Tim Elliott

Voluminous felsic continental crust is, as far as we know, unique to Earth, yet the timescales of its earliest growth remain debated. Archean mantle evolution is complementary to crustal growth but is largely unconstrained for strontium isotopes due widespread Rb-Sr disturbance. Here, we perform high spatial resolution radiogenic Sr and Ca isotope measurements in magmatic plagioclase megacrysts from 3.7–2.8 Ga Archean anorthosites and leucogabbros. We report mantle-like Ca isotope signatures and the most unradiogenic terrestrial ⁸⁷ Sr/ ⁸⁶ Sr ratio yet measured on Earth, in plagioclase ( ⁸⁷ Sr/ ⁸⁶ Sr _initial = 0.700050 ± 0.000017, 95% confidence interval) from the 3.73 Ga Manfred Complex of the Narryer Terrane, Western Australia. These data are consistent with ⁸⁷ Sr/ ⁸⁶ Sr homogenisation between the Earth and Moon ( ⁸⁷ Sr/ ⁸⁶ Sr _initial ≈ 0.699061) at ca. 4.515 Ga during the Moon-forming giant impact. The appearance of a depleted mantle signature in the terrestrial strontium record post-3.5 Ga suggests that extensive continental growth began relatively late.

Most mid-ocean ridge basalts (MORB) are depleted in highly incompatible elements relative to the primitive mantle and are termed normal (N)-MORB. Some MORB, erupted at ridge segments distal from mantle hot-spots, are enriched in incompatible elements. The origin of these enriched (E)-MORB is debated, although many studies have proposed that recycled oceanic crust shapes their compositions. Uranium (U) and molybdenum (Mo) isotope ratios have been argued to trace the contribution of recycled oceanic crust in the source of N-MORB, which has high δ238U and low δ98/95Mo relative to the bulk silicate Earth (BSE). Here, we provide U and Mo isotopic data on E-MORB samples from the northern mid-Atlantic ridge (13° & 45° N). We analysed hand-picked, leached MORB glass, yielding 234U/238U near secular equilibrium, therefore reflecting samples unperturbed by surface processes. Samples have uniform δ238U and δ98/95Mo, with means of −0.307 ± 0.032 ‰, 2sd, and −0.14 ± 0.04 ‰, 2sd, respectively, both within uncertainty of BSE, and distinct from N-MORB. These data, as well as unremarkable Ce/Pb and radiogenic Pb isotopic compositions in E-MORB globally, are incompatible with their sources containing recycled oceanic crust or continental derived sediments. Instead, our data fit with a model of low degree partial melting of the uppermost mantle that metasomatises the sub-oceanic lithosphere. Given BSE-like U isotopic compositions of E-MORB, that are isotopically unfractionated during low degree partial melting, we suggest that the initial melting event must have occurred prior to the recycling of isotopically distinct in U oceanic crust into the upper mantle (i.e., prior to ca. 600 Ma, the estimated time of deep ocean oxygenation). Metasomatised portions of oceanic lithospheric mantle preserve these ≥600Ma U isotopic compositions, which are subducted and stirred back into the convecting upper mantle, ultimately to be sampled at ridges as E-MORB. Molybdenum isotopic compositions of E-MORB are in line with such a model but also reflect isotopic fractionation to higher δ98/95Mo during low degree partial melting of ≥600 Ma upper mantle, that counter acts the lowering of δ98/95Mo in the upper mantle by an on-going process of plate recycling.

The concentration of dissolved oxygen in the deep oceans has varied over Earth History, but the timing of the transition from anoxic to oxic deep oceans is debated. Under modern-day, oxic, deep ocean conditions, alteration of the upper sections of mafic oceanic crust with U-rich seawater leads to U enrichment, low Th/U ratios, and heterogeneous 238U/235U ratios relative to fresh mid-ocean ridge basalt (MORB). Given the solubility behaviour of U, its uptake into altered oceanic crust (AOC) is expected to be smaller and less isotopically fractionated when deep oceans were anoxic and thus U-poor. Determining when, in the geological record, the U elemental and isotopic systematics of ancient oceanic crust first resemble modern day AOC should indicate when deep oceans became oxic. We provide U concentration, Th/U, and U isotopic data on upper-crustal sections of three ophiolites from 750 to 480 Ma, spanning the period inferred for deep ocean oxygenation (~ 850 to 400 Ma). The ophiolites at 480 and 540 Ma have high U contents, low Th/U ratios, and variability in 238U/235U ratios like modern-day AOC, reflecting seawater alteration of oceanic crust under oxygenated seawater conditions. In contrast, the 750 Ma ophiolite does not show the distinctive decreasing Th/U with increasing U concentrations trend of modern AOC and has fewer samples with 238U/235U ratios perturbed from mantle values, reflecting alteration under largely anoxic deep ocean conditions. This is also supported by Fe3+/FeT ratios in these samples that are like unaltered modern MORB. Thus, our data suggest oxygenated deep oceans at some time between 750 and 540 Ma, either reflecting a full transition or intermittent deep ocean oxygenation events within an otherwise anoxic deep ocean.

