Experimental synthesis of Fe-bearing olivine at near-solidus temperatures and its decomposition during longtime heating

Abstract

Several investigations have demonstrated that olivine may be used to simulate geochemical and cosmochemical reactions. Since olivine in extra-terrestrial samples has varying forsterite numbers and natural olivine contains inevitable impurities, synthetic olivine with the requisite forsterite number has been prepared for various experimental research. This study aimed to synthesize Fe-bearing olivine via synthetic experiments conducted at near-solidus temperatures and elucidate the formation and decomposition mechanisms of the obtained Fe-bearing olivine. Specifically, we attempted to synthesize Fo₆₀ [Forsterite number = 100 × Mg/(Mg + Fe) = 60] olivine using a mixture of analytical-grade SiO₂, MgO, and Fe₂O₃. To clarify the stability of the obtained olivine, the temperature range (1350-1500 °C) and heating durations (1.5 or 15 h) were controlled under a constant oxygen fugacity of QFM-1 log units.

The target olivine (Fo₆₀) was obtained via heating for 1.5 h at 1500 °C, corresponding to the solidus line temperature. However, maintaining the olivine obtained at this temperature for 15 h resulted in a much higher forsterite number owing to the formation of oxidized minerals of olivine (magnetite and pyroxene). Similar oxidation products were also obtained following heating at 1450 °C for 15 h. These results indicated the primary formation of Fe-bearing olivine during 1.5 h of heating and its subsequent decomposition to magnetite and pyroxene owing to the high-temperature oxidation at near-solidus temperatures during heating for 1.5-15 h. These findings highlight a strategy for synthesizing Fe-bearing olivine and the associated mechanism and provide experimental insights into the decomposition of olivine in natural igneous rocks.

INTRODUCTION

Olivine is one of the most abundant earth-forming minerals. Several parts of the rocks in the upper mantle are considered to have been formed from forsterite, which is a magnesium-rich end member of olivine (Anderson, 1970). There are several naturally occurring olivine samples with high Mg/Fe ratios. For example, the forsterite number [Fo# = 100 × Mg/(Mg + Fe)] of natural olivine from San Carlos (Arizona, USA), the Kamuikotan Zone (Hokkaido, Japan), and Sado Island (Niigata Prefecture, Japan), vary in the ranges Fo_89.2-91.1 (Fournelle, 2011), Fo_85-87 (Niida, 1977), and Fo_79-89 (Yokoyama et al., 1992), respectively. In contrast, naturally available fayalitic olivine samples from northern Newfoundland-Labrador, Canada, show a forsterite number in the range Fo_6-13 (Wheeler, 1965).

Olivine is also abundant in meteorites and other extra-terrestrial materials. Natural olivine samples obtained from the earth have been used in experiments to simulate extra-terrestrial conditions (Sekine et al., 2015; Giese et al., 2019; Vacher et al., 2019). However, compared with earth-forming samples, meteoritical olivine samples usually exhibit highly varying Mg/Fe ratios (Menzies et al., 2005; Howard et al., 2010, 2014; Le Corre et al., 2015). For example, Mg/Fe variation in the Soko-Banja, Manych, and Sharps meteorites were Fo_60-80, Fo_25-100, and Fo_50-100, respectively (Menzies et al., 2005), and Fo₆₀ olivine was the most abundant forsterite number in these chondrites (Menzies et al., 2005). Natural olivine also often contains inevitable impurities. Given that Fe ions may be important indicators of alteration and metamorphism in planetesimals (Tomeoka et al., 1989), olivine with different Mg/Fe ratios is required for simulating extraterrestrial conditions. Additionally, spectroscopic observations of crystalline olivine dust in a protoplanetary disk are essential for investigating the thermal environment of the disk. The Mg/Fe ratio can also affect the spectral properties at near-IR and optical wavelengths (Dyar et al., 2009; Isaacson et al., 2014). Thus, preparing olivine with varying Mg/Fe ratios is essential for obtaining fundamental data for some observative investigations. In several experimental investigations, synthetic olivine samples with a specific Fo# were prepared with the desired Mg/Fe ratio for use while avoiding impurities (Takatori et al., 1993; Isobe and Yoshizawa, 2014).

Numerous previous studies have synthesized Mg-Fe olivine (Brindley and Hayami, 1965; Hanson et al., 1991; Mitchell et al., 1998; Ito et al., 2003; Ando et al., 2007; Ni et al., 2007; Tsubokawa and Ishikawa, 2017; Wang et al., 2017). Generally, solid-state reactions have been used for a long time now to synthesize forsterite (Mg₂SiO₄) crystals using MgO and SiO₂ powder (Brindley and Hayami, 1965). Compared to the method of growing single crystals, this method can synthesize minerals from commercially available chemicals, and does not demand longtime heating and seed crystals. The synthesis of forsterite-fayalite solid solutions in previous studies via solid-state reactions involving various chemical reagents is shown in Figure 1. It is evident from Figure 1 that the heating temperature is usually lower than the solidus temperature (Bowen and Schairer, 1935) required to obtain the targeted forsterite number (Ni et al., 2007; Dyar et al., 2009; Rani et al., 2014; Nash et al., 2016; Lilova et al., 2018; Pinto et al., 2019); otherwise, the formed olivine could melt based on the investigation by Nitsan (1974).

