Komatiites, greenstone volcanic rocks, and mafic dyke swarms are constituents of early earth magmatic activity, crucial for understanding the chemical evolution of the Archean mantle. The composition of the subcontinental lithospheric mantle (SCLM) is systematically modified throughout the Earth’s history by the addition of geochemically diverse oceanic and continental crustal materials through subduction and can be sampled through intraplate mafic/ultramafic volcanic activities. Here, we present a first report on the multiple sulfur isotope characteristics of the mafic dyke swarms and komatiites from the Dharwar craton in southern India and discuss the geochemical modifications of SCLM through crustal recycling. δ34SV-CDT values of the samples are all negative ranging from -0.15 to -2.91‰. Δ33S values for all the samples are close to 0 with the lowest value of -0.060‰ and highest of 0.146‰. Δ36S values are mostly negative with very few exceptions, ranging from -1.184 to 1.111‰. Near zero values of Δ33S and negative values for δ34S indicate an early formed mantle reservoir with a possible mixture of sulfur from subducting oceanic sediments. Together with trace element geochemistry, we suggest a depleted MORB source mantle (DMM) modified by oceanic crustal components and a depleted mantle (DM) modified by recycled continental crustal sediments as the two end members of the mantle source that produced mafic dyke swarms in the Late Archean to Proterozoic Dharwar craton.
Recycling of continental crust plays a major role in the evolution of subcontinental lithospheric mantle (SCLM), especially through late Archean to early Proterozoic [1–3]. The early Earth dynamics including the cooling of mantle and the onset of plate tectonics that modified the composition of mantle by subduction is understood through the limited Archean record . Gradual destruction of the early formed crust at the convergent boundaries has significantly altered the composition of the primitive mantle from Late Archean to Proterozoic. However, periodic emplacement of mantle materials in the form of mafic dyke swarms associated with large igneous provinces (LIPs) into the Archean crust has since long survived and is exposed in the Precambrian terranes all over the world . These, along with komatiites, ultramafic volcanic rocks that form the significant constituents of the Archean greenstone belts, are key to our understanding of the geochemical evolution of Archean mantle and early earth dynamics [6–8]. The geochemical heterogeneity of the mantle is primarily derived through the incorporation of geochemically distinct oceanic and continental crustal materials through subduction . This is especially significant since the sediments with signatures of ancient oceanic environment as well as those reflecting the Archean atmospheric chemistry are recycled into the Archean mantle and provide distinct isotopic characteristics. Prior to the great oxidation event (GOE) at about 2.3 Ga, the anoxic atmosphere led to the formation of mass-independent fractionation signature of sulfur (S-MIF) primarily produced by SO2 photolysis in low oxygenated conditions [10–13]. On the other hand, mass-dependent sulfur isotope fractionations (S-MDF) are found in older sediments which are attributed to the sulfate aerosol or microbial sulfate reduction (MSF) [14, 15]. The varying fractionation of sulfur isotope in continental and oceanic sediments remains unchanged and yields key information regarding the origin and diagenesis of sulfur compounds in the late Archean [16, 17]. Crustal recycling in the late Archean is thought to be rigorous and extensive leading to the modification of the SCLM. Components of the crustal materials influence the geochemical composition by mixing with the encompassing mantle following subduction and are eventually returned to the crust by various intrusive/extrusive events . However, the mass-dependent and independent signatures are well-preserved and are traceable by using multiple sulfur isotope geochemistry. Mafic dykes related to large igneous provinces and komatiites in the Archean terranes preserve large volumes of mantle derived materials, and sulfur isotope geochemistry can provide key information regarding the evolution of SCLM. The Dharwar craton of southern India is a well-preserved Archean craton where mafic dykes of various generations and greenstone belts containing ultramafic komatiites are exposed and preserve the evidence of large-scale volcanic activity in the Precambrian. Here, we consider the sulfur isotope composition of mafic dykes and komatiites from the Dharwar craton of southern India to address the modification of the Archean mantle. The multiple sulfur isotope analysis of these Archean to Proterozoic rocks helps us to understand the presence of recycled crust and the sulfur budget of the mantle beneath the Dharwar craton.
