There is a paucity of evidence preserved in the rock record regarding Earth’s earliest enriched crust and its complementary depleted mantle during the Hadean. In recent years, vestiges of these early reservoirs have been inferred by examination of Hf isotope systematics compiled from zircons. The Singhbhum craton of Eastern India, for example, preserves only the existence of an enriched (εHf <0) crustal reservoir during the Hadean–Eoarchean, with the notable absence of a depleted mantle reservoir signature (εHf >0) until ca. 3.5 Ga. Here we report a new Sm-Nd isochron for the Lower Lava greenstones of the western Iron Ore Group from the Singhbhum craton, confirming a 3.42 ± 0.14 Ga crystallization age with an initial εNd of +5.7 ± 2.5. This is the highest positive εNd value derived from an isochron of this age. We infer that this depleted mantle source is a vestige complementary to the primary crust following planetary differentiation. Furthermore, we present U-Pb zircon ages for a 3.39 ± 0.02 Ga tuff that lies stratigraphically above the Lower Lava and <30 cm below an extensive conformable banded iron formation (BIF). This age implies that the western Iron Ore Group’s BIF is the largest economic-grade iron formation of its Paleoarchean age, suggesting that free atmospheric oxygen existed as more than just whiffs at this time.

While there is evidence to suggest that Earth differentiated into an enriched crust and a depleted mantle reservoir early (>4.5 Ga; e.g., Harper and Jacobsen, 1992), the fate of these complementary reservoirs and the role they have played in the evolution of Earth remains a subject of great interest. Over the decades, several studies have reported 142Nd anomalies, with some authors proposing the earliest enriched crust became reworked to form cratons (O’Neil and Carlson, 2017) while also possibly remaining stored at the base of the lower mantle (Carlson and Boyet, 2008). However, although 143Nd systematics provides traces for an extremely depleted mantle reservoir in the Archean (Collerson et al., 1991), there are few proposals for the potential survival of a >4.5 Ga depleted mantle source (e.g., Jackson et al., 2010; Caro and Bourdon, 2010). The inaccessibility of Earth’s earliest depleted mantle, the lack of well-preserved mantle-derived Archean rocks, and overprinting by continuous surface recycling have all contributed to our limited understanding of these ancient reservoirs (e.g., Armstrong, 1991), although recently, the most ancient accessible reservoirs of Earth’s mantle have been implicated in studies in basalts of Baffin Island (northern Canada) and West Greenland (see Caro and Bourdon, 2010; Jackson et al., 2010).

The purpose of this study is to apply the 147Sm-143Nd isotopic system (based on the long-lived decay of 147Sm to 143Nd, decay constant λSm = 6.54 × 10−12 yr−1) on well-preserved greenstones from the western Iron Ore Group greenstone belt, Eastern India, in order to establish their age as well as to evaluate the significance of their initial 143Nd/144Nd ratios with respect to the evolution of the depleted mantle. Located in the Singhbhum craton, the western Iron Ore Group is part of a regional (55 × 35 km) NNE-plunging asymmetric synclinorium structure in the Jamda-Koira Valley (Fig. 1; Fig. S1 in the Supplemental Material1). Greenstones make up two significant lithologies within the western Iron Ore Group stratigraphy, with the lowermost stratum designated as the “Lower Lava” and the uppermost stratum the “Upper Lava” (Fig. S2). Together, these greenstones encompass an apparent 8-km-thick volcano-sedimentary and economic-grade banded iron formation (BIF)–bearing greenstone belt succession (e.g., Basu et al., 2008). Samples were collected from both the eastern and western limbs of the Lower Lava western Iron Ore Group synclinorium as well as the overlying Upper Lava deposited on an unconformity (Fig. 1; Fig. S2). All the western Iron Ore Group greenstones are characterized by their low-grade greenschist facies (quartz + albite + chlorite), primary igneous augite and pigeonite, lack of penetrative deformation, preservation of original volcanic textures (i.e., hydrothermal metamorphism), dominant calc-alkaline affinities with subordinate tholeiite, massive-pillowed morphologies, and average basaltic-andesite composition.