Figure 1. Experimental temperature and forsterite numbers of the products obtained in previous studies. Solid gray squares indicate the solidus temperature for each forsterite number of olivine. Solid green triangles indicate the experimental temperature in this study. Other symbols indicate the experimental temperature in previous studies.

Arai and Nagai (1963) synthesized olivine with varying forsterite numbers, starting with mixtures of analytical-grade SiO₂, MgO, and Fe₂O₃, while controlling the MgO to Fe₂O₃ ratio. This study aimed to clarify the mechanism of incorporation of Fe into the olivine solid solution during heating experiments. The heating temperatures were 1450 and 1350 °C for the syntheses of Fo_60-100 and Fo_0-50 olivine, respectively. These temperatures are near or above the solidus temperatures for Fo₆₀ and Fo_0-30 olivine, respectively. Thus, the synthesized olivine could partially melt. However, Arai and Nagai (1963) did not control the oxygen fugacity and heated for only 2 h, although similar synthetic experiments obtained Fe-bearing olivine with longer heating durations (2-24 h) (Ni et al., 2007; Dyar et al., 2009; Nash et al., 2016). Therefore, the stability of Fe-bearing olivine at these temperatures is yet to be confirmed.

Dyar et al. (2009) synthesized Fo₀ olivine, pure fayalite, at a near-solidus temperature (1200 °C) and identified hematite as one of the products. This impurity was a product of the oxidation of fayalitic olivine. Though Dyar et al. (2009) considered that the impurity could be formed owing to laser irradiation during Raman analysis, high temperature oxidation could also induce the formation of such iron oxides as decomposition products (Ejima et al., 2013).

To obtain Fe-bearing olivine via synthetic experiments involving heating at near-solidus temperatures and to elucidate the formation and decomposition mechanism of Fe-bearing olivine, the Fo₆₀ olivine was synthesized under various near-solidus temperatures (1350, 1450, and 1500 °C) and for two heating durations (1.5 and 15 h), as Fo₆₀ was abundant in some meteorites. The products obtained were then characterized via X-ray diffraction (XRD) and scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM/EDS) to investigate the mineral composition, crystallinity, and forsterite numbers of the products.

MATERIALS AND METHODS

Synthesis method

Olivine was synthesized by sintering the chemical reagents in an electronic furnace maintained at the solidus temperature. In brief, to obtain olivine with the desired composition (Fo₆₀), analytical grade SiO₂ (quartz, 99.9%, Wako Chemicals Ltd., Japan), MgO (periclase, 98.0%, Wako Chemicals Ltd., Japan), and Fe₂O₃ (hematite, 99.9%, Wako Chemicals Ltd., Japan) (1.0:1.3:0.4 in molar ratio and 0.170:0.149:0.181 g per 0.5 g) in the powdered form were mixed using an agate mortar. The excess molar ratio of MgO in the mixture was in accordance with a previously described procedure (Arai and Nagai, 1963). The chemical mixture was pressed at 20 MPa into pellets with diameter and thickness of approximately 10 and 4 mm, respectively, while the mixture for heating at 1500 °C for 15 h was pressed into pellets of the same size at 40 MPa for 3 min. Those prepared by heating at 1500 °C for 15 h were likely denser than the others; however, the actual density was not measured. In the next step, the pellets were placed in a Pt-Rh crucible and heated in the electronic furnace installed at Japan Agency for Marine-Earth Science and Technology (JAMSTEC) under 1 atm (101.325 kPa) and at an oxygen fugacity of QFM-1 log units regulated using an H₂-CO₂ gas mixture (Canil, 1997). The pellets were initially placed in the furnace at 1000 °C approximately within 5 min. Subsequently, the temperature was increased to the desired level (1350, 1450, and 1500 °C) at a heating rate of 3 °C/min. After maintaining the temperature for 1.5 or 15 h, the pellet samples were recovered from the furnace and quenched to room temperature (25 °C). Hereafter, each sample is presented with the maintained temperature and heating duration; for example, ‘pellets-1500-1.5’. In this study, olivine was synthesized by controlling the heating temperature and duration at a constant oxygen fugacity of QFM-1 log units.