2. Geological Background
The Dharwar craton in southern India is one of the well-studied Archean cratons around the world and is mainly composed of crustal blocks that evolved during a period from c.3600 Ma to c.2500 Ma [18–21]. The craton had been traditionally divided into two crustal subcratons, namely, the Western Dharwar craton (WDC) and Eastern Dharwar craton (EDC) (Figure 1(a)). Western Dharwar craton, being the older crustal block, the long and substantial geologic history, is expressed as older TTGs, various mafic/ultramafic magmatic activities, and is characterized by the presence of abundant sediment record in the greenstone sequences as compared to the Eastern Dharwar craton [19, 22]. The shear zone marking the eastern margin of the Chitradurga greenstone belt is considered as the structural boundary between the two subcratons [23, 24]. Peucat et al.  divided the craton into three subblocks based on the age of the lithologic units, thermal records, and accretionary histories as western, central, and eastern blocks, separated by major shear zones. The craton is chiefly composed of tonalite-trondhjemite-granodiorite (TTG) type gneisses, volcano-sedimentary greenstone belts, and calc-alkaline to potassic granitoids ([19, 23, 24, 26] and the references therein). The western block is recognized as the oldest part of the craton dominated by the oldest basement rocks known as Peninsular gneisses (3450–3200 Ma) and extensive greenstone sequence of two generations (Sargur Group and Dharwar Supergroup) and high-potassic plutons. The central block consists of older TTGs (3230–2960 Ma) along with younger transitional TTGs (2700–2600 Ma), linear greenstone belts, and calc-alkaline granitoids [19, 27, 28]. In contrast to these two blocks, the eastern block is composed of younger transitional TTGs (2700–2500 Ma) and thin belts of greenstone sequences along with calc-alkaline plutons [19, 25]. The older greenstone sequence of the WDC, the Sargur Group (3400-3200 Ma), is dominated by komatiites and high Mg-basalts and alkaline basalts along with banded iron formation (BIF) indicating a shallow water environment of formation . The widespread komatiite occurrence at 3350 Ma and the subcontemporaneous basement TTGs are supposedly derived from combined plume-arc setting, and the geochemical characteristics suggests that the plume-derived komatiites are derived from depleted mantle reservoirs . The komatiite-high Mg basalt sequence preserves intraoceanic signatures and/or continental margin characteristics, whereas the alkaline basalts show evidence of subduction-related crustal recycling . The presence of a depleted mantle reservoir at 3350 Ma is attributed to early Archean crustal growth. Such a geochemical signature is also identified in the older mafic dykes that are exclusive to WDC which can be considered as the feeders of the older greenstone sequence [30, 31]. Several generations of mafic dykes ranging from Late Archean to Early Proterozoic are present all along the craton (Figure 1(b)). Also, those of the younger Eastern Dharwar craton are well-studied and categorized by several researchers ([32, 33, 34, 35] and the references therein). [32, 36] confirmed episodes of mafic magmatism in the Dharwar craton during 2368, 2220-2209, 2180, and 1891-1883 Ma. The 2368 Ma giant radiating Bangalore dyke swarm is the most extensive event in the craton, but with limited global extent, the 2220 Ma event is recognized worldwide in other Archean cratons . Younger generations of un-metamorphosed dykes that preserve pristine igneous textures and mineralogy  and older metamorphosed dykes with remnants of igneous textures and mineralogy are recognized in the Western Dharwar craton [30, 31, 38–40].
3. Sampling and Methodology
Mafic dykes of varying orientation and composition from both Western and Eastern Dharwar craton are considered in the current study. The dykes from the EDC are extensive and cross-cutting relations with basement gneisses, and supracrustal rocks can be identified in the field. On the other hand, most of the mafic dykes in the WDC are exposed as continuous hill or massive rocks without an exposed direct contact relationship with the country rocks, except for the one locality (2001D). From the EDC we have analyzed E-W trending 2368 Ma dyke, which is a part of the giant Bangalore dyke swarm, E-W trending 2365 Ma Pennukonda dyke, 2220 Ma NNW-SSE trending Kandlamadugu dyke [32, 41], N-S trending 2216 Ma dyke near Kunigal (southern end of Andhra Karnataka long dyke (AKLD) , and 1885 Ma Pulivendla sill . From the Western Dharwar craton NE-SW trending older meta-dolerite dykes (>2.7Ga?) , NW-SE trending younger dolerites (2.7Ga?–1.9Ga) [38, 39] are considered. In order to compare with the pristine mantle derived rocks in the Dharwar craton, komatiites associated with the Sargur Group greenstone belts are also considered.
For determining major element concentrations, pressed pellet sample preparation technique was adopted. ~5 g of representative portion of the powdered sample was pressed into a pellet using a 15-ton manual hydraulic press, and both qualitative and quantitative elemental concentrations were evaluated. X-ray fluorescence analysis was carried out by using ZSX Primus II instrument (Rigaku) housed at Research Institute for Natural Hazards and Disaster Recovery, Niigata University, equipped with 4 kW end-window X-ray tube and a rhodium anode.
