A₂-Type Granites from the Bastar Craton, South-Central India, and Their Implication in Archean-Paleoproterozoic Tectonics in Indian Peninsula

There is a consensus that subduction zone tectonics in the Earth was initiated during the Mesoarchean time [1, 2]. Subsequently, the Neoarchean and the Paleoproterozoic periods experienced extensive subduction and rift-related magmatism leading to the cratonization of Archean blocks [3–5]. Jayananda et al. [4, 6], Condie and O’Neil [7], Condie [8], Condie et al. [9] correlate the late Archean–Paleoproterozoic transition with the extraction of continental crust from the mantle, with a significant increase in large-ion lithophile and high-field strength elements and a decrease in Sr in the continental crust that reflect in the change in magma chemistry from tonalite-trondhjemite-granodiorite to calc-alkaline–alkaline granite.

There is a general agreement that nearly 70–75% of the present continental crust was formed during the late Archean-Paleoproteozoic transition in which granites are the most abundant lithodemic units. The primary minerals in granites, i.e., K-feldspar, quartz, and plagioclase feldspar with minor ferromagnesian minerals, can be crystallized in numerous tectonic processes; the chemical processes and genesis of granites are essential to understand the tectono-thermal evolution of crustal blocks during Archean–Plaeoproterozoic transition.

Chappel and White [10] introduce the first modern geochemical scheme for granite petrogenesis based on the phase-assemblage and geochemistry. The I-type granites are metaluminous to peraluminous, with high Na₂O and variable SiO₂ content (56-77 wt.%) and are inferred to have formed from an igneous source. The second type of granite is S-type granite, strongly peraluminous in chemistry, enriched in K₂O compared to Na₂O, and has higher SiO₂ content. These granites form via melting metasediments in collisional tectonic settings.

Earlier authors, e.g., Whalen et al. [11], introduce the term A-type granite to recognize another type of granite that is distinctly potassic and has a high Fe number and has high Zr and other high field strength element content. The tectonic setting of A-type granite emplacement is debated. Initially, Loiselle and Wones [12] argued that A-type granites are mainly emplaced in anorogenic settings, although most workers [11, 13, 14] consider such granitoids related to magma underplating, melting, and mixing continental and mafic sources.

Eby [13, 15] classified A-type granites into two subgroups based on its trace element abundances. The A₁ subgroup includes felsic rocks emplaced in the continental rifts and oceanic island arcs and is suggested to be derived from an oceanic island arc basalt in a rift setting. In contrast, the A₂ subgroup is formed by various mechanisms that include partial melting of island arc-continental margin basalts or partial melting of restite igneous rock or granulites. Geochemically, the A₁ subgroup exhibit a lower Y/Nb (<1.2) content compared to higher Y/Nb (>1.2) content in A₂ subgroup.

Dall’Agnol et al. [16] suggest eight types of A-type granitoids based on the Ferroan versus Magnesian (Fe$∗$ scheme), Modified Alkali Lime Index (MALI), Alumina Saturation Index (ASI), and Alkalinity-Index (AI) classification scheme [17]. As the A₂ granites can develop in varied tectonic settings, an age-integrated geochemical study of the same provides essential information on the magmatic processes that contribute to the geochemical evolution of the upper continental crust.

In this contribution, we present the results of detailed geochemical, phase petrological, and U-Pb zircon data from relatively undeformed A₂-type EBC granites. The tectonic models presented in this article imply that the Eastern Bastar Craton granites are developed by melting and assimilating K-rich crustal material and underlying mantle during the stabilization of extended Ur at around ~2500 Ma. The results are critical in understanding the plate-tectonic processes and crustal growth during the Neoarchean-Paleoproterozoic transition.

In present-day India, the Central Indian Tectonic Zone (CITZ) bisects the Indian peninsula into North-Indian-Continental-Block (NICB) and South-Indian-Continental-Block (SICB), respectively [18]. The Aravalli – Bundelkhand Craton constitutes the North-Indian-Continental-Block, Singbhum Craton, Dharwar, and Bastar Craton the South-Indian-Continental–Block [19, 20]. The oldest rocks of these cratons preserve evidence for Archean-Paleoproterozoic evolution of the continental crust and are considered crucial crustal components in Archean-Paleoproterozoic supercontinent assembly and disintegration processes [20–27].

Located in the heart of the Indian subcontinent, the Bastar Craton extends over about two million square kilometers of area. The Eastern Ghats Belt (EGB) and the Central Indian Tectonic Zone demarcate its eastern and north-western boundaries. The Mahanadi graben and the Godavari graben mark the north-eastern and the southern extremity of the craton (Figure 1(a)).

Compared to the other well-studied Archean Cratons across the Earth (Isua Greenstone Belt, Greenland: [38, 39]; Pilbara Craton: [40–42], Kaapvaal Craton: [43, 44], China [45] Dharwar Craton: [46–48]; Singbhum Craton: [49–53]; Bundelkhand Craton: [26, 54], among many more), the geological history of the Bastar Craton is loosely constrained. Ramakrishnan and Vaidyanathan [19] compiled geological and geochronological information for Bastar Craton. Mohanty [20, 55] correlates BC’s geochronological and tectono-thermal information with supercontinent assembly and fragmentation processes. However, a comparative evaluation of the study area implies that most data is available from the western and southern parts of the craton. In contrast, the geochemical and geochronological data from the eastern part are meager due to the influence of ultraleft political outfits, which disrupt comprehensive fieldwork.