Characterization techniques

X-ray diffraction. Each pellet of synthesized olivine was ground and analyzed via X-ray powder diffraction (XRD) using a SmartLab diffractometer with a Hypix-3000 detector (Rigaku, Ltd., Japan) installed at Yokohama National University at 40 kV and 45 mA, with Cu Kα radiation using Ni filter at a scan speed of 5.0°/min and step width of 0.01. Further, mineral phase identification was performed using SmartLab Studio II by searching the diffraction patterns from the ICDD card data via the Hanawalt search method.

Scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDS). The elemental composition of pellets-1500-1.5 was specifically characterized by SEM (JSM-6700F; JEOL, Tokyo Japan) equipped with EDS (EX-23000BU; JEOL) installed at Osaka Kyoiku University, as these pellets showed an ideal XRD pattern compared with the others. The results of the point analysis of the other pellets were shown as the supplementary data. Before the chemical composition analysis, each pellet’s top surface was polished using diamond powder to obtain a flat surface. The pellet samples were then fixed on a 10-mm brass stub using carbon tape and coated via Au spattering. The observations via secondary electron imaging and elemental analysis were conducted at 20 kV using EDS.

RESULTS AND DATA

XRD

While the XRD pattern of pellets-1500-1.5 showed almost crystalline olivine peaks. Strong peaks corresponding to quartz and magnesium oxide were observed in the XRD patterns of pellets-1350-1.5 and pellets-1450-1.5 (Fig. 2a), respectively. The intensities of the quartz and magnesium oxide peaks relative to the intensities of the other peaks for pellets-1450-1.5 were weaker than those in pellets-1350-1.5. Both quartz and magnesium oxides were the initial compounds in the starting chemical mixture; thus, under the conditions for obtaining pellets-1350-1.5 and pellets-1450-1.5, the reaction did not go to completion; thus, some of the initial materials were left. In addition, an unidentified weak peak at nearly 34° appeared only in pellets-1500-1.5 and pellets-1450-1.5, which could be a metastable product during the reaction since the peaks did not appear in the starting material and 15 h-heated samples.

Figure 2. (a) XRD patterns of pellets. (b) XRD patterns of pellets-1500-1.5, pellets-1500-15, pellets-1450-1.5, pellets-1450-15, pellets-1350-1.5, San Carlos olivine, and the initial reagent mixture between 31.5 and 33.0° showing the d₍₁₃₀₎ space.

To investigate the Mg/Fe ratios of the synthesized olivine samples, the (130) lattice spacings (d₁₃₀) of olivine in the pellets were compared based on the obtained XRD patterns. Their peak top position was used for the calculation. Since their broadness likely indicates the varying Fo#, the peak top positions represented the major Fo# values. The d₁₃₀ of olivine in pellets-1500-1.5, pellets-1450-1.5, and pellets-1350-1.5 were 2.792, 2.794, and 2.777 Å, respectively (Fig. 2b). The (130) reflection of pellets-1500-1.5 seemed to have a shoulder peak at slightly higher degree, which could be due to the existence of slightly higher Mg# olivine as explained in the discussion section. According to Vegard’s law, the width of d₁₃₀ changes linearly depending on the chemical composition of the solid solution (Yoder and Sahama, 1965; Arai and Nagai, 1963). The calculated Fo# values of the above-mentioned pellets were 58.7 ± 4.1, 56.0 ± 4.0, and 82.7 ± 4.0, respectively (Yoder and Sahama, 1965). Thus, Fo₆₀ olivine could be formed at 1500 and 1450 °C after 1.5 h. However, unreacted materials were left under the reaction conditions for pellets-1450-1.5. In this study, the sintering condition for pellets-1500-1.5 was the most appropriate for synthesizing the targeted olivine in bulk. The pattern of San Carlos olivine in the RRUFF™ database (Lafuente et al., 2015) is also shown in Figures 2a and 2b for reference.

In addition to olivine peaks, the XRD patterns of pellets-1450-15 and pellets-1500-15 showed peaks corresponding to magnetite and pyroxene (Fig. 2a). Pellets-1500-1.5 contained only olivine, thus magnetite and pyroxene were formed during heating for 1.5-15 h after olivine formation.

The d₁₃₀ of olivine in pellets-1500-15 and pellets-1450-15 were 2.771 and 2.774 Å, respectively (Fig. 2b). Further, their calculated Fo# values, based on Yoder and Sahama (1965), were 91.9 ± 4.0 and 88.0 ± 4.0, respectively. Double peaks were confirmed in pellets-1500-15, possibly due to the solid solution with higher Mg# or Cu Kα₂ radiation. The peak at a lower degree was stronger and was adopted for the calculation. Thus, the Fo# of olivine synthesized at 1500 or 1450 °C with the heating maintained for 1.5 h could decrease in Fe content during heating at these temperatures for 15 h.

