Heavy $\delta$⁵⁷Fe in ocean island basalts: A non-unique signature of processes and source lithologies in the mantle

Abstract

Lithological heterogeneity is a widely accepted feature of the Earth’s mantle, with recycled crustal material accounting for a significant part of heterogeneity in ocean island basalt (OIB) geochemistry. Fe isotopes have been used to link geochemical heterogeneity in OIB sources to distinct mantle lithologies due to their mineral-specific equilibrium fractionation effects, or their composition, such as incorporation of kinetically-fractionated core liquids entrained from the core-mantle boundary.

Here we present Fe isotope data for Samoan shield, and Azores volcanoes, together with a combined phase-equilibria and equilibrium mineral-melt isotope fractionation model. These OIB lavas allow us to study the roles of core-derived and recycled mantle components in generating heavy $\delta$⁵⁷Fe melts. Heavy $\delta$⁵⁷Fe correlates with radiogenic isotope signatures of enrichment by a crustal component and not with Fe/Mn or any indicator of core involvement. However, single-stage melting of a MORB-like eclogitic pyroxenite cannot generate the heavy $\delta$⁵⁷Fe seen in Pitcairn, Azores, and rejuvenated Samoa lavas. Melts of a reaction-zone pyroxenite (commonly suggested to form part of the OIB source), derived from eclogite melts hybridised with peridotite, also fail to generate the heaviest Fe isotopic compositions seen in OIB. Instead, the generation of heavy $\delta$⁵⁷Fe melts in OIB requires: (1) processes that make subducted eclogite isotopically heavier than its pristine precursor MORB (e.g., hydrothermal alteration, metamorphism, sediment input); (2) lithospheric processing, such as remobilisation of previously frozen small-degree melts, or a contribution from lithospheric material metasomatised by silicate melts; and/or (3) melting conditions that limit the dilution of melts with heavy $\delta$⁵⁷Fe by ambient lower $\delta$⁵⁷Fe materials. No single process we consider can generate the heavy $\delta$⁵⁷Fe seen in the Azores, Pitcairn, and rejuvenated Samoan lavas.

Therefore, it cannot be assumed that a pyroxenite lithology derived from recycled crustal material is the sole producer of heavy $\delta$⁵⁷Fe melts in OIB, nor can these signatures be related to contributions from the Earth’s core. Instead, the observation of heavy $\delta$⁵⁷Fe OIB melts cannot be ascribed to a unique source or process. This ambiguity reflects the multitude of processes operating from the generation of recycled lithologies through to their mantle melting: from MORB generation, its low temperature alteration, through mantle heterogeneity development and lithospheric processing, to eruption at ocean islands.

Introduction

Numerous isotopic and trace element studies of ocean island basalts (OIB) suggest that the Earth’s mantle is heterogeneous on scales from hundreds of kilometres, across entire mantle plume systems, to less than a kilometre or metres, as recorded by melt inclusions (e.g., Zindler and Hart, 1986, Weaver, 1991, Hofmann, 1997, Stracke et al., 2005, Maclennan, 2008). The variance in long-lived radiogenic isotopic composition (Sr-Nd-Pb) in most OIB can be explained by mixing between five primary components: depleted mantle, the prevalent mantle (PREMA)/focal zone (FOZO) component, two enriched mantle (EM1/EM2) components and a HIMU component (Stracke, 2012; Fig. 1). Radiogenic isotopes have also been used to argue for enriched (and depleted) components in the mid-ocean ridge basalt source (e.g., Hirschmann and Stolper, 1996, Salters and Dick, 2002, Liu et al., 2008, Byerly and Lassiter, 2014). In addition to these components, osmium isotopes (Walker et al., 1995, Brandon et al., 1998), combined noble gas and short-lived radiogenic isotopic systematics (e.g., ³He/⁴He and ¹⁸²W/¹⁸⁴W anomalies; Mundl et al., 2017, Mundl-Petermeier et al., 2020) and transition element ratios (e.g., Fe/Mn; Humayun et al., 2004) have been used to support the presence of a small amount of a primordial, lower mantle or core-derived component entrained in some OIB.

Linking the heterogeneity in long-lived radiogenic isotopes to mantle lithology has been achieved through combining radiogenic and stable isotopes (e.g., Day et al., 2009, Day et al., 2010), major element compositions of basalts (Jackson and Dasgupta, 2008, Shorttle and Maclennan, 2011) and trace elements in olivine (e.g., Sobolev et al., 2005, Sobolev et al., 2007, Herzberg, 2011, Neave et al., 2018). These techniques have led to the widely accepted view that recycled oceanic crust generates a significant amount of the variation seen in OIB geochemistry, along with small contributions from sediments and continental crust (e.g., Cohen and O’Nions, 1982, Hofmann and White, 1982, Weaver, 1991, Chauvel et al., 1992; see also Stracke, 2012). The recycled basaltic component is thought to be present as discrete eclogite or, more commonly, pyroxenite (olivine-poor, pyroxene-rich) lithologies, the latter possibly formed by solid-state or melt reaction of eclogite with ambient peridotite (Sobolev et al., 2007, Herzberg, 2011). However, the success of using major elements to identify pyroxenite in OIB sources may depend on the type of enriched component invoked (Lambart et al., 2013); and due to uncertainties over the melting conditions and role of crustal processes (magma recharge events, mixing, diffusional reequilibration; Matzen et al., 2017, Hole, 2018, Gleeson and Gibson, 2019) it is unclear whether source compositional differences unambiguously control the trace element concentrations in olivine phenocrysts.