Nevertheless, the data available from the western and southern part of the craton implies that Tonalite-Trondhjemite-Granodiorite-gneisses (TTG) and granites are the major lithodemic unit in the craton [19, 37, 55–59]. Also, the TTGs host a thick succession of supracrustal and granulite belts [19, 60], a set of mafic dike swarms [61, 62], and the Proterozoic Purana basin [63–66].

Variably deformed tonalite–trondhjemite–granodiorite (TTG), transitional TTG (t-TTG) and undeformed granitoids constitute the granite-granodiorite lithodemic units in the Bastar Craton [57]. The older and multiply deformed felsic crusts were dated as old as 3.6 Ga (U-Pb Zircon) by Ghosh [67] and Sarkar et al. [37]. Rajesh et al. [30] reported the occurrence of 3.6-3.5 Ga undeformed granitoids from BC. The ~3.6 Ga crust was intruded by ~2.5-2.2 Ga undeformed granites in and around the Dongargarh, Kanker, Malanjkhand, and the Sukma area [25, 29, 56, 58, 68–70]. A series of variably metamorphosed siliciclastic and acid volcanics called the Bailadila group, the Bengpal group, the Sukma group, the Sakoli group, the Sausar group, and the Dongargarh supergroup overlay older gneisses [58, 70]. The only Neoarchaean greenstone belt is located in the north-eastern part of the craton [71]. The younger supracrustal rocks include undeformed and weakly metamorphosed Meso-Neoproterozoic sedimentary sequences deposited in intracratonic basins [72–74]. Many mafic dyke swarms are intrusive into the gneisses, granitoids, and supracrustal rock [61, 75]. Based on the geochemical and geochronological studies from the western part of BC, Santosh et al. [59] divided BC into two distinct tectonic blocks, namely, the Western Bastar Craton (WBC) and the Eastern Bastar Craton (EBC) that stitched along the N-S trending Central Bastar Orogen (CBO) (Figure 1(a)).

Santosh et al. [59] argued that the southern part hosts the oldest crustal components, unlike the western part of the craton (WBC) and orogenic belt in the central part (CBO) preserve Neoarchean-Paleoproterozoic thermal events. Santosh et al. [59] obtained Paleoarchean (3250–3150 Ma) Nd model ages of mafic enclaves from granites exposed in the western part of the craton [76]. Several authors suggest that granites in the WBC and CBO were emplaced within a time frame between 2544 and 2496 Ma [58, 59, 70]. An anorgenic granite, commonly known as Mul granite, intruded the WBC TTG at ~1600 Ma [77].

Mohanty [20] studied the paleomagnetic data of mafic dykes from the Dharwar-Bastar craton and the Yilgarn craton of Western Australia. Similar lithodemic assemblage, the timing of deformation and metamorphism, when combined with paleomagnetic data, implies a coherent landmass called SIWA comprising the Dharwar-Bastar Craton and the Yilgarn Craton during the late-Archean time [20, 69]. A school of authors ([59, 68, 70]) correlates the emplacement of meta-volcanic and mafic rocks in the WBC’s eastern margin with a westward subduction zone tectonic setting (Figure 13 of Manikambya et al. [70]). Nasipuri et al. [78] report late Neoproterozoic (~500 Ma) anatexis and migmatization of EBC granites related to Pan-African Gondwanaland assembly. However, none of the authors correlate studied EBC granite’s geochemical and geochronological data that can be utilized to develop a tectonic model during the Archean-Paleoproterozoic transition.

Representative back-scattered electron images of silicate phases were acquired using a ZEISS Gemini Scanning Electron Microscope (SEM) with an accelerating voltage of 20 kV with 400-478 picoampere (pA) current at the Indian Institute of Science Education and Research Bhopal. The constituent silicate minerals of granites were analyzed using a 5 WDS CAMECA SX-100 electron probe microanalyzer (EPMA) at the Indian Institute of Technology (IIT) Dhanbad. The accelerating voltage was set to15 kV in a LaB₆ cathode with a beam current of 15 nA and beam diameter between 1 and 5 μm for silicate analysis. The following natural mineral standards are chosen for analysis: MgO for Mg, apatite for Ca, haematite for Fe, rhodonite for Mn, orthoclase for K, rutile for Ti, albite for Na and Si, and almandine for Al.

The current set of sample’s major and trace element content was determined using an X-ray fluorescence spectrometer and an inductively coupled plasma mass spectrometer at CSIR-National Geophysical Research Institute Hyderabad. The analytical protocol, detection limits, and associated errors are after Satyanarayanan et al. [79] and Krishna et al. [80], for ICP-MS and XRF, respectively.

The zircon geochronology from the granites was carried out in the Beijing SHRIMP Centre, the Chinese Academy of Geological Sciences. Zircon grains were separated by crushing and sieving of unweathered granite samples. The sieved powder <60 mesh was panned to separate the heavy minerals. A conventional magnetic separator is used for further purification of the heavy mineral concentrate. Finally, zircons were handpicked under a binocular microscope and mounted in 25 mm diameter epoxy resin disks with the zircon standards TEMORA-2. About -third of the grains were polished to reveal their internal zoning. The CL imaging was obtained using a Carl Zeiss Merlin Compact Scanning Electron Microscope with a GATAN Mono CL4 detector (accelerating voltage 10 kV, beam current 5 nA, and acquisition time 60 μs). The Zircon U-Th-Pb isotopic analyses were performed on a SHRIMP II instrument in the Beijing SHRIMP Centre. The intensity of the primary O₂ ion beam was ~3 nA, 25 μm diameter spot. Before analysis, each grain was rasterized by a primary beam for about 2.5 minutes to remove common Pb. Five scans were made for each spot. The standard zircon M257 (⁠$U=840ppm$⁠, Nasdala et al. [81]), and TEMORA 2 (⁠$P206b/U238$ age =416.8 Ma, Black et al. [82], were used for the U content and $P206b/U238$ calibration, respectively. Peak resolution was about 5000 at 1% peak height, and sensitivity was about 20 cps/ppm/nA Pb on the standard. The standard TEMORA 2 was analyzed after 3 to 4 sample analysis samples to check analytical consistency. Common Pb corrections were applied using the measured ²⁰⁴Pb correction method.