The crystallinity of olivine can be evaluated based on the height and narrowness of the peaks in the XRD patterns of the samples, since the amorphous phases can be formed during the heating experiments (Arai and Nagai, 1963). In addition, glass could have been formed from the partial melting of olivine during the rapid cooling. Although XRD, or SEM/EDS, did not detect much glass in each pellet (Fig. 2), to compare the degrees of crystallinity of the different synthesized olivine samples, the full width at half maxima (FWHM) of the (021) and (130) reflections were obtained (Fig. 3) as shown by Nakato et al. (2008). It was observed that both peaks became higher and narrower with increasing temperature and heating duration. Thus, the crystallinity of the synthesized olivine samples showed dependence on both heating temperature and time. Notably, the crystallinity of pellets-1450-15 was higher than that of pellets-1450-1.5, while pellets-1500-1.5 and pellets-1500-15 showed similar crystallinities (similar to that of natural San Carlos olivine).

Figure 3. Comparison of the crystallinity of each pellet based on the FWHM (full width at half maximum) of (021) and (130) reflections determined from the X-ray diffraction patterns shown in Figure 2.

SEM/EDS

The surface of pellets-1500-1.5 was porous (Fig. 4a) and the grain boundaries were sintered with each other, whereas that of pellets-1500-15 h was denser (Fig. 4b), which could be due to the molding pressure on the starting material. EDS mapping (Fig. 4a) showed an almost homogeneous distribution of Mg, Al, Si, Fe, and O in the examined area of pellets-1500-1.5. The area was observed in low magnification (×350) to investigate the polished surface of the pellets in bulk; thus the peripheral part of the figure became somewhat vague. Direct evidence of loss of each element through the capsule and melt segregation was not confirmed. Acicular or tabular glass, which could be formed from the partial melting of olivine during the cooling, was also not detected. Further, some points showed Fe-rich composition and were not olivine, but possibly the Fe-rich solid solution of magnesium-iron oxide formed during the experiment (Fig. 5). The average value of the Fo# obtained via EDS point analysis was 56.4 ± 8.2, which is almost consistent with that of the XRD results. These Fo# values for pellets-1500-1.5 were slightly lower than the target value (Fo₆₀), possibly owing to the formed iron oxides via the removal of Fe from the formed olivine, as described in the discussion section, although it was deniable that the points might be the unreacted Fe-oxides due to insufficient mixing of the starting reagents. However, the XRD patterns did not show the presence of such iron-rich minerals in the pellets. Thus, these iron-rich points may be minor inclusions in the pellets.

Figure 4. A secondary electron image (SEI) and elemental mappings (O, Mg, Al, Si, and Fe) of the observed area on (a) pellets-1500-1.5 and (b) pellets-1500-15. The mapping image was obtained at 20 kV and 1.0 nA, and the dwell time was 0.2 msec, 512 × 384 pixels for 30 minutes.

Figure 5. Ternary diagram (Si+Al-Mg-Fe) of the analyzed points on pellets-1500-1.5. Solid squares show the composition of analyzed points and the orange line represents the ideal composition of olivine and pyroxene. The plotted data units were the mole ratio of each cation. Each plot was obtained randomly on the pellet at 20 kV and 1.0 nA for 15 s.

Si- and Fe-concentrated points were observed on the surfaces of pellets-1450-1.5 and pellets-1350-1.5 (Supplementary Figs. S1 and S2; Figs. S1-S4 are available online from https://doi.org/10.2465/jmps.220913). Plots on the Mg-Fe binary line were possibly due to the Mg-Fe oxide solid solution; thus, unreacted SiO₂ and Fe-rich oxide points could be the significantly Si- and Fe-concentrated points. Further, the average values of Fo# of olivine in the pellets were 68.7 ± 11.8 and 72.4 ± 15.6, respectively. These numbers were not consistent with the calculated results based on XRD (56.0 ± 4.0 and 82.7 ± 4.0, respectively).

Fe-rich and pyroxene-composition minerals were observed in pellets-1500-15 and pellets-1450-15 (Figs. S3 and S4). The distribution of Fe-rich points in the ternary diagram followed a similar trend in both pellets; however, pyroxene-composition plots were significantly confirmed on pellets-1450-15 (Fig. S4), and high-Mg# olivine on pellets-1500-15 (Fig. S3). Unreacted SiO₂ shown in pellets-1450-1.5 (Fig. S1), described in the above section, possibly contributed to the primary formation of pyroxene on pellets-1450-15. Additionally, the average Fo# values of the formed olivine samples in pellets-1500-15 and pellets-1450-15 were 85.5 ± 17.1 and 74.4 ± 14.8, respectively, which were also not consistent with those obtained via XRD (91.9 ± 4.0 and 88.0 ± 4.0, respectively). Measurements using XRD were bulk analyses; however, only the polished surfaces of the pellets were analyzed via EDS. The inconsistencies in the average Fo# values obtained by XRD and EDS possibly indicated the heterogeneity of the mineral constituents or local heterogeneity of elements inside pellets-1450-1.5, pellets-1350-1.5, pellets-1500-15, and pellets-1450-15. However, some plots could be obtained from the mineral boundaries, measuring the several mineral compositions. In addition, both broadness of the peaks in XRD (Fig. 2b) and scattered plots by SEM/EDS (Figs. 5 and S1-S4) could indicate the existence of varying Fo# olivine. The Fo# values obtained via XRD could be the bulk characteristics of the products and were considered as the respective values for each pellet, as discussed below.