Komatiites are exposed in WDC with typical ultramafic lava flow textures like spinifex and cumulate textures . The studied dykes show either typical ophitic to subophitic texture with plagioclase laths and clinopyroxene or large poikilitic plagioclase with clinopyroxene and rarely with orthopyroxene. The younger dolerite dykes in WDC show pristine texture and mineralogy , whereas the meta-dolerite dykes show remnant ophitic texture and pyroxene is mostly altered to amphibole . The sulfide mineral present in the dolerite dykes is mainly pyrite (Figure 1(c)), whereas meta-dolerite dykes are chiefly composed of pyrrhotite and chalcopyrite (Figure 1(d)).
The major element concentrations for 18 studied dyke samples and 4 komatiite samples are given in Table 1. Typically, these dykes are subalkaline tholeiitic in nature, and SiO2 content for all the samples varies between 49.1 and 55.8 wt% (Figure 2). Elemental sulfur concentration varies from 72.4 to 1312 ppm (see Table 1). Multiple sulfur isotope compositions for the studied samples are shown in Table 2. The δ34S values of the dykes are all negative ranging from -0.15 to -2.91‰ (Figure 3(a)). Meta-dolerite dykes from WDC are characterized by δ34S values ranging from -0.84 to -2.48‰; for younger dolerite dykes, δ34S values range from -1.08 to -2.91‰; and for the dykes from EDC, δ34S values range from -0.15 to -1.90‰. The δ34S values of the komatiites from WDC range from -0.70 to -1.42‰. Δ33S values for all the samples are close to 0 with the lowest value -0.060‰ and highest of 0.146‰. Meta-dolerites have Δ33S values ranging from -0.024 to 0.087‰, Δ33S values for younger dolerite dykes range from -0.060 to 0.146‰, all the dykes from EDC have positive values and range from 0.037 to 0.129‰, and for the komatiite, they range from 0.067 to 0.094‰. Δ36S values are mostly negative with very few exceptions, ranging from -1.184 to 1.111‰ (Figure 3(b)). Meta-dolerite dykes are characterized by negative Δ36S values ranging from -1.076 to -0.221‰; for younger dolerite dykes, Δ36S values range from -1.113 to 1.111‰; for the dykes from EDC, Δ36S values range from -1.184 to 0.177‰; and the komatiites are characterized by Δ36S values ranging from -0.988 to 0.073‰.
5.1. Source Characteristics
The geochemical characteristics of the mafic dykes and komatiites along with the sulfur isotope characteristics allow us to understand the nature of their source rock. Major element geochemistry of the dolerite dykes in EDC  suggested a subtholeiitic trend and stated that it is possible to discriminate the older 2368 Ma dykes and younger 2220-2180 dykes based on the total alkali and silica content as basaltic-andesite and basaltic, respectively. The rare earth element patterns are also described as swarm specific with many of the older dyke samples in EDC (2368 Ma) showing slightly incompatible element-enriched patterns on the primitive mantle normalized plot, with distinctive negative Nb anomalies and a slightly fractionated REE patterns. This is linked to the derivation from a mantle region previously influenced by subduction zone magmatism. The N-S trending 2216 Ma dyke displays enrichment of light rare earths indicating the origin from an enriched mantle source . The 1885 Ma Pulivendla sill is considered to be a part of the global-scale mantle upwelling or an enhanced mantle plume activity that is also recognized in other Archean cratons .
Geochemical evolution of subcontinental lithospheric mantle beneath the Dharwar craton suggests a slow transition from depleted mantle to a mantle that is enriched by the subduction of oceanic and continental-derived sediments [39, 40]. The depleted mantle source characteristics can be observed in the trace and rare earth element concentrations of the komatiites  and meta-dolerite dykes , and the gradual enrichment of the mantle is seen in the younger dolerite dykes from WDC  and in those of EDC  and the references therein (Figure 4). In the incompatible element ratio diagram of La/Nb vs. Ba/La and discrimination diagram of Sr/Nd vs. Nb/La, the samples are distributed in the primitive mantle source and trending towards continental crustal sediment input (Figures 5(a) and 5(b)). Although the younger dykes in the EDC show indications of crustal contamination, the older dykes in the WDC are thought to be derived from a metasomatized mantle that was modified by the addition of sediments or altered oceanic crust as early as in Paleoproterozoic [6, 40] and the references therein. In addition to the limited crustal contamination, the chalcophile element enrichment as seen from the Ni/MgO ratios and (Th/Yb)PM vs. (Cu/Zr)PM variation diagrams (Figures 5(c) and 5(d)) and the percentage of partial melting of mantle source also play a key role in the sulfur saturation of the source magma.