The computer program, A-X (http://ccp14.cryst.bbk.ac.uk/ccp/web-mirrors/crush/astaff/holland/ax.html), is used to obtain structural formulae of minerals. The mineral abbreviations are after [83]. The chemical analysis and structural formulae of the constituent minerals are given in Table 1.

The software package GCDKIT [84] is used to plot the geochemical data. The complete range of primary, trace, and rare-earth-element content of 16 samples is given in Table 2.

As garnet is not present in the studied samples, the commonly cited Fe-Mg exchange and net transfer reactions could not be used to determine the P-T conditions for magma emplacement and magma generation. Still, the temperature for magma crystallization is estimated using empirical formulations that involve zircon solubility in silicate melt [85, 86].

The Fortran program package, PERPLE_X [87], and internally consistent thermodynamic dataset of minerals [88] are used for P-T pseudosection modeling to constrain magmatic assemblages P-T stability. The activity composition relationship for plagioclase is after Newton et al. [89]. The biotite activity composition relation was taken from White et al. [90]. Quartz is set to be saturated. Fe is assumed to be Fe²⁺ in all the modeling, leading to minimum equilibrium temperature estimation. Since biotite and K-feldspar are present in the studied sample, the amount of H₂O is adjusted from biotite’s modal abundances in the present set of samples. In the first step, fifteen petrological thin sections of representative samples were scanned in a scanning electron microscope to estimate the relative proportions of biotite and K-feldspar. As biotite and K-feldspar are only phases to accommodate K₂O as a significant component, and plagioclase feldspar is likely to contain very little K₂O (<1 wt.%), the total K₂O content (wt.%) obtained from whole rock analysis is allocated to biotite and K-feldspar based on their relative modal proportions. As the amount of K and Na in the I site in the biotite structure is half of the maximum amount of fluid (OH, Cl, F) present in the ideal biotite stoichiometry, the reallocated K₂O for biotite is doubled to fix the minimum amount of H2O required to stabilize biotite with predicted abundances similar to the observed modal abundance. The Albite mole% of plagioclase feldspar $Na/Na+Ca×100$ and Fe mole% of biotite $Fe/Fe+Mg×100$ were computed to constrain the P-T conditions.

The computer program ISOPLOT [91] is used for data processing. The representative geochemical data with analytical error (⁠$1σ$⁠) is shown in Table 3.

Adjacent to the Bhawanipatna–Ampani section, Odisha, India, the Eastern Ghats Belt granulites are juxtaposed with Archean-Paleoproterozoic cratonic granites and Neoproterozoic migmatites. Nasipuri et al. [78] and Bhadra et al. [92] document extreme melting, migmatisation, and deformation of the EBC granites adjacent to the craton–granulite contact. The melting and migmatization intensity gradually diminish towards the interior, where nonanatectic-foliated granites can be traced.

For the present study, all representative samples are collected from the nonanatectic-foliated granites near the Ampani basin (Figure 1(b)). The nonmigmatite granites are mostly leucocratic in appearance and are variably deformed in the field. Biotite–quartz ribbon-defined fabric is prominent in granite samples collected near Ampani Basin (Figure 1(c)). The geographical locations of the samples are given in Supplement 1.

Quartz, K-feldspar, plagioclase feldspar, and biotite in decreasing abundances are major magmatic phases present in studied samples. Apatite, sphene, ilmenite, and zircon are accessory minerals. The modal abundances of minerals calculated using the CIPW norm routines in GCDKit [84] are given in Supplement 2.

Generally, the majority of study samples exhibit medium to coarse-grained porphyritic texture, where large K-feldspar and plagioclase feldspar grains are embedded in a recrystallized matrix of quartz and K-feldspar. Biotite flakes wrap the K-feldspar and plagioclase feldspar porphyroclasts (Figure 2(a)). A majority of the K-feldspar grains exhibit perthite texture which implies cooling after crystallization of feldspar (Figure 2(b)). Most quartz grains exhibit undulose extinction and chess-board twinning deformation microstructure (Figure 2(c)). Plagioclase feldspar is mostly albite (⁠$Abmole%=69–72$⁠). K-feldspar grains are sanidine with a negligible Ca-feldspar component. Regarding the composition of ferromagnesian phases, the biotite’s Fe mol % and Ti content vary between 60 to 67 and 0.03 to 0.10 (11 O a. p.f.u), respectively.

The SiO₂ and Al₂O₃ content in the studied sample varies between 61 and 69 wt.% and 13 and 15 wt.%, respectively. The FeO_t content (3.55 to 7.87 wt.%) is usually higher than the MgO content (0.50 to 2.65 wt.%) of the current set of samples. Regarding alkali elements, the CaO, Na₂O, and K₂O contents vary in the range of 1.38 to 4.11 wt.%, 1.56 to 3.05 wt.%, and 3.12 to 6.05 wt.%, respectively. The TiO₂ and MnO content is less than 1.5 wt.% and one wt.%, respectively. In a bivariate plot involving SiO₂ and other major element oxides, MgO, FeO, CaO, TiO₂, P₂O₅, and Na₂O exhibit a negative correlation with SiO₂, whereas the Al₂O₃ and K₂O exhibit scattered patterns (Figure 3).