INTERPRETATION AND DISCUSSION

In this study, Fo₆₀ olivine (Fo_58.7±4.1 based on XRD and Fo_56.4±8.2 based on SEM/EDS) was synthesized via heating at the temperature corresponding to the solidus line (1500 °C) for a short period (1.5 h). The crystallinity of this olivine sample was the same as that of natural olivine. Reportedly, when pellets contain a glassy constituent, a hump could exist between 20 and 35° (Raza et al., 2018; Rowe and Brewer, 2018; Ndjock et al., 2020). However, our samples did not show any noticeable humps (Fig. 2). Thus, heating at 1500 °C for 1.5 h was identified as the best condition for synthesizing Fo₆₀ olivine using the present system. However, the olivine thus obtained also contained a small amount of Fe-oxide. The appearance of impurities, such as Fe-oxide in pellets-1500-1.5; quartz and periclase in pellets-1450-1.5 and pellets-1350-1.5; and pyroxene and magnetite in pellets-1500-15 and pellets-1450-15; has been interpreted in detail in the following sections.

Formation of Fe-bearing olivine

The melting points of SiO₂, MgO, and Fe₂O₃ are 1725, 2800, and 1560 °C, respectively (Zhang et al., 2011). Thus, diffusion of each element to form olivine initially proceeded under sub-solidus conditions. While the reaction proceeded, the reduction of Fe³⁺ in Fe₂O₃ to Fe²⁺ in Fe₃O₄ during 1000-1400 °C, in FeO over 1400 °C could also occur (Ketteler et al., 2001). The Fe²⁺ ion could then be captured at the cation sites of olivine. Plots on the Mg-Fe binary line in Supplementary Figures S1 and S2 could indicate the formation of the solid-solution of (Mg²⁺, Fe²⁺)O or (Mg²⁺, Fe²⁺)Fe³⁺₂O₄ via the diffusion of the Fe²⁺ ions. If wüstite (FeO) was formed during the process, the formation of olivine could partially proceed with the wüstite melt due to its low melting point (1420 °C, Adham and Bowes, 2018) in the high-temperature experiments (1450 and 1500 °C).

Further, the Mg/Fe ratio of pellets-1350-1.5 obtained via XRD was higher (Fo_82.7±4.0) than those of the olivine samples obtained at 1450 °C (Fo_56.0±4.0) and 1500 °C (Fo_58.7±4.1). Significant Fe-rich plots were on the Mg-Fe binary line in pellets-1350-1.5 (Fig. S2); however, the plots disappeared in pellets-1450-1.5 (Fig. S1) and pellets-1500-1.5 (Fig. 5). If reducing of Fe³⁺ to Fe²⁺ at 1350 °C was sufficient, more Fe²⁺ ions should diffuse into the cation site of Fe-bearing olivine. Thus, these observations indicated that a higher temperature could be more efficient for reducing Fe₂O₃ (Chen et al., 2020), implying that heating at a temperature of 1350 °C for 1.5 h could be insufficient for reducing all Fe³⁺ to Fe²⁺ (Ketteler et al., 2001).

At 1450 °C, the crystallinity of the sample obtained after heating for 15 h was superior to that obtained after heating for 1.5 h; however, the Fo# increased from 56.0 ± 4.0 to 88.0 ± 4.0 by XRD accompanied by the formation of pyroxene and magnetite during the heating process. At 1500 °C, the crystallinity did not depend on the heating duration, but the long heating duration led to an increase in the Fo# of the sample (from 58.7 ± 4.1 to 91.9 ± 4.0) as well as the formation of pyroxene and magnetite. The effects of heating for 1.5-15 h on the formed olivine are described in detail in the section below.