Non-traditional stable isotope systems, such as Fe, provide an important, alternative method for identifying lithological heterogeneity (of both enriched and depleted components) in the OIB source. Unlike incompatible trace elements and radiogenic isotopes of incompatible elements, both of which are likely to be disproportionately affected by contributions from enriched mantle components (e.g., Burton et al., 2012), the Fe abundances of melts derived from pyroxenitic and peridotitic source mineralogies are similar (e.g., Sobolev et al., 2005). Therefore, neither pyroxenite nor peridotite lithologies should dominate the Fe isotopic composition of erupted melts, and the Fe isotopic composition of the lavas should reflect the relative contributions from these different lithologies (Williams and Bizimis, 2014). Since Fe isotopes show mineral-specific fractionation effects, they provide a unique opportunity to explore the petrological and mineralogical characteristics of mantle heterogeneity, and could help unravel common melting lithologies and processes in the mantle (Williams and Bizimis, 2014, Konter et al., 2016, Nebel et al., 2018, Nebel et al., 2019, Gleeson et al., 2020).

Equilibrium inter-mineral Fe isotope fractionations are driven by the different bonding environments of Fe in mineral structures (Macris et al., 2015, Young et al., 2015). Strong, short Fe–O bonds (due to low coordination numbers and/or the presence of oxidised Fe³⁺) concentrate heavy Fe, i.e., more ⁵⁷Fe favoured over ⁵⁴Fe (Sossi and O’Neill, 2017), which is reported as higher $\delta$⁵⁷Fe, where

$$\delta {}^{57}\mathrm{Fe}=\left(\frac{{({}^{57}\mathrm{Fe}/{}^{54}\mathrm{Fe})}_{\mathrm{sample}}}{{({}^{57}\mathrm{Fe}/{}^{54}\mathrm{Fe})}_{\mathrm{IRMM}-014}}-1\right)\times 1000$$

This equilibrium fractionation effect suggests that an olivine-dominated peridotite will be isotopically lighter than a more pyroxene-rich (pyroxenite) recycled component (Macris et al., 2015, Sossi and O’Neill, 2017) due to the presence of a small amount of Fe³⁺ in pyroxene, and hence stronger Fe–O bonds, relative to olivine. Similarly, pure iron metal (Fe⁰) is expected to be isotopically lighter than lower mantle silicate (Fe^{2+, 3+}) due to the difference in Fe valence state between the phases (Shahar et al., 2016), although this effect is predicted to be small at pressures and temperatures relevant to core equilibration (Shahar et al., 2016, Liu et al., 2017) and unlikely to be relevant to OIB. Kinetic fractionation effects may also contribute to Fe isotopic variation in OIB melt sources, with thermodiffusion (Soret diffusion) in material diffusing from the outer core into the lowermost mantle recently proposed to generate heavy $\delta$⁵⁷Fe in entrained plume material (Lesher et al., 2020).

Qualitatively consistent with an isotopically heavy recycled source component, OIB show variable $\delta$⁵⁷Fe relative to average N- and T-MORB (the latter two dominated by peridotite melting), generally extending to heavier compositions, such as over 0.25‰ in Samoa and Pitcairn. These heavy Fe isotopic compositions are suggested to relate to pyroxenitic mantle components in the OIB source. There are some N-MORB with $\delta$⁵⁷Fe as heavy as 0.2‰, which could also be consistent with indications of small amounts of enriched pyroxenite or eclogite in the MORB source (c.f., Hirschmann and Stolper, 1996). The radiogenic isotope systematics of hotspots (plumes) that display heavy $\delta$⁵⁷Fe signatures show mixing between a common peridotitic mantle component and recycled crustal endmembers (Konter et al., 2016, Nebel et al., 2019; Fig. 1), and in several cases Fe isotopes correlate positively with indices of recycling (e.g., Sr-Nd-Pb isotopes; Nebel et al., 2019). However, it is unclear whether mixing between different mantle components is represented to the same extent in Fe isotopes as in long-lived radiogenic isotopes, and whether Fe and radiogenic isotopic systems can be linked to identify the mineralogy (e.g., pyroxene enrichment) of different mantle components identified in Fig. 1. In using heavy Fe isotopic compositions to better understand mantle heterogeneity, both the sources and processes generating heavy $\delta$⁵⁷Fe melts in OIB need to be considered.