Here we report the western Iron Ore Group Lower Lava to preserve the most-depleted initial ε143Nd signature derived from a Paleoarchean isochron. We also present new U-Pb zircon geochronology for a stratigraphically overlying tuff whose age agrees with the Lower Lavas isochron age. Lastly, we examine the significance of the new 143Nd results presented in this study within the evolutionary context of Singhbhum’s Hfzircon isotopic record.

The data used to construct the Sm-Nd isochrons for both the Lower and Upper Lava units are provided in Table 1. See the Supplemental Material for the isotopic methodology. The Lower Lava of the eastern and western limbs defines a ten-sample whole-rock isochron registering an age of 3420 ± 140 Ma and an initial 143Nd/144Nd ratio of 0.50848 ± 0.00013 (initial εNd = +5.7 ± 2.5), with a low value of 0.98 for the mean square of weighted deviates (MSWD) (Fig. 2A). Sm/Nd ratios for these lavas are subchondritic, which is a result of the bulk rock being light rare earth element (LREE) enriched—possibly from fractionation during melting of the mantle source. A relatively small range of 147Sm/144Nd from 0.12 to 0.17 (Table 1) for the Lower Lava derives from the variable amounts of feldspar, altered glass, and primary clinopyroxene, and a few western limb samples having minor secondary amphibole.

Precise secondary ion microprobe U-Pb ages of 22 zircons (grains ~350 µm long) were measured from a tuff unit that lies above the Lower Lava but just 30 cm conformably below the BIF (Fig. 1; Fig. S2; sample 4/03). The concordant 3392 ± 29 Ma tuff age (Fig. 2B; Table S2 [see footnote 1]) confirms the BIFs Paleoarchean antiquity (see Basu et al., 2008), stratigraphically agreeing with the Lower Lava’s 3.42 Ga age. This tuff age is remarkable in being close to that of the BIF that exceeds a 220-m thickness in the type area with single ore bodies as much as 3 km long along strike and several hundred meters wide. This makes the western Iron Ore Group BIF the largest for its age (~5 × 1010 tons), and it remains a significant economic-grade iron-ore deposit (>60 wt% Fe2O3; Beukes et al., 2008).

In addition to the Lower Lava isochron age agreeing with the tuff age, it is also consistent with the age of the younger Bonai granite, ca. 3.37 Ga (see Asokan et al., 2021), which has an intrusive relationship with the western limb of the western Iron Ore Group syncline (Fig. 1). The 3.42 ± 0.14 Ga age of the Lower Lava Sm-Nd isochron and the U-Pb zircon tuff age place the antiquity of the western Iron Ore Group into the same emplacement timeline as the 3.51 Ga U-Pb zircon age for the southern Iron Ore Group (Mukhopadhyay et al., 2008) and eastern Iron Ore Group (Jodder et al., 2021) greenstone belts in Singhbhum (Fig. S1). The Upper Lava greenstones, collected on a 4 km traverse, lie on a five-sample whole-rock isochron that yields an age of 2654 ± 104 Ma and an initial 143Nd/144Nd ratio of 0.509332 ± 0.000097 (initial εNd = +2.7 ± 1.9; Fig. S3). This isochron also displays a small range of Sm/Nd ratios from 0.12 to 0.17 (Table 1) with a MSWD of 0.57.

