Incompatibility of argon during magma ocean crystallization

https://doi.org/10.1016/j.epsl.2020.116598Get rights and content

Highlights

  • Argon is incompatible during deep magma ocean crystallization.

  • Trapped liquids control retention of argon in crystallizing magma oceans.

  • Basal magma oceans should form relatively rich in noble gases.

Abstract

We report results from multi-anvil (MA) and laser-heated diamond anvil cell (LH-DAC) experiments that synthesize high-pressure phases, including bridgmanite, ferropericlase, stishovite, and ultramafic liquid, in the presence of an argon-rich fluid. The goal of the experiments is to constrain the equilibrium distribution of argon in magma ocean environments. Argon concentrations in LH-DAC experiments were quantified by electron microprobe analysis, while argon concentrations in MA experiments were quantified by laser-ablation mass spectrometry and electron microprobe analysis. Our LH-DAC experiments demonstrate that argon solubility in ultramafic liquid is near or above 1.5 wt.% at conditions between 13–101 GPa and 2300–6300 K. Argon concentrations in bridgmanite and ferropericlase synthesized in LH-DAC experiments range from below detection to 0.58 wt.%. Argon concentrations in bridgmanite and ferropericlase synthesized in MA experiments range from below detection to 2.16 wt.% for electron microprobe measurements and laser-ablation measurements. We interpret this wide range of argon concentrations in minerals to reflect the variable presence of argon-rich fluid inclusions in analytical volumes. Our analyses therefore provide upper limit constraints for argon solubility in high-pressure minerals (<0.015 wt.%) across all mantle pressures and temperatures. The combination of relatively high argon solubility in ultramafic liquid (∼1.5 wt.%) and low argon solubility in minerals implies argon incompatibility (DbridgmanitemeltAr < 0.01, DferropericlasemeltAr < 0.01) during magma ocean crystallization and that the initial distribution of argon, and likely other neutral species, may be controlled by liquids trapped in a crystallizing magma ocean. We thus predict a basal magma ocean would be enriched in noble gases relative to other regions of the mantle. Moreover, we predict that the noble gas parent-daughter ratio of magma ocean cumulates pile will increase with crystallization, assuming refractory and incompatible behavior for parent elements.

Introduction

Ocean island basalts contain materials that were isolated from the convective mantle extremely early in solar system history and have remained geochemically distinct to the modern day (Mukhopadhyay, 2012; Peto et al., 2013; Caracausi et al., 2016; Rizo et al., 2016; Mundl et al., 2017; Williams and Mukhopadhyay, 2019). Evidence for early-isolated or primordial mantle materials largely comes from studies of W isotopes and the noble gases, including helium (He), neon (Ne), argon (Ar), krypton (Kr) and xenon (Xe). Noble gases are uniquely suited to trace Earth's earliest history because they are a group of highly atmophile, largely inert elements (Brooker et al., 2003; Heber et al., 2007; Jackson et al., 2013; Graham et al., 2016). Furthermore, the group is rich in radiogenic isotopes, produced from both short- and long-lived decay systems. Indeed, it was the observation of a high 3He/4He component within intraplate magmas that first suggested the mantle contains a primordial, less degassed geochemical component (Jenkins, 1978; Kaneoka and Takaoka, 1980; Kurz et al., 1982). This view of the mantle has been reinforced by continued analyses of Ne and Ar isotopes produced by radio-decay (Honda et al., 1991; Farley and Craig, 1994; Burnard et al., 1997; Trieloff et al., 2000; Moreira et al., 2001; Kurz et al., 2009).

More recently, Xe isotopes now constrain the formation of high 3He/4He materials to within the lifetime of 129I, or within the first 80 Ma of solar system history (Mukhopadhyay, 2012; Peto et al., 2013; Caracausi et al., 2016). A separate timing constraint is gained from the discovery of Ne derived from the solar nebula in high 3He/4He materials (Williams and Mukhopadhyay, 2019), as solar nebular gases were dissipated from the solar system within the first 10 Ma of solar system history (e.g., Mamajek, 2009). These timing constraints combine to indicate that high 3He/4He materials were first isolated during the era of accretion. This is the time when magma ocean formation and crystallization was a dominant mechanism for mass transfer within terrestrial planets.

