Neodymium Isotope Constraints on the Origin of TTGs and High-K Granitoids in the Bundelkhand Craton, Central India: Implications for Archaean Crustal Evolution

The Earth’s continents started to form four billion years ago in the Archaean Eon, and they grew larger through time. Over a very long time, ancient continents have joined, broken apart, and drifted around the globe multiple times into new configurations because of plate tectonics [1–3]. To understand the formation of the oldest continents, we need to study their fragments and correlate their chemical compositions and ages, which have been a great challenge and topic of debate for decades.

We know that, in the early Archaean, the main crust formation process was the episodic melting of basaltic crust that led to the formation of TTGs (tonalite-trondhjemite-granodiorites) and building of the first continents [4]. The conditions under which the first continental crust formed were different from those presently operative. The mantle was hotter with smaller basaltic plates and faster convection rates [5]. Different conditions resulted in distinct geochemistry and isotope signatures of Archaean granitoids [6–11]. The geochemical composition of granitoids depends on various factors, including the physical and chemical conditions of the source, amount of residual minerals, varying anatectic conditions, stages of magmatic differentiation, and tectonic setting. Therefore, their compositional variation provides crucial information on the plate tectonic processes and crustal evolutionary history of the Early Earth [7, 10, 12–15]. The isotope systematics and geochemical signatures of granitoids provide insights into the age and nature of their source and role in crustal evolution.

Today, there is a general agreement that partial melting of hydrated basalts during the Archaean has produced the early felsic crust consisting mainly of silicic and sodic TTGs. However, the tectonic setting of TTGs is still controversial, and the views range from stagnant lid tectonics and plume tectonics involving mantle upwelling to arc tectonics related to subduction [16–20]. Some researchers favour partial melting of subducting oceanic slab and believe that the formation of Archaean protooceanic crust (hydrous metabasalt) and subsequent deep burial via subduction (a process similar to modern plate tectonics) can explain the generation of TTG melts at relatively great depths [17, 18, 21, 22]. Other researchers believe that TTGs could form in overthickened mafic crust [23] in island arc settings [24–26] or plume-related basalt plateaux [27–30]. Studies also suggest that the vertical growth of Archaean oceanic plateaux above mantle plumes can allow hydrothermal alteration of basalts required for TTG formation [31–34]. Recent studies suggest that TTGs can also be generated by fractional crystallization or by fractionation of melts derived from the enriched lithospheric mantle [35–37].

During the Meso- to Neoarchaean (3.0-2.5 Ga), the episodic melting in basaltic sources leading to the formation of TTGs declined, and a new type of multisource high-K calc-alkaline granitoids emerged as a consequence of changes in the Earth’s geodynamics [14, 19]. These high-K granitoids are the second-most voluminous component of the Archaean crust after the TTGs. They are multisourced, ranging from the mantle or mixed mantle and crustal sources to pure crustal sources. They are formed by increasing crust-mantle interactions due to the onset of modern-type plate tectonics [8–10, 12, 38, 39]. Joshi et al. [9] divided high-K granitoids into two main groups, the Low Silica, High Mg (LSHM) group and the High Silica, Low Mg group (HSLM). The former includes mantle-derived sanukitoid granitoids and related hybrid granitoids formed by increased mantle-crust interactions, and the latter anatectic granitoids formed by pure crustal melting.

Geochemical signatures and geochronology of TTGs and high-K granitoids divide the Archaean crustal evolution into four main stages: (1) In the Eoarchaean, TTGs formed by episodic melting within the thin or thickened basaltic crust. 2) In the Paleoarchaean, TTG magmatism continued along with the formation of diorites and anatectic high-K granites, which caused the thickening of the crust. (3) In the Mesoarchaean, TTG formation continued and the first granitoids with mantle contributions, known as sanukitoids, appeared at around 3.0 Ga. (4) Neoarchaean was the time of granitoid diversification when TTG formation decreased and high-K magmatism increased indicating a significant geodynamic change that culminated in the Archaean-Proterozoic boundary [12]. In the Indian Shield, five cratons are the results of Paleoarchaean to Neoarchaean crustal growth (Figure 1). In this study, we focus on the Bundelkhand craton between the Aravalli and Bastar cratons in central India. This craton is a key research target because it includes Paleo- to Neoarchaean fragments of TTG crust embedded in abundant high-K granitoids formed at the Archaean-Proterozoic boundary. Therefore, the craton gives important knowledge on the evolution of the Paleoarchaean crust and the stabilisation of a supercontinent.