The concentration of plagioclase-compatible elements, e.g., the studied samples’ Sr content, varies between 96 and 327 ppm. In contrast, the concentration of plagioclase feldspar incompatible elements and K-feldspar compatible elements, like Rb and Ba, varies between 116 and 218 ppm and 936 and 2164 ppm, respectively. The average Rb/Sr ratio implies K-feldspar is the dominant phase compared to plagioclase feldspar. In the primitive mantle normalized spider plot (McDonough and Sun [93]), high-field-strength elements like Th, U, and Pb exhibit a positive anomaly, and Nb-Ti exhibit a prominent negative anomaly (Figures 4(a) and 4(b)). The ΣREE of the studied samples varies between 290 and 715 ppm. In chondrite normalized REE plot (McDonough and Sun [93]), all samples exhibit an LREE enriched (⁠$La/SmCN=3.04to6.78$⁠)–HREE-depleted trend (⁠$La/YbCN=10.90−28.21$⁠) with a prominent negative Eu anomaly (0.42-0.70) (Figures 4(a) and 4(b)).

The zircon content in the current samples yields magmatic temperature between 700°C and 900°C (average temperature $819°C±57°C$⁠) [85]. The zircon saturation models after [86] yield temperature $770°C±64°C$ (Table 2).

The P-T pseudosection analysis results in the NCKFMASH system are shown in Figure 5(a). The assemblage K-feldspar–plagioclase–biotite–ilmenite is stable between 650 and 750°C, 0.3 to 0.6 GPa. Rutile with K-feldspar–plagioclase–biotite–ilmenite become stable at <650°C. Garnet stabilizes with K-feldspar–plagioclase–biotite–ilmenite, at >0.7 GPa. Similarly, temperature increase leads to stabilization of melt and garnet at >750°C, 0.3-0.6 GPa. Biotite destabilized with increasing $temperature>750°C$ and is entirely consumed at ~800°C. The assemblage K-feldspar–plagioclase–garnet–ilmenite–sanidine is stable between 800 and 900°C, 0.3 to 0.6 GPa. Finally, K-feldspar, garnet and, plagioclase are destabilized within a narrow temperature range of 900°C to 950°C, leaving silicate melt–ilmenite residue. As the observed samples do not contain garnet, the garnet-in univariant reaction line constrains the maximum temperature and pressure the rock has experienced. The P-T conditions are further constrained by modeling the chemical composition of plagioclase and biotite (Figure 5(b)).

The modeled isopleths for Ab mol % have a moderately steep positive slope with an increasing value towards lower temperature. The modeled isopleths for biotite, i.e., Fe mol %, have a steep slope, increasing Fe mol % values towards lower temperature. The modeled composition of plagioclase and biotite, similar to those computed from the measured chemical composition, i.e., $Abmole%=~75$ and $Femol%=58−61$⁠, intersect between 700°C and 750°C, 0.4-0.6 GPa.

Zircon grains are selected from representative samples that preserve the new magmatic character of the studied samples. A total of 44 spots were analyzed to constrain the geochronology of EBC granites. The representative CL images of the zircon grains are given in Figures 6(a)–6(c). The ²⁰⁷Pb/²⁰⁶Pb ratio is considered for age calculation. More than >80% of data falls near Concordia for the studied samples. The age distribution of zircon concentrates around ~2500 Ma if all analysis is taken together.

The zircon grains separated from the sample RG-37 are prismatic and euhedral to subhedral in shape (Figure 6(a)). The grain size varies from 100 μm to 500 μm. The larger grains $length>500μm$⁠, (RG-37: 1, 3, 4, and 6) exhibit a higher length : width ratio, 4 : 1 to 5 : 1, preserving prismatic crystal faces. The relatively more minor grains (RG-37: 2, 7, 8, 9, and 11) are subhedral and are characterized by a length : width ratio, 2 : 1 to 3 : 1. A majority of the grains preserve concentric zonation parallel to the prismatic crystal faces. The CL images show core-rim structure in most zircon grains with distinct chemical zoning indicating magmatic Origin (Figure 6(a)). The magmatic zoning was faint or not visible in the marginal part of most of the grains. However, the prismatic shape reflecting magmatic character was still preserved in most of the zircon grains in RG-37. Six analyses from magmatic domains and seven analyses from recrystallizing domains from eleven zircon grains are selected for isochron diagrams. The recrystallized domains exhibit a less variable U content (772-1182 ppm) and lower Th content (84–234 ppm). In contrast, the magmatic domains exhibit variable U content (100-1758 ppm) and higher Th content (170-504 ppm). The following analyses 1.1 and 4.1 (Table 3) imply Pb-loss, though the other four analyses (2.1, 6.1, 7.1, and 11.1) from zircon core yield the oldest and near-concordant ages (2460-2483 Ma) with a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of $2477±5Ma$ (⁠$MSWD=2.1$⁠) (Figure 7(a)). The spot age of analysis 6.1 is younger than the other three analyses, possibly related to slight resetting of the U-Pb system during deformation. The other three analyses show a weighted mean age of $2479±6Ma$ (⁠$MSWD=1.18$⁠) (Figure 7(b)). The weighted mean age of $2479±6Ma$ from the zircon core constrains the timing of granite emplacement. The remaining seven analyses from recrystallized domains show ²⁰⁷Pb/²⁰⁶Pb age varying between 2131 and 2421 Ma. In a U-Pb Concordia diagram, these seven zircon rim analyses and four zircon core analyses define a discordant line with an upper intercept at $2427±62Ma$ and a lower intercept at $2092±130Ma$ (⁠$MSWD=1.3$⁠) (Figure 7(c)). If the weighted age of magmatic grains $2479±6Ma$ is used to anchor the upper interception, the lower intercept is constrained at $2116±110Ma$ (⁠$MSWD=1.4$⁠) (Figure 7(d)). Though the error is too large to constrain recrystallization timing, the lower intercept can be considered the timing of zircon recrystallization during Paleoproterozoic tectonics. The tectonic significance of the same is discussed in the Paleogeographic implication section.