Oxidation of olivine during heating

Fo_70-80 olivine can coexist with the melt at 1450 or 1500 °C under QFM-1 log unit (Nitsan, 1974). Thus, Fo₆₀ olivine formed in 1.5 h could be the metastable phase. The higher Fo# olivine formed by additional heating for 15 h could be the more stable phase compared to Fo₆₀ olivine. In this regard, there are possibly two significant processes for the decomposition of Fo₆₀ olivine to form higher Fo# olivine, pyroxene, and magnetite during the heating for 1.5-15 h. First, the high-temperature oxidation process of Fo₆₀ olivine, which can lead to the formation of decomposition products, such as pyroxene and iron oxides at approximately 820-1150 °C (Koltermann, 1962; Haggerty and Baker, 1967; Kohlstedtand and Vander Sande, 1975; Ashworth and Chambers, 2000; Gualtieri et al., 2003; Ejima et al., 2013). Specifically, magnetite formation requires the presence of Fe³⁺ ions; however, the ideal chemical formula of olivine does not contain Fe³⁺. Therefore, the oxidation of olivine in a solid state can be accomplished by the diffusion of electrons and cations in the absence of oxygen (Koltermann, 1962), as follows:   

\begin{align*} &4\text{Fe}^{2+}_{2}\text{SiO}_{4}^{\text{solid}} \\ &\text{Olivine}\\ &\quad\rightarrow 4\text{Fe$^{2+}$SiO}_{3}^{\text{solid}} + \text{Fe$^{2+}$Fe$^{3+}_{2}$O}_{4}^{\text{solid}} + 2\text{e}^{-} + \text{Fe}^{2+}. \\ &\qquad\ \ \text{Pyroxene} \qquad\quad\ \text{Magnetite} \end{align*}

Thus, the oxidation of olivine in this regard may have resulted in the formation of magnetite. However, the formed pyroxene in pellets-1450-15 showed highly Mg-rich points in terms of composition (Fig. S4), although the decomposition of Fe-bearing olivine seems to result in the formation of Fe-bearing pyroxene. The high Mg# pyroxene could have been formed after the diffusion of Fe into the liquid as described in the next paragraph.

Second, the partial melting could cause the formation of higher Mg# olivine and decomposition products such as pyroxene and magnetite. The Fe²⁺ ions in the primarily formed Fo₆₀ olivine can become concentrated in the melt, given that smaller ions, such as Mg²⁺, are known to be preferentially incorporated into the solid lattice rather than the liquid phase (Goldschmidt, 1937). Fe-rich and varying Fo# (Fo_40-70) points observed on pellets-1500-1.5 (Fig. 5) were likely due to the initial incorporation of Fe²⁺ into the partial melt and their removal from the formed olivine solid solution to the melt by diffusion, and then, further concentration of Fe²⁺ in the melt proceeded during the heating for 1.5-15 h. The removal of Fe from the formed olivine in pellets-1500-1.5 could lead to the formation of a higher Mg# olivine. The shoulder peak at (130) reflection (Fig. 2b) might be due to the initial formation of a slightly higher Mg# olivine. Further, pyroxene and magnetite could be formed from the Fe-rich melt during air cooling after recovering from the furnace; besides these, high-temperature oxidation in the solid state should be considered. And then, Fe-rich pyroxene should have been formed from the Fe-concentrated melt, but Mg-bearing pyroxene was rather confirmed in Figures S3 and S4. Thus, the formation of pyroxene could be the high-temperature oxidation product of Fo₆₀ olivine or the higher Mg# olivine as described above (Haggerty and Baker, 1967; Koltermann, 1962). Magnetite was likely formed from high-temperature oxidation of olivine and crystallization from the Fe-rich melt during cooling. An increase in the Mg/Fe ratio in olivine residues during heating for 1.5 to 15 h (Figs. 5, S1, S3, and S4) indicated the removal of Fe ions and the retention of Mg ions in the residual olivine samples.

A similar olivine oxidation process has been observed as a geological phenomenon in lava from some volcanoes (Ejima et al., 2013; Ueki et al., 2020). The olivine in lava from the Kasayama volcano was found to have an Mg-rich core (Fo_95-97) and Fe³⁺-bearing precipitated minerals on its surface (Ejima et al., 2013). In the case of Kasayama volcano, it was estimated that primary crystallized Fe-bearing olivine from magma was likely formed at approximately 1200-1400 °C as described by Matthews et al. (2021) and was secondary heated above 800 °C after solidification of the lava in the air (Ejima et al., 2013). Thus, in nature, the high-temperature oxidation of olivine can proceed at a lower temperature than primary crystallization temperature of magma and can result in the formation of magnetite and pyroxene alongside Mg-rich relict olivine (Haggerty and Baker, 1967; Koltermann, 1962; Ejima et al., 2013).

The formation of olivine decomposed products such as Mg-rich relict olivine, magnetite, and pyroxene was also confirmed in the present study. However, the oxidation temperature, time, and oxygen fugacity were quite different between nature and the present study. In particular, high-temperature oxidation can occur even at 800 °C during several hours in nature. Although temperatures over 1350 °C and considerably longer heating durations were not the focus of this investigation, maintaining temperatures of 800-1350 °C for a prolonged period of time might provide insights for further petrological interpretations.