Recent studies have proposed a petrogenetic explanation of these diverse granitoid types based on whole-rock major and trace element geochemical and mineral chemistry [9, 40–42]. However, very few studies [43, 44] have focused on isotopic systematics to explain the plausible sources of the various granitoid varieties emplaced in the craton, and even fewer have compared the geochemical signatures with isotope results. In this contribution, we provide insights into the petrogenesis of the different types of the Bundelkhand granitoids by comparing new whole-rock Nd isotope results with compiled geochemical signatures and U-Pb ages. The aim is to correlate the geochemical fingerprints with the Nd isotope signatures to enlighten their source and tectonic setting. The Sm-Nd method has its limitations since the model ages are dependent on the suitability of the mantle model and the analytical uncertainties are relatively large. However, the εNd(⁠$t$⁠) values and Nd-depleted mantle model ages have been successfully used in determining crustal formation ages, representing the approximate time when crustal blocks were first created by mantle-derived magmatism.

This study is aimed at providing significant information for defining the (1) Paleoarchaean and Neoarchaean evolution of the TTG crust, (2) stabilisation of the continental crust at the end of the Neoarchaean, and (3) the age, sources, and crustal contribution in the type formations of the Archaean-Proterozoic boundary. The last outcome can significantly contribute to the aims of the International Commission on Stratigraphy (ICS) in its efforts to define the border of the Archaean and Proterozoic Eons.

There are five major Archaean cratons in two distinct crustal blocks in the Indian Shield (Figure 1), the Northern and Southern Blocks separated by the Central Indian Tectonic Zone (CITZ). The former consists of the Bundelkhand and Aravalli cratons, while the latter consists of the Dharwar, Singhbhum, and Bastar cratons. Figure 1 shows a generalized geological map of the Bundelkhand craton with an inset showing the different cratons of the Indian Shield. The semicircular body of the Bundelkhand craton is bounded by the Paleoproterozoic (2.0–1.8 Ga), Gwalior (Northwest), Sonrai (South) and Bijawar (Southeast) basins which are overlain by the Mesoproterozoic Vindhyan Supergroup that occurs on three sides of the craton. The northern side of the Bundelkhand craton is covered by Indo-Gangetic alluvial plains [45–47].

The dominant lithology in the Bundelkhand craton (Figure 1) consists of Neoarchaean potassic granites, which are intruded into Paleo-Mesoarchaean TTG gneisses and supracrustal units. The E-W-trending Bundelkhand tectonic zone (Figure 1) divides the craton into three segments, viz., (i) the central Bundelkhand granite-greenstone terrane, (ii) the southern Bundelkhand granite-greenstone terrane, and (iii) the north Bundelkhand granitoid terrane [48–51]. The central Bundelkhand granite-greenstone terrane extends from Mahoba to Babina and consists of TTGs, volcano-sedimentary sequences, and high-K granitoids (including sanukitoids) [9, 40, 43, 44, 52–59]. The southern Bundelkhand granite-greenstone terrane consists of quartzite, BIF, chlorite schist, and marble and extends from Rungaon to Girar [48, 51, 60–62]. Mafic dyke swarms and giant quartz veins traverse the above lithologies and represent the last magma-related hydrothermal activity in the Bundelkhand craton [52, 63–66].

Table 1 presents compiled geochronological data from the Bundelkhand craton [9, 43, 44, 48, 52, 54–58, 67–70]. The oldest rocks in the Bundelkhand craton are Paleo- to Neoarchaean TTGs showing distinct formation episodes at ca. 3.56 Ga, 3.44 Ga, 3.3 Ga, 3.2 Ga, and 2.71-2.68 Ga. Singh and Slabunov [48] and Slabunov and Singh [62] dated felsic volcanics from Babina and Mauranipur at 2.54, 2.56, and 2.81 Ga, respectively, while the mafic-ultramafic rocks from Babina have yielded a Sm-Nd isochron age of 3.4 Ga [50]. The TTGs occur as east-west-trending rafts within abundant high-K granitoids, the dominant rock type of the craton emplaced at the 2.5 Ga Archaean-Proterozoic boundary.

