Mass-independent fractionation of oxygen isotopes during thermal decomposition of divalent metal carbonates: Crystallographic influence, potential mechanism and cosmochemical significance

Abstract

Few physical or chemical processes defy well-established laws of mass-dependent isotopic fractionation. A surprising example, discovered two decades ago, is that thermal decomposition of calcium and magnesium carbonate minerals (conducted in vacuo, to minimise back-reaction and isotopic exchange) causes the oxygen triple-isotope compositions of the resulting solid oxide and CO₂ to fit on parallel mass-dependent fractionation lines in ln(1 + δ¹⁷O) versus ln(1 + δ¹⁸O) space, with anomalous depletion of ¹⁷O in the solid and equivalent enrichment of ¹⁷O in the CO₂. By investigating the thermal decomposition of other natural divalent metal carbonates and one synthetic example, under similar conditions, we find that the unusual isotope effect occurs in all cases and that the magnitude of the anomaly (Δ′¹⁷O) seems to depend on the room temperature crystallographic structure of the carbonate. A lower cation coordination number (as associated with smaller cation radius) correlates with a Δ′¹⁷O value closer to zero. Local symmetry considerations may therefore be influential. Relative to a reference fractionation line of slope 0.524 and passing through VSMOW, solid oxides produced by thermal decomposition of orthorhombic carbonates were characterised by Δ′¹⁷O = −0.367 ± 0.004‰ (standard error). The comparable figure from rhombohedral examples was −0.317 ± 0.010‰, whereas from the sole monoclinic (synthesised) specimen it was −0.219 ± 0.011‰. The numerical values are, to some extent, dependent on details of the experimental procedure. We discuss potential origins of the isotopic anomaly, including the possibility of hyperfine coupling between ¹⁷O nuclei and unpaired electrons of transient radicals (the ‘magnetic isotope effect’). A new mechanism based on the latter process is proposed. The associated transition state is compatible with that suggested by recent quantum chemical and kinetic studies of the thermal decompositions of calcite and magnesite. An earlier suggestion based on the magnetic isotope effect is shown to be incompatible with the generation of a ¹⁷O anomaly, regardless of the identity of the carbonate. We cannot exclude the possibility that a Fermi resonance between states leading to dissociation may additionally affect the magnitude of Δ′¹⁷O in some cases. Our findings have cosmochemical implications, with thermal processing of carbonates providing a potential mechanism for the mass-independent fractionation of oxygen isotopes in protoplanetary systems.

Introduction

With few – though important – exceptions, chemical or physical processes that modify oxygen stable isotope distributions in nature, whether under equilibrium conditions or by kinetic mechanisms, cause the ¹⁷O/¹⁶O ratio to change by approximately half the corresponding change in ¹⁸O/¹⁶O. This is related to the mass difference between ¹⁷O and ¹⁶O (1.0042 Da) being approximately half that between ¹⁸O and ¹⁶O (2.0042 Da). Isotope ratio modifications that follow this pattern of proportionality are usually referred to as ‘mass-dependent’ fractionations and are described by well-established laws (Urey, 1947; Bigeleisen and Goeppert-Mayer, 1947; Bigeleisen and Wolfsberg, 1957; Young et al., 2002; Dauphas and Schauble, 2016). In contrast, Thiemens and Heidenreich (1983) reported that ozone generated by electrical discharge in molecular oxygen deviates from this relationship, with the ¹⁷O/¹⁶O and ¹⁸O/¹⁶O ratios changing by an equal amount. Since this discovery of a chemically produced ‘mass independent’ (or ‘non-mass-dependent’) isotope effect, several other examples have been documented, usually involving gas phase photochemistry. These include the reaction of CO with the ∙OH radical in Earth's atmosphere (Huff and Thiemens, 1998; Röckmann et al., 1998); photodissociation of CO₂ (Bhattacharya et al., 2000) and of CO (Chakraborty et al., 2012). The occurrence of mass independent isotope effects in nature provides useful tracers and insights to the present and past atmosphere, climate and even the origin of life (Thiemens, 2013; Thiemens and Lin, 2019, Thiemens and Lin, 2021).

Surprisingly, thermal decomposition of calcium and magnesium carbonates is also associated with mass-independent fractionation of the oxygen isotopes (Miller et al., 2002), if the decomposition is conducted under vacuum, to minimise the potential for back-reaction and isotopic exchange between the resulting solid oxide and CO₂. Anomalous depletion of ¹⁷O in the solid oxide is accompanied by a corresponding enrichment of ¹⁷O in the CO₂. No generally accepted explanation for this finding has since been proposed. Here, we report an investigation of whether the unusual fractionation pattern occurs during thermal decomposition of other anhydrous divalent metal carbonates and consider whether the magnitude of the isotopic anomaly relates to specific characteristics of the individual carbonates. Most of the empirical data were obtained shortly after the initial report, but have not been published hitherto. Some years later, we also tested a hypothesis based on a hyperfine coupling mechanism (proposed and subsequently published by Buchachenko, 2013) for explaining the experimental findings. Although that test and its outcome were mentioned at a conference (Miller et al., 2012), no data were reported in the abstract; neither have they been published since. We therefore present the details here, together with the (also hitherto unpublished) oxygen triple-isotope measurements of oxides from thermal decomposition of eight different divalent carbonate minerals, for comparison with the Ca and Mg examples reported by Miller et al. (2002).

On the basis of the Buchachenko (2013) mechanism, details of which are discussed below (Section 5.6), all divalent metal carbonates would be expected to exhibit mass-independent fractionation of the oxygen isotopes during the thermal decomposition process, unless the corresponding cation in the univalent state contains no unpaired electrons. In this case, the intermediate cannot be a radical pair and therefore its reactions are not spin-state selective. Because Cu⁺ has electronic configuration [Ar]3d¹⁰, a definitive empirical test of the hypothesis was to perform controlled thermal decomposition of CuCO₃ in vacuo and measure the oxygen triple-isotope compositions of the resulting CuO and CO₂. If the proposed mechanism is correct, no ¹⁷O anomaly would be produced.

Unfortunately, CuCO₃ does not occur in nature. The first reliable synthesis – from heating the hydroxycarbonate minerals azurite and malachite at 500 ± 10 °C for approximately 21 h duration in CO₂ atmosphere at 20 ± 1 kb pressure – was reported by Ehrhardt et al. (1973), with the structure being published the following year (Seidel et al., 1974). CuCO₃ is a grey, monoclinic crystalline solid (space group Pa-C²_S, cation coordination number 5), reputedly stable in dry air at room temperature for several months. To conduct the experimental test, it was therefore necessary to first prepare a small quantity (~20 mg) of CuCO₃, based on the Ehrhardt et al. (1973) procedure.