The oxidation state of europium in silicate melts as a function of oxygen fugacity, composition and temperature

Highlights

• Eu^3 +/Eu^2 + in melts can be predicted using an empirical equation.

• Eu^3 +/Eu^2 + in a natural melt is not preserved on quenching to a glass.

• Eu^2 + + Fe^3 + = Eu^3 + + Fe^2 +.

Abstract

Europium L_III-edge X-ray absorption near edge structure (XANES) spectra were recorded for a series of synthetic glasses and melts equilibrated over a range of oxygen fugacities ( f O₂s, from − 14 to + 6 logarithmic units relative to the quartz–fayalite–magnetite, QFM, buffer) and temperatures (1250–1500 °C). Eu^3 +/ΣEu (where ΣEu = Eu^2 + + Eu^3 +) values were determined from the spectra with a precision of ± 0.015. Eu^3 +/ΣEu varies systematically with f O₂ from 0 to 1 over the range studied, increases with decreasing melt polymerisation and temperature, and can be described by the empirical equation: $\mathrm{E}{\mathrm{u}}^{3+}/\mathrm{\Sigma }\mathrm{E}\mathrm{u}=\frac{1}{1+{10}^{-0.25logf{\text{O}}_2-6410/T-14.2\mathrm{\Lambda }+10.1}}$, where Λ is the optical basicity of the melt and T is the temperature in K. Eu^3 +/ΣEu in glasses and melts equilibrated at the same conditions are in excellent agreement for Fe-free systems. For Fe-bearing compositions the reaction Eu^2 + + Fe^3 + = Eu^3 + + Fe^2 + occurs during quenching to a glass and the high temperature value of Eu^3 +/ΣEu is not preserved on cooling; in situ measurements are essential for determining Eu^3 +/ΣEu in natural melts.

Introduction

Europium is the only rare earth element (REE) that can occur in significant proportions as an oxidation state other than REE^3 + in natural silicate melts (e.g., Drake, 1975). The oxidation state of Eu is of geochemical interest because Eu^2 + and Eu^3 + partition differently (e.g., Drake, 1975) and thus the behaviour of Eu in a magma depends on the oxygen fugacity (f O₂) and the phase(s) crystallising. Chondrite-normalised REE patterns with anomalous Eu concentrations, quantified as Eu/Eu*$=\mathrm{E}\mathrm{u}/\sqrt{\left[\mathrm{S}\mathrm{m}\times \mathrm{G}\mathrm{d}\right]}$, are frequently observed and are attributed to the redox-variability of Eu (e.g., Morris et al., 1974a, Drake, 1975, Hoskin et al., 2000). Experimental studies have calibrated the fO₂-dependency of Eu partitioning between feldspars and melt and shown D_Eu^2 + (where D_i is the ratio of the weight fraction of i in the crystal to that in the melt) to be an order of magnitude greater than D_Eu^3 + (Aigner-Torres et al., 2007, Drake, 1975, McKay et al., 1994, Wilke and Behrens, 1999). Conversely, Eu^2 + is far less compatible than REE^3 + in pyroxenes (McKay et al., 1994, Wadhwa, 2001; see Eby, 1975 for exceptions). The resulting positive or negative Eu anomalies can provide an indication of crystal fractionation (e.g., Brophy and Basu, 1990, Hoskin et al., 2000). Alternately, these anomalies could be used to estimate the fO₂ at which a sample crystallised. For example, the partitioning of Eu between pyroxene and melt as a function of f O₂ has potential as an oxybarometer, particularly for meteoritic samples (Karner et al., 2010, McKay et al., 1994, Shearer et al., 2006). Anomalous concentrations of Eu are also commonly exhibited by zircon and similarly could record the f O₂ of crystallisation (Burnham and Berry, 2012, Hoskin and Schaltegger, 2003, Trail et al., 2012).