The granitoids of the Bundelkhand craton show typical geochemical characteristics of TTGs, sanukitoids, and anatectic granitoids. Full geochemical datasets, ages, and explanations for the Bundelkhand granitoid samples that are analysed for Nd in this study are available in the works of Joshi et al. [9, 40]. In this paper, we address the possible source characteristics of the Bundelkhand granitoids in terms of the geochemical signatures or fingerprints suggested for Archaean granitoids based on data compiled from several cratons [12].

• (1)

  Mantle signature (low SiO₂ contents and high contents of mantle-compatible elements Mg, Cr, and Ni) can be observed in mantle-derived granitoids

• (2)

  Basaltic crust signature (high Na, high SiO₂ contents, and low contents of mantle-compatible elements) in granitoids derived from basaltic precursors

• (3)

  Enriched mantle signature (high Mg-K-Ba-Sr-P and LREE), inherited from Archaean mantle overprinted by crust-mantle interactions, is typical for sanukitoid granitoids. This signature cannot be a consequence of fractional crystallization, because it is independent of the SiO₂ content and can be also found in mafic rocks, such as lamprophyres, which are often associated with sanukitoids

• (4)

  Garnet fingerprint (low Y and HREE, high Gd/Er) derived from garnet-bearing sources, where garnet is retained in the residue (e.g., deep lower crust). Thus, the presence or absence of this fingerprint reflects the depth of melting

• (5)

  Continental crust signature (high SiO₂ and K₂O contents, mantle and enriched mantle signatures absent) in continental crust-derived granitoids

The geochemical classification scheme of Joshi et al. [9] in Figure 2 divides the Bundelkhand granitoids into two main groups: high-Na TTGs (~3.5-2.6 Ga) and high-K granitoids (~2.5 Ga). The TTG group carries the basaltic crust signature and consists of low-HREE and high-HREE end members (garnet fingerprint present or absent, respectively). The key geochemical signatures of the low-HREE group are LILE enrichment and depletion in HFSE, Y, and Yb, whereas the high-HREE group shows elevated HFSEs and enrichment in LREE. The silica and magnesium contents of the high-K granitoids further divide them into LSHM (Low Silica, High Mg) and HSLM groups. The LSHM group shows enriched mantle signatures and mafic magmatic enclaves. The strength of the mantle signature may vary from high in sanukitoid granitoids to lower in hybrid (“Closepet-type”) granitoids formed by stronger crust-mantle interactions. The HSLM group consists of anatectic granitoids with a continental crust signature. Specific geochemical characteristics divide this group into low-HREE monzogranites from deep HREE-depleted crust, monzogranites (biotite granites) from the middle crust, and A-type granites with strong negative Eu anomalies from the upper crust.

The detailed field descriptions, major and trace element geochemistry, and U-Pb zircon geochronology of TTGs and high-K granitoids analysed for Nd isotope compositions in this study can be found in the work by Joshi et al. [9, 40]. A summary of the sample locations and descriptions of the studied granitoid varieties is given in Table 2 while field and petrographic images are shown in Figures 3 and 4. The TTGs are exposed as E-W-trending slivers in Mahoba, Mauranipur, and Babina. They are deformed coarse-grained heterogeneous bodies, typically exhibiting alternating layers of leucocratic and melanocratic bands (Figure 3(a)) and are occasionally homogeneous (Figure 3(b)). The TTGs are inequigranular and consist of coarse-grained plagioclase (45 vol%), quartz (22 vol%), K-feldspar (18 vol%), biotite (10 vol%), and amphibole (5 vol%) (Figures 4(a) and 4(b)) with minor amounts (1 vol%) of titanite, epidote, apatite, and zircon.

The high-K group consists of grey or pink, fine- to coarse-grained, and occasionally porphyritic granitoids (Figures 3(c)–3(f)). Sometimes, the high-K granitoids are slightly deformed and show an alignment of mafic minerals. In the LSHM varieties, the K-feldspar dominate over plagioclase. Mineral abundances depict a higher quartz content (27 vol%) as compared to TTGs, plagioclase (35 vol%), K-feldspar (15 vol%), and the presence of amphibole (15 vol%) and biotite (7 vol%) as major minerals (Figures 4(c)–4(f)). The common accessory minerals (1 vol%) in LSHM varieties include zircon, titanite, epidote, and apatite. The HSLM varieties have a very similar mineral composition [K-feldspar (35 vol%), quartz (40 vol%), plagioclase (22 vol%), and biotite (2 vol%)] with accessory phases like amphibole, titanite, epidote, apatite, and zircon (1 vol%), but they have low to negligible amphibole content and, occasionally, fine-grained groundmass with phenocrysts of quartz and feldspar (Figure 4(f)).