Long-term high-temperature heating, on the other hand, should be discouraged in order to avoid superfluous oxidation of produced olivine to get the desired Fo₆₀ olivine. In the current work, applying high temperatures (e.g., 1500 °C) in a short period (e.g., ∼ 1.5 h) may be sufficient to form the desired Fo₆₀ olivine. The effect of a small number of impurities in the pellets depends on the precision requirement of future experiments. Magnetic separation could be used to eliminate unnecessary magnetite or some iron oxides in pellets-1500-1.5 (Buchmann et al., 2018) if pure olivine samples are required. If more highly pure samples are required, the other metastable compounds should be identified and removed using an appropriate procedure.

CONCLUDING REMARKS

This study synthesized Fe-bearing olivine via heating experiments using a mixture of analytical-grade SiO₂, MgO, and Fe₂O₃. To obtain Fo₆₀ olivine and elucidate the formation and decomposition process at near-solidus temperatures, these chemicals were heated at temperatures of 1350-1500 °C for 1.5 to 15 h under a constant oxygen fugacity of QFM-1 log units. The olivine with the targeted Fo# was obtained following heating at the solidus line temperature (1500 °C) for 1.5 h. During heating, olivine was synthesized via a solid-solid reaction involving each reagent. Fe²⁺ was likely supplied via the reduction of Fe₂O₃ at high temperatures. However, maintaining this high temperature for 15 h resulted in magnetite and pyroxene partially replacing the initially formed olivine. Magnetite requires Fe³⁺; thus, Fe²⁺ in olivine can be oxidized to form magnetite. The Fo# of the residual olivine after heating for 15 h was much higher than that of olivine obtained at 1.5 h, supporting the hypothesis of the high-temperature oxidation of olivine. Further, pyroxene remained as a residue of olivine decomposition. The results of this study provide a method for synthesizing Fe-bearing olivine at near-solidus temperatures and also highlight the stability of the formed olivine, while maintaining the temperature. The method in the present study demanded short-time heating duration (1.5 h) for obtaining Fe-bearing olivine, and the crystallinity was near the natural earth olivine samples. However, the formed olivine showed varying forsterite number points in SEM/EDS observation scale and contained trace Fe-rich minerals. These heterogeneity and impurities were probably due to the high-temperature oxidation of the formed olivine. These findings also provide insights into the decomposition of olivine in natural igneous rocks, since some igneous samples provide evidence of high-temperature oxidation of olivine. In addition, Fo₆₀ olivine might be exploited in simulations of meteorite parent body processes. Olivine with a certain forsterite number, such as Fo₆₀, is available in lieu of natural olivine since the forsterite number of olivine in meteorites differs from that of earth samples.

ACKNOWLEDGMENT

We thank Naoki Yoshihara for providing access to the XRD, Shungo Kawagata and Ichihito Narita for providing access to the SEM/EDS, and Sayuri Kubo for supporting the preparation of pellets. This research was supported by the Sasakawa Scientific Research Grant from The Japan Science Society Grant number 2020-6037, and JSPS KAKENHI Grant Numbers JP18K03722, JP17H06458, and JP17H06455.

SUPPLEMENTARY MATERIALS

Supplementary Figures S1-S4 are available online from https://doi.org/10.2465/jmps.220913.