Zircon grains from this sample are mostly euhedral and elongated grains. Length varies from 300 to 500 μm and has a higher length : width ratio (from 3 : 1 to 4 : 1) (Figure 6(b)). Most zircons show distinct oscillatory zoning, suggesting the magmatic origin of these grains. Recrystallized rims with no zoning or blurred zoning laced the magmatic rims. The thickness of recrystallized rims are variable and often too small to analyze. A total of eighteen analyses, including fourteen from the magmatic core and four from the recrystallized rim, are chosen for further comparison. The U and Th content in the recrystallized rim vary between 331 and 1085 ppm and 33 and 276 ppm, respectively. In contrast, the magmatic cores exhibit 95–702 ppm of U and 64–559 ppm of Th. The fourteen analyses from magmatic core exhibit ²⁰⁷Pb/²⁰⁶Pb ages between 2396 and 2506 Ma. After excluding the four discordant analyses (4.1, 8.1, 9.1, and 14.1), the remaining ten analyses yielded a weighted mean age of $2471±7Ma$ (⁠$MSWD=1.2$⁠), which is considered as granite emplacement age (Figure 7(e)). The ²⁰⁷Pb/²⁰⁶Pb ages from rim (7.1, 13.1, 15.1, and 16.1) vary between 2116 and 2438 Ma. The analyses from 13.1 and 16.1 show Pb loss, but analyses 7.1 and 15.1 are plotted near the Concordia line with ²⁰⁷Pb/²⁰⁶Pb age of $2337±10Ma$ and $2267±6Ma$⁠, respectively. The ten analyses from the magmatic core and two from the rim define a discordant line with upper and lower intercept at $2458±44Ma$ and $2142±110Ma$ (⁠$MSWD=1.1$⁠) (Figure 7(f)). If the upper intercepted age is anchored at $2471±7Ma$⁠, the lower intercepted age is $2149±100Ma$ (⁠$MSWD=1.03$⁠) (Figure 7(g)). Similar to the sample RG-37, the combination of analysis for RG-11 points towards Paleoproterozoic tectonic activity error is large enough to constrain the timing preciously.

Most of the zircon grains are euhedral to subhedral shape. The length varies from 100 to 200 μm with a length : width ratio of 3 : 1 to 5 : 1 (Figure 6(c)). Five zircon grains (RG-59: 1, 2, 3, 4, and 8) are perfect euhedral prismatic crystals with long axes >200 μm with length : width is mostly 5 : 1. These perfect euhedral grains exhibit oscillatory zoning that implies differential U and Th incorporation during crystal growth in supraliquidus conditions. The remaining zircon grains are also euhedral-subhedral in shape, but the zoning is blurred or erased, implying recrystallization. In some grains, the magmatic zones are cross-cut by homogeneous (e.g., grain 4 in Figure 6(c)). The recrystallized zones show higher U content (294 to -615 ppm) than zones showing magmatic character (113 to 322 ppm), though they exhibit overlapping Th content 140 to 254 ppm, and 114 to 284 ppm, respectively. Accordingly, the recrystallized domains are characterized by a low Th/U ratio (0.41-0.58) compared to the magmatic domains (0.66-1.67). However, both the domains show similar ²⁰⁷Pb/²⁰⁶Pb ages, i.e., 2448 to 2483 Ma for recrystallizing domains and 2446 to 2497 Ma for magmatic domains, implying that U-Pb system of zircons has not been resetted. The probability density function exhibits a single peak for the analysed spot with a mean age of $2474±6Ma$ (⁠$MSWD=1.6$⁠) (Figure 7(h)). In a U-Pb Concordia diagram, the U-Pb ages plot near $~2480±11$ ages (Figure 7(h)). As the similar ²⁰⁷Pb/²⁰⁶Pb ages imply possible U-Pb isotope resetting, the U-Pb Concordia age from RG-59 is discarded from further discussion.

The chemical composition of studied samples is calc-alkaline in an AFM plot (Supplement 3a). The Fe number of the current set of samples plot in the ferroan field in a Fe# versus SiO₂ (wt.%) diagram similar to Fe# values reported for A-type granites [17] (Figure 8(a)).

In an A/NK versus A/CNK plot after [17], most samples overlapped over the metaluminous–peraluminous boundary to the peraluminous field (Figure 8(b)). Similarly, in a (Na₂O + K₂O–CaO) versus SiO₂ diagram, the samples are straddled between the alkali-calcic and calcic-alkali subgroups (Supplement 3b, [17]).