The isotopic composition of lavas associated with mantle plumes has previously been interpreted in the light of core–mantle interaction, suggesting that mantle plumes may transport core material to Earth’s surface^1–5. However, a definitive fingerprint of Earth’s core in the mantle remains unconfirmed. Precious metals, such as ruthenium (Ru), are highly concentrated in the metallic core but extremely depleted in the silicate mantle. Recently discovered mass-independent Ru isotope variations (ε ¹⁰⁰Ru) in ancient rocks show that the Ru isotope composition of accreted material changed during later stages of Earth’s growth⁶, indicating that the core and mantle must have different Ru isotope compositions. This illustrates the potential of Ru isotopes as a new tracer for core–mantle interaction. Here we report Ru isotope anomalies for ocean island basalts. Basalts from Hawaii have higher ε ¹⁰⁰Ru than the ambient mantle. Combined with unradiogenic tungsten (W) isotope ratios, this is diagnostic of a core contribution to their mantle sources. The combined Ru and W isotope systematics of Hawaiian basalts are best explained by simple core entrainment but addition of core-derived oxide minerals at the core–mantle boundary is a possibility.

The prevailing view of the formation of porphyry copper deposits along convergent plate boundaries involves deep crustal differentiation of metal-bearing juvenile magmas derived from the mantle wedge above a subduction zone. However, many major porphyry districts formed during periods of flat-slab subduction when the mantle wedge would have been reduced or absent, leaving the source of the ore-forming magmas unclear. Here we use geochronology and thermobarometry to investigate deep crustal processes during the genesis of the Late Cretaceous–Palaeocene Laramide Porphyry Province in Arizona, which formed during flat-slab subduction of the Farallon Plate beneath North America. We show that the isotopic signatures of Laramide granitic rocks are consistent with a Proterozoic crustal source that was potentially pre-enriched in copper. This source underwent water-fluxed melting between 73 and 60 Ma, coincident with the peak of granitic magmatism (78–50 Ma), porphyry genesis (73–56 Ma) and flat-slab subduction (70–40 Ma). To explain the formation of the Laramide Porphyry Province, we propose that volatiles derived from the leading edge of the Farallon flat slab promoted melting of both mafic and felsic pre-enriched lower crust, without requiring extensive magmatic or metallogenic input from the mantle wedge. Other convergent plate boundaries with flat-slab regimes may undergo a similar mechanism of volatile-mediated lower-crustal melting.

The naturally occurring stable isotopes of potassium (⁴¹K/³⁹K, expressed as δ⁴¹K) have the potential to make significant contributions to vertebrate and human biology. The utility of K stable isotopes is however conditioned by the understanding of the dietary and biological factors controlling natural variability of δ⁴¹K. This paper reports a systematic study of K isotopes in extant terrestrial endothermic vertebrates. δ⁴¹K has been measured in 158 samples of tissues, biofluids and excreta from 40 individuals of four vertebrate species (rat, guinea pig, pig and quail) reared in two controlled feeding experiments. We show that biological processing of K by endothermic vertebrates produces remarkable intra-organism δ⁴¹K variations of ca. 1.6 ‰. Dietary δ⁴¹K is the primary control of interindividual variability and δ⁴¹K of bodily K is +0.5 to 0.6 ‰ higher than diet. Such a trophic isotope effect is expected to propagate throughout trophic chains, opening promising use for reconstructing dietary behaviours in vertebrate ecosystems. In individuals, cellular δ⁴¹K is related to the intensity of K cycling and effectors of K homeostasis, including plasma membrane permeability and electrical potential. Renal and intestinal transepithelial transports also control fractionation of K isotopes. Using a box-modelling approach, we establish a first model of K isotope homeostasis. We predict a strong sensitivity of δ⁴¹K to variations of intracellular and renal K cycling in normal and pathological contexts. Thus, K isotopes constitute a promising tool for the study of K dyshomeostasis.