In this contribution we focus on the partitioning behavior of Ar during magma ocean solidification in order to constrain the initial distribution of noble gases within the mantle-atmosphere system. Strong partitioning of Ar into mineral phases would lead to the sequestration of large amounts of Ar into Earth's primordial mantle, leaving an Ar-poor atmosphere in response. Partitioning behavior can be quantified as either the equilibrium concentration or solubility ratio of Ar in two phases (DminmeltAr=[Ar]mineral[Ar]melt=Sol.mineralArSol.meltAr).

The large majority of experimental and observational data indicate that Ar and other noble gases are extremely incompatible (Dminmeltnoblegas < 10−3) within the uppermost mantle (<3 GPa) (Chamorro et al., 2002; Brooker et al., 2003; Heber et al., 2007; Delon et al., 2019; Jackson et al., 2013; Graham et al., 2016). Although Watson et al. (2007) report high Ar solubility in olivine and pyroxene at low Ar fugacity (ƒAr). Rapid cooling of large igneous bodies can lead to 50% trapped magma in the cumulate pile (Tegner et al., 2009). High degrees of noble gas incompatibility within lower pressure mantle environments, combined with the likely large amount of trapped liquids within magma ocean cumulate piles, implies that the initial noble gas budget of the uppermost mantle was determined by the amount of trapped melt within the cumulate pile. The major source of uncertainty regarding the partitioning of noble gases during magma ocean crystallization relates to their behavior under higher pressures within the mantle.

No direct partitioning studies of noble gases have been completed at the pressures applicable to the deepest magma oceans, but partitioning behavior can be inferred from the solubility ratio of noble gases in the applicable phases at high pressure, namely ultramafic liquid, ferropericlase (FP), and bridgmanite (Brg). Several studies have reported Ar contents of quenched ultramafic liquids (QUL) that were reacted within Ar pressure media using laser-heated diamond anvils cells (LH-DAC) (Chamorro-Perez et al., 1998; Bouhifd and Jephcoat, 2006) (we use the term QUL because molten silicate often quenches to nanometer scale crystals with small amounts of Fe alloy in LH-DAC experiments, Supplementary Fig. 1 and 2). These studies uniformly conclude that Ar solubility in ultramafic liquids plateaus at a concentration of 0.2 wt.% near 5 GPa but sharply decreases to low values (<0.03 wt.%) under higher pressures (∼10 GPa). Higher pressure crystallization of magma oceans is dominated by Brg, and more recent experiments suggest that Ar solubility in Brg may be relatively high, ranging up to 1 wt.% Ar (Shcheka and Keppler, 2012). The reported high Ar solubility in Brg is suggested to reflect abundant, large radius oxygen vacancies that can host also large radius, neutrally charged noble gases.

The combination of high Ar solubility in Brg and low solubility in ultramafic liquid implies Ar, and potentially other neutral species, would be compatible during crystallization of deep magma oceans. A ratio of reported Ar solubility for Brg and ultramafic liquid implies DBrgmeltAr30. If true, crystallization of Brg from a magma ocean would create a relatively Ar-rich lower mantle. Similarly, fractionation of Brg from a basal magma ocean would create a lowermost mantle that is highly depleted in Ar and potentially other neutral species.