Binary plots involving a combination of major and trace elements are also helpful for discriminating the S-I-M-A granites. Accordingly, in a (FeO_t/MgO) versus Zr + Nb + Ce + Y (ppm) binary diagram, all samples plot in the A-type granite field (Figure 8(c)). A similar conclusion, EBC granites as A-type granites, can be made from (Na₂O + K₂O)/CaO versus Zr + Nb + Ce + Y, Nb versus 1000 Ga/Al, and Ce versus 1000 Ga/Al plot (Supplement 3c-e).

Several authors suggest postcollisional A-type granite magmas are usually hot and dry compared to the S- and I-type granite magma [10, 11, 13, 94–96]. Following the empirical expressions after Watson and Harrison [85], the zircon content in the current set of samples yields averages magmatic temperature (⁠$Tzr$⁠) $819±57°C$⁠. In contrast, the $TZr$ varies between 857 and 671°C (average magmatic temperature $770±65°C$⁠) if models after [86] are used. The average temperature estimated using the formulations of Watson and Harrison [85], i.e., 819°C, is marginally lower than the general cut-off temperature, i.e., ~830°C, for A-type granite [96].

In the P-T pseudosection, the temperature estimation from the intersection of compositional isopleths of plagioclase and biotite, i.e., ~750°C, at 0.4 to 0.6 GPa, is 50° to 70°C, lower than the temperate (⁠$Tzr$⁠) estimated from zircon saturation in silicate melts. Also, the magmatic temperature estimations for the current set of samples compared with the experimental wet-granite solidus imply the presence of ~five wt.% H₂O in the silicate melt (Figure 9(a)). The observation is consistent with the relatively anhydrous nature of magma parental to A-type granites [97]. In addition, the estimated P-T conditions, ~750°C–819°C, at 0.4–0.5 GPa correspond to a geothermal gradient of 1500°C/GPa [98] (Figure 9(b)), typically observed in crustal domains experiencing magmatic underplating and upwelling of hot-asthenosphere [99–103].

The discrepancies between the P-T estimates from experiments formulations, i.e., zircon saturation in silicate melt [85, 86] and forward thermodynamic modeling, i.e., P-T pseudosection analysis (Figure 5), primarily depend on the differences in the mineral end-member thermodynamic properties. Palin et al. [104] and Hernández-Uribe and Palin [105] suggest the analytical uncertainties be limited between 0.1 GPa and ± 50°C (⁠$2σ$⁠). Also, authors, Nasipuri et al. [106] and Palin et al. [104], suggest the adequate equilibrium volume, i.e., microdomain equilibrium versus whole-rock equilibrium, and the nature of mineral phase distribution affect the P-T calculation. As the temperature differences (50°C-70°C) are very close to the general analytical uncertainties, ±50°C, we suggest that the local bulk rock chemistry and considering all iron as Fe²⁺ in biotite stoichiometry may shift the temperature estimation in P-T estimation towards its lower end. Nevertheless, the average $TZr$ estimated from zircon saturation in silicate melt models, and P-T pseudosection is ~100 to 150°C higher than the average magmatic temperature of S-type granites [95].

The mean ages reported in this study, $2479±6Ma$ (RG-37), $2471±7Ma$ (RG-11), are constrained as granite emplacement age in the eastern part of the craton. Accordingly, the single peak with a mean age of $2474±6Ma$ and the U-Pb Concordia age of $2480±11Ma$ probably represent fluid flow and recrystallization of zircons from RG-59 during granite emplacement between 2479 and 2471 Ma (constrained from RG-37 and RG-11). The magma emplacement ages 2479 Ma–2471 Ma, in the Eastern Bastar Craton, are similar to the other cratons in India [46, 47], China [107], parts of Australia [108], Antarctica [109, 110], and Swaziland-Limpopo Belt [111, 112], indicating a global magmatic activity and a possible connection between these cratons during the Neoarchean–Paleoproterozoic transition.

The possible sources and magma generation mechanism for A-type granite is widely debated [11–13, 15, 16, 94, 95, 113].

Three petrogenetic models, i.e., (i) partial melting of crustal rocks [96, 114, 115], (ii) high degree melting and fractionation of mantle-derived mafic rocks [116], and (iii) by mixing of these two end-member magmas, are suggested for their generation [16, 113, 117–119]. Authors, e.g., Clemens et al. [96] and Whalen et al. [11], suggested that melting of I-type source rocks or residual granulites can produce A-type magma by relatively fluid-absent partial melting of the lower crust. Usually, the melt derived from the crustal melting materials is peraluminous, as the dehydrated restite minerals like garnet and aluminosilicate control the magma compositions [120]. The melting of hornblende or clinopyroxene bearing metaigneous rocks mostly from the mantle produces meta-aluminous A-type granites [117, 120].

In this study, most of the samples plot near the peraluminous-metalumineous field in an A/NK vs. A/CNK plot (Figure 8(b), Frost et al. [17]) imply that magma is parental to the EBC granites generated via partial melting of crustal materials. The current samples’ major element chemistry, when plot in an Al₂O₃/(FeO + MgO)–3 × CaO–5 × (K₂O/Na₂O) ternary plot after Laurent et al. [121], implies high-K mafic rocks, and metasediments are a possible source for studied samples (Figure 10(a)). Further, the studied sample’s elevated Th and Pb content implies mantle and crustal source mixing during EBC granite emplacement.