The geochemical characteristics of younger dolerites from WDC indicate the derivation from a source magma enriched by the subduction of continental crustal sediments and are geochemically coherent with the 2368 Ma and 2221 Ma dyke swarms of the EDC . On the other hand, the geochemical signature along with Sm-Nd isotopic characteristics of the older dolerites suggests that they are derived from a depleted mantle source modified by the addition of oceanic crustal sediments. The meta-dolerites in WDC show only a nominal LILE enrichment in primitive mantle-normalized multielement spidergram and flat or undepleted REE pattern in the chondrite normalized rare earth element geochemistry. The geochemical characteristics assign a more primitive mantle source for this suite of dykes. The meta-dolerites are considered to be older events and possibly the feeders for the early greenstone volcanism recorded in WDC . Komatiite and komatiitic basalts with typical ultramafic lava flow textures like spinifex and cumulate textures are exposed in the greenstone belts of WDC . Even though these are affected by greenschist to lower amphibolite facies metamorphism and rarely preserve primary olivine and pyroxene, they constrain the composition of Archean mantle. The trace and rare earth element characteristics together with the Nd isotopic characteristics suggest their derivation from a depleted mantle. The existence of such depleted mantle in the early Archean that is attributed to the preceded continental crustal growth makes them an ideal reference to understand the modification of mantle in the late Archean to early Proterozoic.
5.2. Multiple Sulfur Isotope Characteristics
Multiple sulfur isotope signatures in Archean are mainly associated with biological, photochemical, or hydrothermal processes . Mass-independent fractionation (MIF) signatures of sulfur are extensively generated until ~2.4 Ga mainly through photochemical reactions  and the references therein and are preserved during the recycling of ancient crust  and the references therein. The photochemical signatures reflected in the sulfur isotopes can be overprinted by late Archean microbial sulfur cycles. Such overprinting dilutes the atmospheric mass-independent signatures and leads to mass-dependent isotope fractionation (MDF) . The samples in the current study show near zero values of Δ33S ranging from -0.001 to 0.146‰. Δ33S ≈0 is linked to a deep primordial component ; however, variation and magnitude of Δ33S are attributed to the mixing with Archean and Proterozoic sediments [51, 52]. Labidi et al.  suggested that the mantle components that record subduction processes primarily reflect the sediment recycling. Negative values of Δ33S and a negative slope of Δ33S/Δ36S observed in certain samples can be linked to the sulfur aerosols, whereas the high positive values exhibit a clear mass-independent fractionation characteristic probably due to biogeochemical processes [14, 45, 53]. This along with the negative values for δ34S indicates an early formed mantle reservoir with a possible mixture of sulfur from subducting oceanic sediments from Archean to Proterozoic. Labidi et al.  stated that the mean MORB value is -0.91 ± 0.50‰ and contrasts with the modern sulfur isotopic mantle values ranging between ~0‰ and 4‰ as given by  which is a result of homogenization by hydrothermal fluids or posttectonic isotopic exchange [56, 57]. The Sm-Nd isotopic evolution of the komatiites and dykes from WDC and EDC suggests a systematic enrichment of SCLM beneath the Dharwar craton. The komatiites are formed from a depleted mantle reservoir after the continental crust building process , and the late Archean to early Proterozoic mafic dykes indicate the derivation from a mantle source enriched by the subduction of the mafic oceanic crust in the late Archean to a mantle enriched by continent-derived sediments towards the early Proterozoic [39, 58]. The distinct negative δ34S values obtained for the current samples also reaffirm the early Earth’s mantle reservoir signatures. The results from the current study are compared with the previous results  from the volcano-sedimentary succession of Dharwar craton (Figure 6) to substantiate the similar characteristics in younger mantle-generated events like mafic dykes. Mishima et al.  suggested that the early Archean (3.0-2.7 Ga) sedimentary rocks exhibit only weak mass-independent fractionation (MIF) signatures or even mass-dependent (MDF) signatures as compared to the MIF dominance of those formed between 2.7 to 2.5 Ga. This variation of the Δ33S and Δ36S in the Archean is attributed to the changing atmospheric chemistry. Ono et al.  and others suggested that the photolysis in an anoxic environment in the early Archean produced two isotopically distinct sulfur aerosols (positive and negative Δ33S) which were later incorporated into the surface sediments. Although the current samples show both negative and positive Δ33S values, the lack of significant MIF signatures can be explained by the overprinted microbial sulfate reduction signatures. The narrow variation of δ34S of the Archean sediments from -1.26 to 2.89‰ is likely to be produced by the microbial sulfate reduction, which was introduced into the ocean floor that suffered later subduction activities . This again leads to the possibility of recycling of sediments with varying isotopic signatures. Archean sulfur biogeochemical records from these sedimentary rocks exhibit systematic changes in the global atmospheric signatures while serving as a tracer to understanding the volcanic activities that operated before and after the great oxidation event.
