemical characteristics of crustal sources.

Generally, A-type granites have a higher 10000×Ga/Al ratio and Nb content than I- and S-type granites (Whalen JB et al., 1987). Granites from the Lunte area have a low 10000*Ga/Al ratio and Nb content. In the A-type granite discriminant diagram, all the samples fall into the zone representing I+S-type granites (consistent with the samples from the Kapatu area) and are notably different from the A-type granites (Fig. 8a). In addition, some studies showed that Th and Y contents, as well as their correlation with Rb, are important reference indices for distinguishing S-type granites from I-type granites (Chappell BW, 1999). In the genetic-type diagram of granites, all samples from the Lunte area fall into the zone representing S-type granites (Fig. 8b). Typically, Th,Y, and Rb of S-type granites are negatively correlated, while Th, Y, and Rb of I-type granites are positively correlated. The Th, Y, and Rb of granites in the Lunte area are notably negatively correlated, suggesting the affinity with S-type granites (Figs. 8c, d). According to mineralogical studies,garnets, muscovite, and other aluminum-rich minerals are common in granitoid in the Lunte area, which is consistent with the characteristics of S-type granites. Based on the above analysis, similar to the granites from the Kapatu area, the Paleoproterozoic granites that are widely developed in the Lunte area most likely belong to peraluminous S-type granites.

Fig. 8. 10000×Ga/Al vs. Nb (a, modified from Whalen JB et al., 1987), Al2O3-(K2O+Na2O) vs. CaO-TFeO+MgO (b, modified from White AKR and Chappell BW, 1977), Rb vs. Th (c) and Rb-Y (d; c and d after Chappell BW and White AJR, 1999) diagrams of granites from the Lunte area in northeastern Zambia (data of the Kapatu from Gu AL et al., 2021).

6.3. Magma source

S-type granites are commonly interpreted as magmas derived from the partial melting of metasedimentary crust(Sylvester PJ, 1998). Experimental petrological studies show that the CaO/Na2O ratio of S-type granites formed from the partial melting of different source areas notably differs(Sylvester PJ, 1998). It is relatively low (＜ 0.3) for S-type granites formed from the remelting of argillaceous rocks but is relatively high (＞ 0.3) for S-type granites formed from the remelting of sandy rocks. The CaO/Na2O ratio of granite samples in the study area is 0.26-0.38, with an average of 0.32, from which no obvious indication of source areas can be deduced. However, as shown in the discrimination diagram of granite source areas (Fig. 9a), most of the granite samples fall into the partially melting zone of metamorphic argillaceous rocks. In the mixed simulation results of argillaceous rock and basalt melting, the samples fall into the zone close to argillaceous rock melting, and all samples show the characteristics of the clay-rich source zone (Fig. 9b). In addition, the characteristics of REEs and trace elements are very similar to those of the upper crust. Therefore,comprehensive analysis shows that the source areas of granite magmas in the Lunte area are mainly formed by the partial melting of crustal argillaceous rocks. However, some samples in the Kapatu area fell into the melting zone of metamorphic gray sandstones, indicating that the Bangweulu Block has diverse granite source areas.

Fig. 9. A/MF vs. C/MF diagram (a) and Rb/Ba vs. Rb/Sr diagram (b) of granites from the Lunte area in northeastern Zambia (a, after Altherr R et al., 2000; b, after Sylvester PJ, 1998; data of the Kapatu area are quoted from Gu AL et al., 2021).

6.4. Tectonic setting

The collision between the Bangweulu Block and the Tanzania Craton during the Paleoproterozoic led to the formation of the Usagaran-Ubendian orogenic Belt, but the regional tectonic evolution is still controversial. Brewer MS et al. (1979) claimed that the high-K calc-alkaline felsic igneous rocks developing at 1850 Ma in the Bangweulu Block represent an active continental margin arc environment related to the Ubendian orogenic belt. Andersen LS and Unrug R(1984) believed that the metamorphic volcanic rocks related to the high-K calc-alkaline series in the northwestern Bangweulu Block were formed in a subduction system. De Waele B et al. (2006) and Sun HW et al. (2019) argued that the northeastern Bangweulu Block was in a passive continental margin environment during the Proterozoic.Kazimoto EO et al. (2015) suggested that many collisions occurred between the Bangweulu Block and the Tanzania Craton during the Paleoproterozoic and that the entire area was in a continent-continent collision environment. Ren JP et al. (2019a) suggested that the Bangweulu Block might have been formed in an Andean-type active continental margin environment.