The conclusion of low Ar solubility in ultramafic liquid under high pressures, however, has been challenged. First, the physical mechanism behind the sharp decrease in solubility has remained elusive (Guillot and Sator, 2012). It is expected that noble gases dissolve into larger atomic-scale voids in silicate melts and minerals (Carroll and Stolper, 1993; Jackson et al., 2015). Compression of silicate liquid should decrease the size of atomic-scale voids and therefore may limit noble gas solubility. However, densification of melt is a complex, continuous process (Lee, 2011; Solomatova and Caracas, 2019) and this continuity makes the sharpness of the solubility drop difficult to reconcile with only the effects of compression. Moreover, collapse of He solubility in ultramafic liquids has been reported at the same pressure reported for Ar (Bouhifd et al., 2013). Drops in He solubility related to the compression of silicate liquid should presumably occur at higher pressure relative to Ar given their relative atomic radii. Furthermore, separate experiments aimed at measuring Ar solubility in SiO2 liquid at high-pressure (up to 19 GPa) failed to reproduce the discontinuous solubility drop. Rather, Ar concentration of quenched SiO2 liquid remains high beyond the pressures where the solubility drop had previously been reported for this liquid composition (Niwa et al., 2013). The conclusion of high Ar solubility in Brg also remains unverified, leaving the partitioning of Ar, and other neutral species, uncertain in deep magma ocean environments.

We report two parallel series of experiments to constrain the Ar partitioning during deep magma ocean crystallization. The first series comprises Brg, FP, and stishovite (Stv) synthesized in the presence of an Ar-rich fluid using a multi-anvil (MA) press (Fig. 1 and Supplementary Fig. 1). Composition is systemically varied in the MA series to quantify the effect of oxygen vacancies promoting Ar solubility. The second series comprises QUL, Brg, and FP synthesized in the presence of an Ar-rich fluid using a LH-DAC (Fig. 2, Supplementary Figs. 2 and 3). Together, our experiments demonstrate that, up to 100 GPa and 5000 K, Ar solubility in ultramafic liquid remains relatively high (∼1.5 wt.%), whereas Ar solubility in high pressure mineral phases remains relatively low (<0.015 wt.%) despite variations in oxygen vacancy concentrations. Both of these results contrast with previous reports of Ar solubility in high-pressure materials (Chamorro-Perez et al., 1998; Bouhifd and Jephcoat, 2006; Shcheka and Keppler, 2012), and together our new results imply that Ar is highly incompatible (DminmeltAr < 0.01) during deep magma ocean crystallization. We suggest that the initial distribution of Ar, and potentially other highly volatile elements, within the silicate Earth was controlled by liquids trapped during magma ocean crystallization.

Section snippets

Experimental approach

We conducted a series of experiments using a MA press (Table 1) with the goal of reacting Brg of varying chemistry and point defect populations with Ar-rich fluid. The large volume of MA experiments also enabled analysis by laser-ablation to complement electron microprobe analyses in this subset of our experiments, as detailed below. We also conducted experiments with LH-DACs to document the reactivity of Ar with silicate liquid, Brg, and FP up to P-T conditions that approach the core-mantle

Multi-anvil series: run product description

MA experiments were completed at 23–24 GPa and 2173–2373 K. Run products were dominated by Brg but contained lesser amounts of FP and Stv. Run products were spatially organized into regions of different mineral assemblages (Fig. 1, Supplementary Fig. 1), likely related to the initial distribution of chemical components (e.g., Ar-bearing SiO2 glass or MgO) in the starting composition. Regions include pure mineral types but also Brg with either intergrown Stv or FP (Supplementary Fig. 1). Stv and

Oxygen vacancies in bridgmanite

Atomic-scale porosity facilitates noble gas dissolution into minerals and melts (Carroll and Stolper, 1993; Jackson et al., 2015). Porosity can take the form of an interstitial space or vacancy, and because of the large size of noble gases, larger radii interstices or vacancies, such as oxygen vacancies or ring sites, are expected to promote noble gas solubility (Shcheka and Keppler, 2012; Jackson et al., 2015). Brg can contain a relatively high concentration of oxygen vacancies, depending on