The mantle source for the Lower Lava greenstones represents the highest inferred time-integrated 143Nd/144Nd ratio (with respect to chondrite) of any Archean suite reported to date, using the “real age” isochron technique and not individually calculated model εNd values using other chronometers for the age at the time of its initial value (Fig. 2A). Assuming planetary fractionation occurred at 4.5 Ga, the early depleted mantle reservoir of the Lower Lavas was derived from an initial εNd value of +5.7 at 3.42 Ga that has an estimated present-day 143Nd/144Nd = 0.513817 and 147Sm/144Nd = 0.236 ratio that projects to a modern-day εNd of +23 ± 10. Straddling this evolution line are ultramafics from Labrador (northern Canada) (Collerson et al., 1991) and West Greenland (van de Löcht et al., 2020) that depict the presence of an existing terrestrial mantle reservoir with a similar Sm-Nd fractionation history. Given that a younger mantle separation age requires the existence of increasingly higher present-day εNd values, we propose εNd = +23 to be more reasonable as observed by our present data. Thus, the Lower Lava’s initial εNd value (+5.7) argues for having been derived from a long-lived (>1 Ga) mantle source that evolved within a closed system. This observation is consistent with previous studies that reported Nd isotopic data supporting the presence of chemical heterogeneities in Earth’s early mantle that persisted for at least the first billion years of Earth history, as well as highly depleted mantle reservoirs (e.g., Bennett et al., 1993, 2007; Hoffmann et al., 2010).

It has been problematic to compare the depleted ε143Nd signature reported from the Singhbhum craton by other studies to that recorded by the western Iron Ore Group Lower Lava in the current study. This is because these previously reported ages and initial εNd values were either not derived from an isochron (e.g., Chaudhuri et al., 2017; Pandey et al., 2019; Asokan et al., 2023), derived from an errorchron (e.g., Pandey et al., 2019; Maltese et al., 2022), from samples significantly affected by crustal contamination (e.g., Adhikari 2021a, 2021b; Chaudhuri et al., 2017), or associated with an isochron-derived age that does not agree with other independent stratigraphic chronometers (e.g., Basu et al., 1981; Adhikari et al., 2021b; Adhikari and Vadlamani, 2022). It is important to specify that the assumed initial εNd value of +5.2 for a single Singhbhum granitic sample accepted by Pandey et al. (2019) and reused by Maltese et al. (2022) was calculated not from a reliable isochron but from using a single U-Pb zircon age whose 147Sm-143Nd data also formed an errorchron (very high MSWD) according to Pandey et al. (2019). The results from these studies suggest the samples were not cogenetic and have not remained in a closed system, or they were the result of a mixing line between two or more temporary distinct magmatic events. The strength of the current study is that the initial εNd value (+5.7) and age (3.42 Ga) of the Lower Lava are derived from a well-constrained isochron (closed-system behavior) whose age agrees with two independent stratigraphic chronometers. Plotting 143Nd/144Nd versus 1/Nd (ppm) for both the Lower and Upper Lava does not exhibit correlation, suggesting the Lower Lava and Upper Lava isochrons are not mixing lines. Lastly, the calculated initial εNd values for the individual samples of both the Lower Lava (±0.3) and Upper Lava (±0.1) exhibit low deviations from their isochron-derived initial εNd values, indicating their cogenetic derivation.

Hafnium isotopes can provide an independent examination for the plausibility of the Nd data given that the behavior of the 147Sm-143Nd system parallels that of the 176Lu-176Hf system. The current Hfzircon isotopic record from the Singhbhum craton (Fig. 3B) suggests that an enriched primary crust (implied from the presence of only negative εHf values) evolved not from the addition of juvenile magma from the mantle but rather from extensive reworking of an older existing enriched crust between the Hadean (4.2 Ga) to the Eoarchean (ca. 3.5 Ga) (e.g., Bauer et al., 2020). This older Hadean enriched crust suggests the existence of an isolated complementary depleted mantle reservoir equivalent in age. The simultaneous occurrence of both positive and negative εHf values abruptly beginning at ca. 3.5 Ga signals the sustained emergence of the addition of a juvenile mantle signature (e.g., Sreenivas et al., 2019). This excursion of positive εHf values coincides with the 3.42 ± 0.14 Ga Sm-Nd whole-rock emplacement age of the Lower Lava greenstones. Therefore, the 1-billion-year interval from 4.5 to 3.5 Ga provides a reasonable geological time scale for the closed-system ingrowth of 143Nd to have occurred in order to have imparted their distinctly positive initial (εNd = +5.7) value (Fig. 1A). An early Earth differentiation event for the Singhbhum craton is consistent with the inferred 4.2–4.5 Ga separation model ages (Chaudhuri et al., 2018; Maltese et al., 2022). The emergence of a zircon Hf juvenile signature at 3.5 Ga, with many samples having initial εHf >+6 above the proposed Hf evolution curve for the depleted mantle (Fig. 3B), may indicate that the highly depleted nature of the mantle as recorded by Nd isotopes from the western Iron Ore Group Lower Lava greenstones and Hfzircon isotopes is a characteristic of the Singhbhum craton preserved in both the Nd and Hf isotopic systems.