Approximately 200 mg of whole-rock powder was first decomposed in a mixture of HF-HNO₃-HCl in microwave digestion. The solution was further transferred in Savillex® vials and digested repeatedly with HF-HNO₃-HCl at ~100°C to bring the powder to complete solution. The acid digestion step was repeated as needed to ensure the complete digestion of the sample. Post digestion, the solution was dried and divided into two aliquots by weight. A known amount of ¹⁴⁹Sm and ¹⁴⁵Nd spikes by isotope dilution was added to the first aliquot, and Sm and Nd concentrations were measured by QICP-MS (Thermo Xseries-2) at the Physical Research Laboratory (PRL), Ahmedabad, India. The second aliquot was used to separate Nd by cation exchange columns, following standard ion-exchange procedures [71], ¹⁴³Nd/¹⁴⁴Nd ratios were measured on a Finnigan Neptune MC-ICP-MS at PRL, and the analyses were carried out in static multicollection mode. Mass fractionation corrections were made by normalising ¹⁴³Nd/¹⁴⁴Nd ratios to $146Nd/144Nd=0.7219$⁠. The JMC for the Nd (¹⁴³Nd/¹⁴⁴Nd) isotope standard was measured during analyses which yielded values of $0.710343±0.000002$ (1σ, $n=7$⁠; σ = standard deviation) and $0.511714±0.000004$ (1σ, $n=15$⁠). Total procedural blanks for Nd were several orders of magnitude lower than typical total Nd loads analysed, and hence, no corrections for blanks were made.

The single-stage Nd model age (⁠$TDM1$⁠), also known as “crustal extraction/formation age” or “crustal residence age,” is the time elapsed since the Nd in the rock separated from the mantle. In the case of single-stage Nd model ages, it is assumed that crustal processes like metamorphism and intracrustal melting have not changed the Sm/Nd ratios. Archaean granitoids might have suffered multiple melting episodes of magma mixing, melting, fractional crystallization, and alteration. Therefore, to correct for changes in Sm/Nd ratios produced by crustal processes, two-stage depleted mantle Nd model ages (⁠$TDM2$⁠) were calculated using the present-day depleted mantle ¹⁴³Nd/¹⁴⁴Nd and ¹⁴⁷Sm/¹⁴⁴Nd values of 0.51315 and 0.2136 and decay constant value of $6.54×10−12a−1$⁠, and the rock formation ages [72] suggested that the two-stage model ages give more consistent ages. The depleted mantle (DM) has been calculated using the linear depletion model, which assumes linear depletion from $εNd=0$ at ~4.56 Ga to +10 today. The isotope compositions of samples from Bundelkhand granitoids are presented in Table 2.

The Sm-Nd results on 44 whole-rock samples of TTGs and high-K granitoids are shown in Table 3. We calculated εNd(⁠$t$⁠) values based on U-Pb zircon data from Joshi et al. [9] and unpublished ages, on the same sample or samples within the same geochemical group. Fifteen samples of Paleoarchaean TTGs (3.4-3.3 Ga) show positive εNd (3345 or 3335) values from 0.54 to 1.96 and the Nd two-stage depleted mantle (⁠$TDM2$⁠) model ages from 3592 to 3397 Ma. Neoarchaean TTGs, on the other hand, show mostly negative εNd (2713) values ranging from 0.36 to -2.61 and Nd (⁠$TDM2$⁠) model ages of 3314-2965 Ma. The 2.57-2.53 Ga high-K LSHM granitoid group (15 samples of sanukitoids and hybrids) shows εNd(⁠$t$⁠) values from -3.60 to 3.13 (Table 3). Their Nd (⁠$TDM2$⁠) model ages vary within a relatively smaller range from 3144 to 2619 Ma. The 2.59-2.54 Ga high-K HSLM varieties (18 samples) show εNd(⁠$t$⁠) values ranging between -0.30 and -5.89 and Nd (⁠$TDM2$⁠) model ages ranging from 3190 to 2820 Ma.