Partitioning between feldspar, or other minerals, and melt allows the oxidation state ratio of Eu in the melt to be estimated, because the observed D_Eu can be modelled as an average of D_Sr (a proxy for D_Eu^2 +) and D_Gd or D_Sm (a proxy for D_Eu^3 +) weighted by the relative proportions of Eu^2 + and Eu^3 + in the melt (e.g., McKay et al., 1994). However, such experiments ideally require melts with a compositionally invariant phase on the liquidus and thus cannot fully address the roles of composition and temperature, as neither can be varied freely. Spectroscopic determinations of Eu^3 +/Eu^2 +, or Eu^3 +/ΣEu where ΣEu = Eu^2 + + Eu^3 +, are not subject to this constraint.

¹⁵¹Eu (47.8% natural abundance) Mössbauer spectroscopy has been used to determine Eu^3 +/ΣEu in silicate glasses (Nemov et al., 2007, Virgo et al., 1981), though the limited applications and the long spectral acquisition times (several days) have probably hampered greater uptake of this technique. Eu^3 +/ΣEu can also be determined by electron paramagnetic resonance (EPR) spectroscopy (Eu^2 + is EPR active), however, the method is restricted to Fe-free compositions since the resonances of other paramagnetic ions will overlap with the Eu signal (Morris et al., 1974a, Morris et al., 1974b, Schreiber, 1977).

Eu L_III-edge X-ray absorption near edge structure (XANES) spectroscopy has been used to provide information on the oxidation state ratio and coordination environment of Eu in aqueous solutions, silicate melts and minerals (Antonio et al., 1997, Cicconi et al., 2012, Karner et al., 2010, Mayanovic et al., 2007, Rakovan et al., 2001, Shearer et al., 2011). Eu^2 + spectra are characterised by an intense peak (white line) at ~ 6974 eV and Eu^3 + spectra by a white line at ~ 6982 eV (Rakovan et al., 2001, Shimizugawa et al., 1999, Takahashi et al., 2005). The peaks are assigned to the 2p_3/2 → 5d transitions of Eu^2 + and Eu^3 + (Rakovan et al., 2001, Takahashi et al., 2005). The ~ 8 eV difference in energy between these features is large compared to the core hole width of 3.91 eV (Krause and Oliver, 1979), and hence for mixed valence systems the two peaks are well resolved and the relative contributions from each oxidation state can be determined. This resolution makes quantification of redox ratios potentially easier than for many other elements (e.g., Berry and O'Neill, 2004, Sutton et al., 2005). For example, Eu^3 +/ΣEu has been determined with a precision of 0.01 for a sample of apatite containing only 700 ppm Eu (Rakovan et al., 2001). XANES spectroscopy can also be used to determine oxidation state ratios in situ; oxidation states present in a melt may not be preserved in a glass, even with rapid quenching, due to electron exchange reactions (Berry et al., 2006). In situ experiments on melts have suggested that electron exchange may occur between Eu and Fe (Cicconi et al., 2015).

In order to use Eu anomalies in minerals as an oxybarometer it is necessary to understand the separate effects of f O₂, melt composition, and temperature on Eu^3 +/ΣEu in magmas. The work of Morris et al., 1974a, Morris et al., 1974b represents the most comprehensive study to date of the controls on the oxidation state ratio of Eu in glasses, covering temperatures from 1415 to 1650 °C, 12 log units of f O₂ and over 20 compositions in the system CMAS; they found that Eu^3 + was stabilised by high f O₂s, melt compositions with high Ca/Mg ratios, and low temperatures. More recently, XANES spectroscopy has been used to determine Eu^3 +/ΣEu in a number of glass compositions equilibrated at four values of f O₂, although the data were too limited to model or interpret accurately (Cicconi et al, 2012).

Here we use XANES spectroscopy to determine Eu^3 +/ΣEu as a function of fO₂ for a range of silicate glasses and melts, thereby extending the work of Morris et al., 1974a, Morris et al., 1974b to other melt compositions and to lower temperatures. We also use an X-ray absorption spectroscopy furnace (Berry et al., 2003) to make the first in situ determinations of Eu^3 +/ΣEu in silicate melts as a function of f O₂. Pressure may also affect Eu^3 +/ΣEu although this variable will not be addressed here.