However, the ternary plot after Laurent et al. [121] could not quantify the role of crustal and mantle sources in the origin of granites from the eastern part of the craton. Thus, to separate the relative contribution of the mantle and crustal sources in the EBC granite, several trace element ratios sensitive to crust-mantle differentiation are constructed and compared with the studied samples. Eby [15] suggests that the A₁ group is characterized by trace-element ratios similar to those observed for oceanic-island basalts. In contrast, the A₂ type is characterized by variable trace-element content between the continental crust and island-arc basalts. These granites represent magmas derived from the underplated continental crust experienced in the continent-continent collision. As an incompatible element, Rb primarily resides in the crustal material, unlike the mantle-rock, usually enriched in HREEs. The Y/Nb and Rb/Nb values of the studied samples plot in the A₂ type granite field imply that EBC granites are emplaced during a continent-continent collision (Figure 10(b)). A similar conclusion, i.e., A₂ type granites derived from crustal materials, can be made from the Y-Nb-Ce ternary plot of the present samples (Figure 10(c)). Accordingly, the samples plot on the A₁–A₂ line implies mixing mantle- and crustal-derived magma. The Rb/Sr values of mantle rocks are extremely low and vary between 0.01 to 0.1 (Taylor and McLennan, 1995), the Rb/Sr values of the current set of samples (0.394 to 2.27, $mean=0.85$⁠, $1σ=0.45$⁠) imply the involvement of lower crustal rocks in their genesis. Further, the negative Nb-Ta and Ti anomalies imply magma generation in a subduction zone [122]. In an Rb versus Y+ Nb and Nb versus Y plot, EBC rocks show a syn–postcollisional granite emplacement (Figure 10(d)).

Dall’Agnol et al. [16] reviewed the geochemical characteristics and possible source rocks for alkali-calcic and calc-alkalic metaluminous and peraluminous granites. The calc-alkalic subgroup is mainly characterized by $SiO2>70wt.%$⁠, whereas the alkali-calcic subgroup shows variable SiO₂ content. The authors suggest that a wide range of SiO₂ content in alkali-calcic granites is caused mainly by the variation in the ferromagnesian mineral modal abundances. Cullers and Podkovyrov [123] suggest that the higher degree of assimilation between a mafic magma and felsic crust will produce peraluminous alkali-calcic granites, in contrast to the lower degree of assimilation between mafic and felsic sources and produce metaluminous granites. In contrast to the alkali-calcic series, meta/peraluminous calc-alkaline granites are formed by melting felsic gneiss/granulites with relatively high f (O2) conditions [16, 94, 117, 124]. Combining all the geochemical and petrographical characteristics, i.e., (a) SiO₂ content <70 wt%, (b) greater abundance of ilmenite over magnetite, and (c) peraluminous to metaluminous nature point towards alkali-calcic chemical nature for the EBC granites. Also, the higher Y (⁠$average=60.3ppm$⁠) and Yb (⁠$average=5ppm$⁠) content of the EBC granites implies its origin via slab dehydration and mantle wedge melting in relatively shallow crustal levels (c.f. Manickyamba et al. [70]). Further, the samples’ negative Sr and Eu anomalies constrain melting depth within the plagioclase-stability field. These interpretations for EBC granites are also consistent with Zr saturation temperature (⁠$Tzr$⁠) and phase-equilibria modeling, as the modeled P-T estimates also imply emplacement of EBC granites in relatively shallow depth (0.4–0.6 GPa) and high temperature (>750°C) conditions. The modeled P-T conditions correspond to a geothermal gradient of 1500°C GPa^-1, typically observed in crustal domains experiencing magma underplating [125].

The geochemical observations, i.e., elevated Th and Pb content, few samples, plot precisely on the A₁-A₂ line and relatively high geothermal gradient (1500°C GPA^-1_) implies mixing of crustal and mantle components during the emplacement of A₂-type granites in the study area.

Ur is the oldest known Archean supercontinent, which stabilized ~3000 Ma by assembling the Dharwar and the Singbhum cratons of the Indian subcontinent, the Kaapvaal craton of South Africa and the Pilbara cratons of western Australia [3, 126]. Saha et al. [26] report ~3400–3500 Ma zircon from the Bundelkhand craton and correlate the same with crust formation events in the Dharwar and the Singbhum craton as a part of the original Ur supercontinent. Rogers and Santosh [3] also, suggest that the Yilgarn craton and the Zimbabwe craton accreted with the original Ur ~2500 Ma to form the extended Ur (Figure 11(a)).

It is generally agreed that the configuration of crustal blocks in extended Ur exists till the stabilization of Mesozoic supercontinent Pangea. Mohanty [20] and Mohanty [55] studied the paleomagnetic data from the Dharwar and the Bastar craton and compared the same with paleomagnetic data from Yilgarn craton and east-Antarctica Archean blocks. The author argued that these Archean blocks were in a near-neighborhood position during the Archean–Proterozoic transition. The paleomagnetic data from ~2500 Ma aged mafic dyke from Yilgarn craton points towards its emplacement during normal polarity in the northern hemisphere or a reserve polarity in the southern hemisphere. Mohanty [20] considered the northern hemisphere reconstruction for these mafic dykes to exhibit a close correlation between the Bastar craton and Yilgarn craton, with probably a matching boundary at the current position of the Eastern Ghats Belt at 2500 Ma (Figure 11(b)). The Singbhum craton was amalgamated with the rest of the Indian peninsula at 2400 Ma and with Rengali province [50]. When combined with the geochronological data from the Dharwar, Bastar, and Singbhum cratons (as South India craton) and Yilgarn craton from Australia, the available paleogeographic models imply the existence of an Archean–Paleoproterozoic supercraton (SIWA) within the overall framework of extended Ur supercontinent (Figure 11(b)). The 2100 Ma extensional events probably separate the Yilgarn craton from the Bastar craton and the Napier complex from the Dharwar craton [20, 25, 61]. Mohanty [20] argues for an ocean basin opening at the present-day Eastern Ghats Belt at the east coast of India (Figure 11(c)). Finally, the Archean Napier Complex, Antarctica, accredited with the East Dharwar craton and Domain 1A of Eastern Ghats Belt, at ~1700 Ma as Columbia supercontinent [127, 128]. However, the amalgamation timing between the Bastar cratons and other Archean cratons of Antarctica during the Columbia assembly is still debated [21, 78].