The εNd(⁠$t$⁠) values and zircon ages for the TTG, LSHM, and HSLM granitoids of this study together with compiled data from Bundelkhand, Dharwar, and Aravalli cratons are shown in Figure 5. The positive εNd(⁠$t$⁠) values indicate derivation from a juvenile source or a short-lived mafic crust, while the negative εNd(⁠$t$⁠) values point towards crustal inputs. A large gap between the model and crystallization ages point to large crustal residence times and a dominant role of the older continental crust.

The granitoids of the Bundelkhand craton show a wide range in composition and ages from Paleoarchaean high-Na TTGs to abundant high-K granites emplaced at the Archaean-Proterozoic boundary. The multicationic classification diagram [73], Al₂O₃/(FeO_t+MgO), $3∗CaO$⁠, $5∗K2O/Na2O$⁠, and $2∗A/CNK$⁠; Na₂O/K₂O ratio; and $2∗FeOt+MgOwt%∗Sr+Bawt%$(=FMSB) diagrams [10] for the Bundelkhand granitoids are shown in Figure 6. The TTGs plot in the tonalite and granodiorite fields, whereas the LSHM sanukitoids and hybrids mainly occupy the granodiorite, quartz monzonite, and granite fields (Figure 6(a)). The anatectic HSLM group falls in the quartz monzonite and granite fields (Figure 6(a)). Figure 6(b) shows that the major-element composition of LSHM sanukitoids and hybrid granitoids does not point to an origin through the melting of a single crustal lithology. Several studies have proposed a mixed origin, viz., mixing between sanukitoid and TTGs, contamination of juvenile magma with preexisting crust, and mixing of TTG magma with enriched mantle, for such granitoids [9, 74–76]. On plotting the studied granitoids on the Na₂O/K₂O-FMSB-A/CNK plot (Figure 6(c)), it is noted that the studied granitoids fall well within TTG, sanukitoid, hybrid granitoid and biotite, and two mica granites from the Dharwar craton.

The change from sodic TTGs to potassic granodiorites and granites is considered to be a consequence of increasing crustal contributions. In addition to the composition of the granitoids, certain geochemical fingerprints, differences in U-Pb zircon formation ages and Nd model ages, εNd(⁠$t$⁠) values, and occurrence of inherited zircons are regarded as measures of crustal contamination. In this discussion, we compare the main geochemical signatures of the Bundelkhand granitoids with Nd results (Appendix 1) of this and other relevant studies from the Dharwar and Aravalli cratons [43, 44, 77–84]. To study the sources of granitoids, the selected elements [SiO₂, Na₂O, K₂O, MgO, P₂O₅, Ba+Sr, and (Gd/Er)_N] are plotted against U-Pb zircon ages (Figures 7 and 8) and εNd(⁠$t$⁠) values (Figures 9 and 10). The SiO₂ and MgO contents reflect the mantle vs. crust contributions, Na₂O indicates the role of a basaltic source, K₂O may indicate crustal contamination or origin from an enriched mantle, and P₂O₅ and Ba+Sr are regarded as indicators of contributions from an enriched mantle.

The Bundelkhand TTGs carry the specific “basaltic crust signature” typical for Archaean TTGs: high Na and SiO₂ contents and low contents of mantle compatible elements. The number of samples is not adequate for far-reaching interpretations but allows us some conclusions on the TTGs of the Bundelkhand craton. The element vs. U-Pb zircon age plots (Figure 7) indicate that the Paleoarchaean TTGs (3.3-3.2 Ga) show a wider range in their SiO₂ and MgO contents than the Neoarchaean TTGs (2.7 Ga). Table 4 shows that the Paleoarchaean TTGs have shorter time gaps (~41-220 Ma) between the U-Pb formation ages and Nd model ages than the Neoarchaean TTGs (~354-601 Ma). For the Paleoarchaean TTGs, the element vs. εNd(⁠$t$⁠) plot (Figure 9) shows a narrow range of slightly positive εNd(⁠$t$⁠) values between 1.96 and -0.53, but there is no correlation between the SiO₂ or MgO contents and εNd(⁠$t$⁠) values. Conversely, Neoarchaean TTGs show mainly negative and more variable εNd(⁠$t$⁠) values (between -2.61 and 0.36). The increasing SiO₂ and slightly falling MgO trends with decreasing εNd(⁠$t$⁠) values (Figure 9) point towards derivation from a mafic crust contaminated by preexisting crust.