Conclusions

Two series of high P-T experiments indicate that Ar is incompatible during crystallization of deep magma oceans (DbridgmanitemeltAr < 0.01, DferropericlasemeltAr < 0.01). Our inference of incompatibly is based on upper-limit constraints we derived for Ar solubility in Brg and FP. These results, in combination with literature results (Brooker et al., 2003; Heber et al., 2007; Delon et al., 2019; Jackson et al., 2013), imply that Ar is highly incompatible in all stages of magma ocean

CRediT authorship contribution statement

All authors contributed to the drafting of the manuscript. Jackson performed the experiments, prepared the experiments, completed the microprobe analysis, and completed the first draft of the manuscript. Williams completed the laser-ablation analyses. Fei facilitated the completion of the multi-anvil and LH-DAC experiments. Fei also contributed to the analysis of the XRD data. Mukhopadhyay facilitated the laser-ablation analyses. Du and Bennett contributed to the execution of the experiments.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank Emma Bullock for her assistance with focused ion milling and electron microprobe work at Carnegie, Ali Bouhifd for sharing data for QUL, and Sash Hier-Majumder for discussions on trapped liquid modeling. We additionally thank Steve Parman and an anonymous reviewer for their careful, constructive reviews. CRMJ, ZD, and NRB acknowledge fellowship support from the Carnegie Institution and CRMJ additionally acknowledges startup support from Tulane University. CDW acknowledges support from

References (52)

  • M. Moreira et al.

    Solar neon in the Icelandic mantle: new evidence for an undegassed lower mantle

    Earth Planet. Sci. Lett.

    (2001)
  • K. Niwa et al.

    Argon solubility in SiO2 melt under high pressures: a new experimental result using laser-heated diamond anvil cell

    Earth Planet. Sci. Lett.

    (2013)
  • M.K. Peto et al.

    Heterogeneities from the first 100 million years recorded in deep mantle noble gases from the Northern Lau Back-arc Basin

    Earth Planet. Sci. Lett.

    (2013)
  • A. Rosa et al.

    Krypton storage capacity of the Earth's lower mantle

    Earth Planet. Sci. Lett.

    (2020)
  • Y. Thibault et al.

    The influence of pressure and temperature on the metal-silicate partition coefficients of nickel and cobalt in a model C1 chondrite and implications for metal segregation in a deep magma ocean

    Geochim. Cosmochim. Acta

    (1995)
  • R. Boehler et al.

    High-pressure melting curves of argon, krypton, and xenon: deviation from corresponding states theory

    Phys. Rev. Lett.

    (2001)
  • M. Bouhifd et al.

    Helium in Earth's early core

    Nat. Geosci.

    (2013)
  • M.A. Bouhifd et al.

    Aluminium control of argon solubility in silicate melts under pressure

    Nature

    (2006)
  • J.P. Brodholt

    Pressure-induced changes in the compression mechanism of aluminous perovskite in the Earth's mantle

    Nature

    (2000)
  • R.A. Brooker et al.

    The ‘zero charge’ partitioning behaviour of noble gases during mantle melting

    Nature

    (2003)
  • P. Burnard et al.

    Vesicle-specific noble gas analyses of “popping rock”: implications for primordial noble gases in Earth

    Science

    (1997)
  • A. Caracausi et al.

    Chondritic xenon in the Earth's mantle

    Nature

    (2016)
  • E.M. Chamorro et al.

    Ar and K partitioning between clinopyroxene and silicate melt to 8 GPa

    Geochim. Cosmochim. Acta

    (2002)
  • E. Chamorro-Perez et al.

    Low argon solubility in silicate melts at high pressure

    Nature

    (1998)
  • L.T. Elkins-Tanton

    Magma oceans in the inner solar system

    Annu. Rev. Earth Planet. Sci.

    (2012)
  • K. Farley et al.

    Atmospheric argon contamination of ocean island basalt olivine phenocrysts

    Geochim. Cosmochim. Acta

    (1994)
  • Cited by (0)

    View full text