Integration of the geochemical interpretation and the results of phase diagram modeling, including thermometric estimates from zircon solubility in silicate melt and U-Pb geochronology, enable us to reconstruct the tectonic model for the thermal evolution of the study area.

A detailed compilation of available data indicates that most of the western part of the craton experienced granite magmatism between 2800 Ma and 2500 Ma [59, 70]. Khanna et al. [58] suggest the development of an Andean-type magmatic arc around 2500 Ma, where an oceanic spreading ridge was getting subducted beneath the continental lithosphere. Manikambya et al. [70] suggest a similar model where rocks below that experience slab-dehydration, asthenosphere melting, granite, and rhyolite emplacement in the western part of the Bastar Craton (Figure 13 of Manikyamba et al. [70]). By implication, the oceanic crust present in the present-day EGB should be older than ~2800-2500 Ma granites in the WBC.

However, the present tectonic model, i.e., west-vergent subduction of oceanic lithosphere below the present-day Dongargarh granite, fails to address the origin of A₂ type granites in the eastern part of the craton, as the origin of alkali-calcic peralkaline A₂ type requires slab hydration and mixing of mantle-derived mafic and crustal material [70, 129]. Given the present tectonic scenario, we suggest a modified tectonic model by combining the results of geochemical interpretation, phase-topological data, geochronology, and paleogeographic data.

The paleogeographical data suggest that the Dharwar, Bastar, and the Yilgarn craton were assembled at around 2500 Ma in the SIWA supercraton [20]. We suggest that an Archean Sea separate the Bastar and the Yilgarn craton before 2500 Ma (Figure 12(a)). Like the other Archean cratons, i.e., the Aravalli and the Bundelkhand craton [130], the Archean sea probably receives weathered sediments from the oldest crustal components from the Bastar Craton. Westerly-directed subduction caused granite emplacement ~2800–2500 Ma in the western part of the craton (Figure 12(b), [59]). We also suggest a nearly similar tectonic setup between the tonalites and Yilgarn Craton, where an oceanic crust was gradually subducted below the oldest tonalities present in the southeastern part of the craton. As the subduction continues, the oceanic slab gets dehydrated and melted and finally assimilates with the asthenosphere (Figure 12(c)). Subsequently, the rising asthenosphere will melt the crustal rocks, i.e., TTGs and their weathered products, to produce A₂-type granites in the study area. Accordingly, ~2470 Ma magmatic ages imply A₂ type granite emplacement during the final amalgamation phase between eastern part of the Bastar craton with Yilgarn craton as a part of extended Ur (Figure 12(d)). The Bastar Craton-Yilgarn Craton block experienced rifting mafic dyke emplacement and subsequent disintegration as reflected by the 2100 Ma zircon rim (Figure 12(e)). The 2100 Ma ages probably represent the timing of dyke emplacement in Bastar-Dharwar and Bundelkhand craton [61, 131, 132]. Hifzurrahman et al. [133] correlate the ~2100–1900 Ma zircon growth from North-Indian-Continental-Block with postextended Ur–pre-Columbia tectonics. However, such an integrated study reflecting the disintegration of extended Ur and subsequent amalgamation in Columbia is lacking from the Bastar Craton.

Geochemical characterization implies that EBC rocks are A₂-type granites. The results of P-T pseudosection analysis and U-Pb zircon geochronology point towards the emplacement of these rocks during the Bastar Craton–Yilgarn Craton amalgamation process in a convergent tectonic setting with a geothermal gradient close to 1500°C/GPa at ~2470 Ma as a part of extended Ur configuration. The U-Pb Concordia lower intercept., ~2100 Ma, probably represents the timing of dyke emplacement and disintegration of Archean blocks in extended Ur before the amalgamation of the Paleoproterozoic Columbia supercontinent.

Additional data for academic research purposes is available from the corresponding author for academic research on request.

The authors declare that there is no secondary or financial interest regarding the publication of this paper.

Kausik Satpathi and Ab Majeed Ganaie acknowledge the Department of Science and Technology, and Hifzurrahman acknowledges a Ph.D. fellowship from University Grants Commission, India. Sagar Misra acknowledges Prof. Siva Umapathy, Director, IISER Bhopal, for Institute Post-Doctoral Fellowship. Pritam Nasipuri gracefully thanks Prof. A Bhattacharya for his introduction to the field. Pritam Nasipuri is grateful to Prof. Fernando Corfu for his generous help and suggestions to understand the zircon geochronology about Pan-African orogeny from this area. H-Q Xie thanks Chun Yang, Zhichao Zhang, and Jianhui Liu for mount making, zircon CL imaging, and helping with SHRIMP zircon dating. Zircon geochronological costs were covered by the National Nature Science Foundation of China (grant number 41872200) awarded to Hang-Qiang Xie. Pritam Nasipuri thanks Prof. Siva Umapathy for encouragement and supporting partial analytical cost.