The Lunte area is located in the central part of the Bangweulu Block. The granitic intrusions widely exposed in the area are the major components of the basement complex in the Bangweulu Block. This study shows that the granites in the study area likely formed between 2051±13 Ma and 2009±20 Ma during the Paleoproterozoic and constitute a peraluminous S-type granite as a whole. Generally, the formation of S-type granites is usually related to collisional orogeny and they may be formed in syn-collision and postcollision environments (Sylvester PJ, 1998; Barbarin B,1999). According to the source analysis, the granites in the Lunte area resulted from the partial melting of crustal argillaceous rocks, while the melting of crust-derived metamorphic clastic rocks is one of the main processes for the generation of granitoid during the collisional orogeny (Zhou TF et al., 2010). As shown in the R1 vs. R2 tectonic environment diagram, all the samples fall into the zone of syn-collisional granites derived from crust melting (Fig. 10a).Meanwhile, in the (Y+Nb)vs.Rb diagram, all the samples fall in the syn-collisional or post-collisional granite zone (Fig.10b), and data from the Kapatu area show similar characteristics (Gu AL et al., 2021). From a regional perspective, Lenoir JL et al. (1994) believed that the collision between the Tanzania Craton and the Bangweulu Block at 2100-2025 Ma corresponds to the early collisional orogenic event in the Ubendian Belt. Meanwhile, the formation age of eclogites in the Paleoproterozoic Usagaran Belt in Tanzania is 2010-1996 Ma, which also reflects the collisional orogeny of about 2.0 Ga (Möller A et al., 1995). Boniface N and Schenk V (2012) suggested that the Usagaran Belt was in a continental collision stage during 2.05-1.93 Ga.

Fig. 10. R1 vs. R2 diagram (a) and Rb vs. (Y+Nb) diagram (b) of granites from the Lunte area in northeastern Zambia (a, after Batchelor RA and Bowden P, 1985; b, after Pearce JA et al., 1984; data of the Kapatu area from Gu AL et al., 2021).

The comprehensive study indicates that the granites in the Lunte area were likely formed in the collision environment between the Bangweulu Block and Tanzania Craton.Combining this with the emplacement age of granites in the area, the authors suggest that the granites in the Lunte area are mainly the product of the Usagaran orogeny, which is also the main formation period of the Bangweulu Block.

During the Paleoproterozoic-Mesoproterozoic Columbia supercontinent cycle, many large-scale collisional orogenic events occurred in the world, including the Eburnean orogeny(2.10-2.00 Ga), the Limpopo orogeny (2.00-1.90 Ga), the Capricorn orogeny (2.00-1.85 Ga), and the intracontinental collisional orogeny of the North China Craton (1.85 Ga;Cooper MR, 1990; Zhao GC et al., 2002; Rogers JJW and Santosh M, 2002; Zhai MG, 2010; Smit CA et al., 2014; Sun K et al., 2018). This study shows that the extensively developing granites in the Lunte area, northeastern Zambia were formed during 2.00-2.05 Ga under a tectonic environment of syn-collision or post-collision.

Therefore, the authors infer that the Paleoproterozoic magmatic event in the Bangweulu Block may be the response of the block to the Columbia supercontinent convergence.Further research on the convergence mechanisms is required.

7. Conclusions

(i) The granitoid in the Lunte area, northeastern Zambia were formed during the Paleoproterozoic, with an emplacement age of 2051±13-2009±20 Ma, which is consistent with the Usagaran tectonic event.

(ii) The whole-rock geochemical characteristics show that the granitoid in the Lunte area, northeastern Zambia belongs to strongly peraluminous S-type granites, and the distribution characteristics of trace elements and REEs are similar to those of the upper crust. It is suggested that the source rocks of the granites originated from the partial melting of crustal argillaceous rocks.

(iii) The granitoid in the study area is the product of the collisional environment between the Bangweulu Block and the Tanzania Craton. Based on the history of global tectonic evolution, the authors conclude that the Bangweulu Block may be a part of the Columbia supercontinent, and the magmatism in the study area might belong to the geological response of the Bangweulu Block to the collision orogeny of the Columbia supercontinent.

CRediT authorship contribution statement

Hong-wei Sun conceived of the presented idea. Hong-wei Sun wrote the manuscript with support from Jun-ping Ren.Jun-ping Ren improved the theory. All authors discussed the results and contributed to the final manuscript.

Declaration of competing interest

The authors declare no conflicts of interest.

Acknowledgment

This study was jointly funded by projects of the Ministry of Commerce ([2015] 352 and [2012]558) and the projects of the China Geological Survey (DD20201150 and 1212011220910). Thoughtful comments by three anonymous reviewers and editors have considerably improved the manuscript.