5.3. Implications for Mantle Dynamics and Crustal Recycling
Mafic dykes cool rapidly and preserve the composition of the source mantle. Being very resistant to posttectonic weathering and erosional modifications, they represent the characteristics of the mantle reservoir beneath the Archean craton . The variation in the sulfur isotope characteristics is thought to be due to the addition of recycled oceanic crust and continental crustal sediments that has distinct mass-independent and mass-dependent sulfur isotopic signatures due to their process of formation. Cabral  suggested that the complementary positive and negative Δ33S can originate from sediments incorporating the atmospheric sulfur and an ancient crustal source with oceanic affinities, respectively. This also supports the origin of such mixed isotopic signatures in the studied samples from the aforementioned mantle reservoirs. Based on the observed positive and negative Δ33S pools of the current samples that are attributed to the incorporation of Archean atmospheric sulfur or the microbial activity at the ocean floor, and the Sm-Nd isotopic compositions [6, 39, 40], we suggest a depleted MORB source mantle (DMM) + recycled oceanic crustal components and a depleted mantle (DM) + recycled continental crust sediments as the two end members.
The current study substantiates our understanding about the preservation of atmospheric and biogenic signatures of sulfur in the continental and oceanic crustal sediments, their preservation while subduction, and incorporation into the mantle reservoir before its return to the crust in the form of mantle-derived rocks. The tectonic evolution of Dharwar craton  suggests that the WDC is older part of the craton composed of TTG-greenstone sequences of varying compositions, indicating significant changes in the mantle composition from 3450-2600 Ma. Subduction of an oceanic slab and its subsequent melting is proposed as the magmatic protolith to TTG in WDC . EDC was tectonically amalgamated and stabilized at around 2500 Ma, and the related accretion and subduction further modified the SCLM beneath the Dharwar craton. The meta-dolerite dykes and komatiites in the current study are formed from a depleted mantle source, and older mafic dyke swarms of the WDC (2.7 Ga) are formed from a mantle source that are enriched by the oceanic crustal sediments leading to the formation of dolerite dykes of different geochemical characteristics. The younger dykes of the WDC and those of EDC are compositionally enriched and are thought to have formed due to a long-lived plume activity or a large igneous province [32, 35]. Delavault et al.  suggest that the source material for mantle plume volcanism shows geochemical characteristics of former crust and/or sediments recycled into the mantle by subduction. The sulfur isotope variation in the mafic dykes of EDC might have been caused by the mantle plume that carried recycled crustal components in its source.
The multiple sulfur isotope characteristics of the mafic dykes and komatiites give evidences of crustal recycling and the geochemical evolution of the SCLM beneath the Archean Dharwar craton. The variations in the sulfur isotope signatures are explained by a combination of mass-independent fractionation (MIF) signatures and mass-dependent (MDF) signatures in the late Archean. Further geochemical characterization using systematic radiogenic isotopes is necessary for the detailed investigation on the recycled crustal materials and the identification of mixing end members.
All data used in this research work is included in the manuscript.
A.S. Silpa’s present address is the Department of Earth Sciences, Shimane University, Shimane, 690-8504, Japan.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
The authors thank the guest editor for an efficient editorial handling of the manuscript and two anonymous reviewers for their valuable comments which helped in improving the manuscript to a large extent. SAS acknowledges Japanese Government (Monbukagakusho) scholarship for PhD program at Niigata University. Bulk rock geochemical data was obtained using the XRF facility at Research Institute for Natural Hazards and Disaster Recovery, Niigata University; the authors thank K. S. Kataoka for the support during the analysis. This study was supported by the Grant-in-Aid for Scientific Research on Innovative Areas. M.S-K. acknowledges the partial support for field work through JSPS KAKENHI grant numbers JP23340155 and JP25302008 and analytical support through JSPS KAKENHI grant numbers JP15H05831 and 20KK0081. The authors thank the Japan-India Science Cooperative Programs, lead by T. Toyoshima 2019-2020 and T. Hokada 2013-2014, for partial support in the early stages of field studies in the Dharwar craton. The authors thank Yuichiro Ueno and Shinnosuke Aoyama for the advice and for the help in building the sulfur isotope analytical system at Niigata University. K. Sajeev and S. Kiran for the help rendered during field studies in Dharwar craton.