The Fe isotopic compositions recorded in OIB are heavier (and more variable) than the expected equilibrium isotopic compositions of melts from crustal endmembers contributing to mantle heterogeneity. Isotopic composition estimates for the mantle are $\delta$⁵⁷Fe = 0.03 $\pm$ 0.03‰ (DM; Craddock et al., 2013) to 0.05 $\pm$ 0.01‰ (BSE; Sossi et al., 2016); for average oceanic crust, represented by MORB, are $\delta$⁵⁷Fe = 0.15‰ (Teng et al., 2013, Sossi et al., 2016); and bulk continental crust is indistinguishable from, or lighter than, oceanic crust ($\delta$⁵⁷Fe = 0.08–0.16‰; Johnson et al., 2020). Highly differentiated (SiO₂ $>$ 70 wt%) crust, which could contribute to continentally-derived sediment, records $\delta$⁵⁷Fe $<$ 0.9‰ (Du et al., 2017) although the average $\delta$⁵⁷Fe of rocks with SiO₂ $>$ 60 wt% is around 0.3‰ (Johnson et al., 2020). Experimental and theoretical estimates of fractionation during partial mantle melting are known to be small (e.g., Dauphas et al., 2009, Sossi and O’Neill, 2017, Gleeson et al., 2020), likely $\mathrm{\Delta }$⁵⁷Fe_{melt−mantle} (= $\delta$⁵⁷Fe_melt −$\delta$⁵⁷Fe_mantle) $<$ 0.1‰ depending on Fe³⁺ buffering in the mantle (Dauphas et al., 2014). However, data from Pitcairn and Galapagos Spreading Centre lavas require a mantle component with $\delta$⁵⁷Fe = 0.30‰ (Nebel et al., 2019, Gleeson et al., 2020). The heaviest isotopic values in OIB from Samoa have previously been explained by combining source heterogeneity, partial melting, and fractional crystallisation effects (Konter et al., 2016). However, even in this multi-process scenario, equilibrium fractionation factors for these processes are required to be high (crystallisation with $\mathrm{\Delta }$⁵⁷${\mathrm{Fe}}_{\mathrm{olivine}-\mathrm{melt}}$ = $-0.45$‰, and mantle melting with $\mathrm{\Delta }$⁵⁷${\mathrm{Fe}}_{\mathrm{melt}-\mathrm{mantle}}$ $>$ 0.15‰; Konter et al., 2016), and possibly unrealistically high (Gleeson et al., 2020). It therefore remains unclear if the heaviest Fe isotopic values in OIB ($\delta$⁵⁷Fe $>$ 0.25‰) can be explained by simple melting processes of recycled crustal components embedded in ambient plume mantle.

Samoan rejuvenated lavas (a later, volumetrically less significant stage than the main shield lavas, erupted far from the plume; Natland, 1980) record the heaviest Fe isotopic compositions in the global OIB dataset (Konter et al., 2016). Samoan shield samples also show correlated ³He/⁴He-$\mu$¹⁸²W, with ³He/⁴He $⩽$ 33.8 R/R_a (Jackson et al., 2007a) and $\mu$¹⁸²W $⩾$$-17.3$ (Mundl et al., 2017), proposed to relate to core-equilibrated material (Mundl-Petermeier et al., 2020). Thus, Samoan lavas were selected for further Fe isotopic characterisation. We have expanded the shield lava dataset allowing us to study pyroxenite versus core contributions to heavy $\delta$⁵⁷Fe liquids, as the two contributions could be associated with other distinct geochemical signatures. The Azores was chosen as a second OIB locality because it shows radiogenic isotopic mixing between a common Azores mantle component and recycled components, so could allow identification of recycled mantle pyroxenite. There may also be a minor lower mantle or core component associated with the Azores plume, identified by raised ³He/⁴He relative to MORB ($⩽$18.4 R/R_a; Moreira et al., 2012) and small negative $\mu$¹⁸²W anomalies ($⩾$$-9.9$; Mundl-Petermeier et al., 2020). Unlike Samoa, the Azores is also (1) a cooler plume ($>$ 1400$°$C; Beier et al., 2012, compared to $>$ 1600$°$C in Samoa; Putirka et al., 2018) meaning melts of enriched (and possibly heavy $\delta$⁵⁷Fe) components will be minimally obscured by contemporaneous melting of ambient, relatively depleted mantle; and (2) shows a well spatially-resolved distribution of melts from different components sampled by volcanoes (Béguelin et al., 2017, Beier et al., 2018; Fig. 1). Therefore, Samoa and the Azores offer different perspectives on the links between distinct mantle components and the source of heavy $\delta$⁵⁷Fe lavas.