Figure 7(g) shows that Paleoarchaean TTGs have variable HREE characteristics, as measured by the (Gd/Er)_N ratio, whereas the Neoarchaean samples show high HREE characteristics. The reason for the variation in the HREE contents of TTGs is not sustained yet, but melting at different depths can account for the variation [28, 85, 86]. In the Bundelkhand craton, the high-HREE TTGs occur as metatexite enclaves in diatexitic low-HREE TTGs. This may indicate migration of melts from deep garnet-bearing sources into garnet-free amphibolite facies as suggested for TTGs of Arctic Fennoscandia [28]. Short time gaps between formation and model ages, narrow range of slightly positive εNd(⁠$t$⁠) values, and wide variation of SiO₂ and MgO lacking correlation with εNd(⁠$t$⁠) values suggest that the Paleoarchaean TTGs formed from juvenile of short-lived mafic crust with contributions from newly formed felsic crust wherein distinct isotope compositions had not developed yet. In the Neoarchaean TTGs, long gaps between formation and model ages as well as SiO₂ and MgO correlating with negative εNd(⁠$t$⁠) values indicate interactions of juvenile sources with older sources. The presence of inherited zircons in ~2.70 Ga TTGs [43] further supports our conclusion.

Based on the above, we conclude that the geochemical and Nd isotope signatures of Archaean Bundelkhand TTGs point towards partial melting of juvenile or short-lived mafic crust at different depths with variable contributions from newly formed felsic crust in the Paleoarchaean and older felsic crust in the Neoarchaean. This may reflect the transition from melting in Paleoarchaean oceanic plateau in plume settings to shallow melting in Neoarchaean subduction-related island arcs.

The LSHM granitoids, especially sanukitoids, carry a specific “enriched mantle signature” (high Mg-K-Ba-Sr-P and LREE) that is inherited from Archaean mantle overprinted by crust-mantle interactions. This signature cannot be a consequence of fractional crystallization, because it is independent of the SiO₂ content and occurs in mafic rocks, such as lamprophyres, which are often associated with sanukitoids.

The element vs. U-Pb zircon age plots (Figure 8) indicate that the 2.5 Ga LSHM granitoids have a wider range of element contents than the HSLM group of granitoids. The LSHM granitoids show variable time gaps (~99-590 Ma) between the U-Pb formation age and Nd model age (Table 4). The εNd(⁠$t$⁠) values for the LSHM group consisting of sanukitoid and hybrid granitoids range from -3.64 to 3.13 (Table 4). Sanukitoids show a narrow range of εNd(⁠$t$⁠) values and wide variation in Na₂O, K₂O, MgO, P₂O₅, (Gd/Er)_N, and high Ba+Sr contents but a minor variation in the SiO₂ content (Figure 10). Some of the hybrid granitoids show higher εNd(⁠$t$⁠) values than sanukitoids. Those with positive εNd(⁠$t$⁠) values show mainly higher SiO₂ content and lower MgO content, which supports crustal assimilation. Other signatures are similar to those of sanukitoids.

The enriched mantle geochemical signature (high Mg-K-Ba-Sr-P) is typical for sanukitoid granitoids. The plots in Figure 10 show that these elements do not correlate with SiO₂ content or εNd(⁠$t$⁠) values; therefore, we agree that the signature comes from an enriched mantle. Our results support the previous ideas on the enrichment of the mantle by partial melting in a mantle wedge [87–89] metasomatized by slab-derived melts or fluids from a subducting oceanic slab [87, 90, 91]. Two-stage enrichment, firstly by subduction and secondly by low-degree melting and metasomatism in the mantle would best explain the lack of correlation between Nd isotope and geochemical signatures [85]. Some researchers [92–94] have suggested that sanukitoids were formed by interactions between mantle and TTG melts. However, the discrepancy between ages as well as geochemical and Nd isotope signatures contradicts this suggestion in the Bundelkhand craton. According to Halla [95], radiogenic Pb isotope signatures that developed in the old preexisting crust may erode and enter the mantle, overprint the mantle composition, and return to the crust by juvenile magmatism. Such crustal isotope signatures in mantle-derived rocks form proof of crustal recycling and subduction processes. This implies that crustal signatures may be observed in geochemistry but not in the isotope systematics (in the case of young crust) and vice versa (in the case where crustal isotope signatures are inherited from the mantle).