The difference of three ε143Nd units over the ~770 m.y. period between the emplacement of the Lower Lava (5.7 at 3.42 Ga) and that of the Upper Lava (2.7 at 2.65 Ga) could be interpreted to reflect either the sampling of different mantle sources or the progressive recycling of LREE-enriched material into the mantle source (e.g., Frost et al., 2023). If the latter is correct, then the ε143Nd became suppressed, preventing the growth of increasingly positive values. Such time-integrated closed system behavior is observed by many mantle-derived samples from other silicate bodies (e.g., Moon and Mars; Borg et al., 2011; Lapen et al.,2017) that lack an active surface recycling plate tectonic regime (e.g., Armstrong, 1991; Bowering and Housh, 1995). The difference between the ε143Nd values of the inferred present-day western Iron Ore Group depleted mantle source for the Lower Lava (+23) and those of present-day mid-ocean-ridge basalts (MORBs) (+10) suggests that crustal recycling has suppressed Earth’s mantle by a minimum of 13 εNd units over 3.4 b.y. of geologic time. However, this assumes that the depletion recorded in the Singhbhum craton is representative of a global phenomenon that is a mantle vestige of the early differentiation of the primary crust.

We obtained a 3.42 Ga Sm-Nd isochron age for the Lower Lava, which agrees with the 3.39 Ga zircon U-Pb age obtained for the stratigraphically overlying tuff. The Lower Lava’s initial 143Nd suggests derivation from a source that represents one of the best-preserved examples of early Earth’s Hadean LREE-depleted mantle reservoir, whose 143Nd evolution mirrors that of the 176Hf-enriched crust as preserved in Singhbhum’s zircon record. The new Nd isochrons reported for the Lower Lava and the unconformable Upper Lava bracket the depositional age of the western Iron Ore Group basin between 3.42 and 2.65 Ga. Lastly, the 3.39 Ga zircon tuff age immediately below the conformably overlying huge BIF deposit indicates large variations of free atmospheric oxygen during the Paleoarchean.

1Supplemental Material. Figures S1–S4, Tables S1–S5, and materials and methods. Please visit https://doi.org/10.1130/GEOL.S.25361074 to access the supplemental material; contact editing@geosociety.org with any questions.

We thank Mouhcine Gannoun for measuring the Sm-Nd isotopic compositions. Basu collected the western Iron Ore Group samples in the field with help from P.K. Bandyopadhyay of Presidency College, Kolkata, and D.K. Bose of the iron ore mines of Bihar-Orissa. H. Zou aided in the U-Pb data acquisition of the zircons from the tuff at the University of California, Los Angeles (UCLA). R. Chakrabarti helped in final picking of these zircons at the University of Rochester (New York). We thank Robert T. Gregory at SMU for providing manuscript edits. Wright received support from the Roy M. Huffington Scholarship account, the Stable Isotope Laboratory of Southern Methodist University, the Clifford W. Matthews chair account, the American Federation of Mineralogical Societies, and the Society of Independent Professional Earth Scientists. Basu partially supported this study while at the University of Rochester for field sampling and U-Pb zircon analyses with the UCLA ion probe. We thank Rob Strachan for the handling of this manuscript and the anonymous reviewers for their valuable comments that improved this manuscript.

Gold Open Access: This paper is published under the terms of the CC-BY license.