The HSLM anatectic granitoids show a “continental crust signature” (high SiO₂ and K₂O contents, mantle and enriched mantle signatures absent) pointing to pure crustal origin. The element vs. U-Pb zircon age plots in Figure 8 show less variation in the element contents than the LSHM group. The time gap between formation and model age varies between 258 and 644 Ma. The εNd(⁠$t$⁠) values for the HSLM group are all negative and range from -0.30 to -5.89 (Table 4). These values are consistent with the derivation from reworking of a heterogeneous crustal source, pointing to the involvement of older heterogeneous crust. The HSLM group shows a relatively narrow range of εNd(⁠$t$⁠) values and minor variation in the element contents without correlation with εNd(⁠$t$⁠) (Figure 10), which supports a pure crustal source without a mantle contribution. The absence of inherited zircon grains in HSLM varieties led Singh et al. [44]) to conclude that these anatectic granitoids were formed in conditions that inhibit zircon survival; however, Joshi et al. [9] reported ²⁰⁷Pb/²⁰⁶Pb xenocrystic zircon ages of 3568 Ma and 2787 Ma from HSLM granitoids. The involvement of the crustal component in the generation of LSHM as well as HSLM granitoids is further supported by the occurrence of xenocrystic zircons within both the granitoid groups (LSHM and HSLM), which further suggests that crustal reworking was at its peak during the Neoarchaean.

The location of the Indian cratons on a world map modified after Bleeker [96] and the geochronology of TTGs, sanukitoids, and anatectic granitoids in the Bundelkhand, Aravalli, Singhbhum, Bastar, and Dharwar cratons in India are shown in Figure 11. The North China craton is included in the figure because it shows pulses of 2.5 Ga LSHM and HSLM magmatism comparable to the Bundelkhand craton. Next, we describe how the results from the Bundelkhand craton relate to Archaean crustal evolution in general.

In the Eoarchaean, TTGs with high SiO₂, high Na, MgO, and variable HREE formed by episodic melting within the thin or thickened basaltic crust. Eoarchaean ages have not been reported in the Bundelkhand craton, but the oldest documented U-Pb ages of 3.55 Ga for zircons and 3.59 Ga for zircon xenocrysts together with the oldest model ages of 3592 Ma suggest that the crust formation in the craton started close to the Eoarchaean-Paleoarchaean boundary [54, 58]. Eoarchaean TTGs are found, for example, in the Slave Province, West Greenland, and North China craton.

In the Paleoarchaean, TTG magmatism continued along with the formation of diorites and anatectic high-K granites. The Singhbhum and West Dharwar cratons show abundant TTG formation in the Paleoarchaean, and the Bundelkhand craton includes abundant rafts of Paleoarchaean TTGs. The geochemical and Nd isotope signatures of 3.4-3.2 Ga Bundelkhand TTGs point towards partial melting of juvenile or short-lived mafic crust at different depths (reflected by variable (Gd/Er)_N ratios) with some crustal contributions. Paleoarchaean TTG crust shows slightly positive εNd(⁠$t$⁠) values that do not correlate with crustal geochemical signatures. The time gaps between crystallization and model ages are short. These indicate contributions from a newly formed crust that has not yet developed a distinct isotope composition. Figure 11 shows that in India, especially the Bundelkhand, Singhbhum, and Western Dharwar cratons were active during the Paleoarchaean. Figure 5 shows that the εNd(⁠$t$⁠) values from the Bundelkhand craton are comparable to those from the Dharwar craton.

In the Mesoarchaean, the TTG formation continued, but the Bundelkhand craton was quieter until the Neoarchaean. The first LSHM sanukitoid granitoids appeared at 3.0 Ga in the Carajas Province, Brazil [92, 93] and the Pilbara craton, Australia [97]; however, no such granitoids have been reported from the Bundelkhand craton.

In the early Neoarchaean, 2.7 Ga TTG crust shows negative εNd(⁠$t$⁠) values that correlate with crustal geochemical signatures. The long gaps between the crystallization and model ages indicate significant contributions from the older crust. These discoveries may reflect the transition from the Paleoarchaean primitive oceanic plateau setting to island arc accretion along an older protocontinent, possibly related to the formation of the supercontinent Kenorland. During the beginning of the Neoarchaean, the TTG formation decreased and high-K magmatism increased indicating a geodynamic change, probably approaching modern-style plate tectonics with the mantle wedge above subduction zones. The Bundelkhand craton was quiet until the Archaean-Proterozoic boundary when the formation of extensive multisource batholiths involving both mantle- and crust-derived materials (LSHM and HSLM granitoids) started. The most important representatives of this event are found in the Bundelkhand and North China cratons [9, 43, 44, 98, 99].

The geochemical and isotope signatures of the 2.5 Ga LSHM granitoid group correspond to major juvenile inputs (high content of mantle compatible elements with or without older crustal components, as reflected by the εNd(⁠$t$⁠) values varying from negative to positive). Sanukitoid granitoids inherited their special signature (high K-Mg-P-Ba-Sr) from the metasomatically enriched mantle, whereas direct crustal contributions have modified the hybrid granitoids. The enriched mantle signatures in the LSHM granitoids have been related to the onset of modern-style plate tectonic processes [12], as recycling of crustal material into the mantle can enrich the subcontinental lithospheric mantle wedge with incompatible elements and crustal isotope signatures, which is generally attributed to subduction. The LSHM sanukitoid and hybrid granitoids formed from crust-enriched mantle reservoirs with or without mixing with anatectic melts. This scenario points to convergent collisional settings. Figure 11 shows that sanukitoid and hybrid granitoids were formed also in the Dharwar and, especially, in the North China craton (for references, see the figure caption). The formation of abundant multisource granitoid batholiths worldwide marks the stabilisation of the supercontinent Kenorland at the Archaean-Proterozoic boundary.

The Bundelkhand craton is a key locality for studies on the abundant multisource granitoid batholiths that stabilised the supercontinent Kenorland by 2.5 Ga. The Nd isotope of this study supports and provides further constraints for the previous observations based on the geochemistry of the craton and supports results from the previous Hf isotope studies [9, 43, 54, 55]. Figure 12(a) shows the geochronology (U-Pb zircon ages as well as Nd and Hf model ages) of the Archaean TTG enclaves across the Bundelkhand craton from west to east in their key localities in Babina, Mauranipur, and Mahoba. The youngest Neoarchaean and oldest Paleoarchaean TTGs are found in the Babina area; however, most of the TTGs are Paleoarchaean (3.6-3.2 Ga). Figure 12(b) shows a similar diagram for sanukitoid and hybrid types of high-K granitoids as well as anatectic granitoids. The figure shows that abundant granitoids formed within a short time span and stabilised the Bundelkhand craton at the end of the Neoarchaean. The range in Hf model ages indicates an inheritance from older sources. For example, the Hf model age range for an anatectic granitoid from Mauranipur is as old as that of the oldest Mauranipur TTGs. This indicates that the high Hf model ages in the high-K granitoids were inherited from the Paleoarchaean TTGs, the sources of the anatectic granitoids. Some Paleoarchaean TTGs may have even inherited Eoarchaean Hf signatures. Overall, the Nd model ages follow the trend of the Hf model ages.

Combined geochronology, geochemical signatures, and Nd isotope compositions allow us to draw the following conclusions on the evolution of Archaean granitoids of the Bundelkhand craton:

• (1)

  The geochemical and Nd isotope signatures of Archaean Bundelkhand TTGs point towards partial melting of juvenile or short-lived mafic crust at different depths with contributions from newly formed felsic crust in the Paleoarchaean and older felsic crust in the Neoarchaean. This may reflect the transition from melting in Paleoarchaean oceanic plateaux (3.4-3.2 Ga) in plume settings to shallow melting in Neoarchaean island arcs (2.7 Ga) in subduction settings

• (2)

  The 2.5 Ga high-K granitoids formed at convergent subduction settings by partial melting of the mantle wedge and preexisting crust. Sanukitoids and hybrid granitoids originated in the mantle, the latter showing stronger crustal contributions, whereas abundant anatectic granitoids were the results of the melting of the continental crust

The data used in the manuscript is attached as tables.

The authors declare that they have no conflicts of interest.