REFERENCES

• Adham, K. and Bowes, G. (2018) Natural Graphite Purification Through Chlorination in Fluidized Bed Reactor: Proceedings of the First Global Conference on Extractive Metallurgy. In Extraction 2018 (Davis, B.R., Moats, M.S., Wang, S., et al. Eds.). The Minerals, Metals & Materials Series. Springer, Cham, 2505-2512. https://doi.org/10.1007/978-3-319-95022-8_211
•  Anderson,  D.L. (1970) Petrology of the mantle. Mineralogical Society of American Special Paper, 3, 85-93.
•  Ando,  M.,  Hiura,  K.,  Tsunooka,  T.,  Kagomiya,  I. and  Ohsato,  H. (2007) Synthesis of high-quality forsterite. Japanese Journal of Applied Physics, 46, 7112-7116.
•  Arai,  Y. and  Nagai,  S. (1963) Synthesis of olivine and acid solubilities of the products. Gypsum & Lime, 67, 245-252 (in Japanese with English abstract).
•  Ashworth,  J.R. and  Chambers,  A.D. (2000) Symplectic reaction in olivine and the controls of intergrowth spacing in symplectites. Journal of Petrology, 41, 285-304.
•  Bowen,  N.L. and  Schairer,  J.F. (1935) The system, MgO-FeO-SiO₂. American Journal of Science, s5-29, 151-217.
•  Brindley,  G.W. and  Hayami,  R. (1965) Kinetics and mechanism of formation of forsterite (Mg₂SiO₄) by solid state reaction of MgO and SiO₂. Philosophical Magazine, 12, 505-514.
•  Buchmann,  M.,  Schach,  E.,  Tolosana-Delgado,  R.,  Leißner,  T., et al (2018) Evaluation of magnetic separation efficiency on a cassiterite-bearing skarn ore by means of integrative SEM-based image and XRF-XRD data analysis. Minerals, 8, 390.
•  Canil,  D. (1997) Vanadium partitioning and the oxidation state of Archaean komatiite magmas. Nature, 389, 842-845.
•  Chen,  Z.,  Zeilstra,  C.,  van der Stel,  J.,  Sietsma,  J. and  Yang,  Y. (2020) Review and data evaluation for high-temperature reduction of iron oxide particles in suspension. Ironmaking & Steelmaking, 47, 741-747.
•  Dyar,  M.D.,  Sklute,  E.C.,  Menziez,  O.N.,  Bland,  P.A., et al (2009) Spectroscopic characteristics of synthetic olivine: An integrated multi-wavelength and multi-technique approach. American Mineralogist, 94, 883-898.
•  Ejima,  T.,  Akasaka,  M.,  Nagao,  T. and  Ohfuji,  H. (2013) Oxidation states of Fe and precipitates within olivine from orthopyroxene-olivine-clinopyroxene andesite lava from Kasayama Volcano, Hagi, Yamaguchi, Japan. Journal of Mineralogical and Petrological Sciences, 108, 25-36.
•  Fournelle,  J.H. (2011) An Investigation of “San Carlos Olivine”: Comparing USNM-distributed material with commercially available material. Microscopy and Microanalysis, 17, 842-843.
•  Giese,  C.,  Ten Kate,  I.L.,  Plumper,  O.,  King,  H.E., et al (2019) The evolution of polycyclic aromatic hydrocarbons under simulated inner asteroid conditions. Meteoritics & Planetary Science, 54, 1930-1950.
•  Goldschmidt,  V.M. (1937) The principles of distribution of chemical elements in minerals and rocks. Journal of Chemical Society, 655-673.
•  Gualtieri,  A.F.,  Gemmi,  M. and  Dapiaggi,  M. (2003) Phase transformations and reaction kinetics during the temperature-induced oxidation of natural olivine. American Mineralogist, 88, 1560-1574.
•  Haggerty,  S.E. and  Baker,  I. (1967) The alteration of olivine in basaltic and associated lavas. Part I: High temperature alteration. Contributions to Mineralogy and Petrology, 16, 233-257.
•  Hanson,  D.R.,  Young,  M. and  Ryerson,  F.J. (1991) Growth and characterization of synthetic iron-bearing olivine. Physics and Chemistry of Minerals, 18, 53-63.
•  Howard,  K.T.,  Benedix,  G.K.,  Bkand,  P.A. and  Cressay,  G. (2010) Modal mineralogy of CV3 chondrites by X-ray diffraction (PSD-XRD). Geochimica et Cosmochimica Acta, 74, 5084-5097.
• Howard, K.T., Alexander, C.M.O’D. and Dyl, K.A. (2014) PSD-XRD modal mineralogy of type 3.0 CO chondrites: initial asteroidal water mass fractions and implications for CM chondrites. In Lunar and Planetary Science Conference XXXXV Abstract #1830.
•  Isaacson,  P.J.,  Klima,  R.L.,  Sunshine,  J.M.,  Cheek,  L.C., et al (2014) Visible to near-infrared optical properties of pure synthetic olivine across the olivine solid solution. American Mineralogist, 99, 467-478.
•  Isobe,  H. and  Yoshizawa,  M. (2014) Formation of iron mineral fine particles by acidic hydrothermal alteration experiments of synthetic martian basalt. Journal of Mineralogical and Petrological Science, 109, 62-73.
•  Ito,  K.,  Sato,  H.,  Kanazawa,  H.,  Kawame,  N., et al (2003) First synthesis of olivine single crystal as large as 250 carats. Journal of Crystal Growth, 253, 557-561.
•  Ketteler,  G.,  Weiss,  W.,  Ranke,  W. and  Schlögl,  R. (2001) Bulk and surface phases of iron oxides in an oxygen and water atmosphere at low pressure. Physical Chemistry Chemical Physics, 3, 1114-1122.
•  Kohlstedtand,  D.L. and  Vander Sande,  J.B. (1975) An Electron microscopy study of naturally occurring oxidation produced precipitates in iron-bearing olivines. Contributions to Mineralogy and Petrology, 53, 13-24.
•  Koltermann,  M. (1962) Der thermische zerfall fayalithaltiger olivine bei hohen temperaturen. Neues Jahrbuch für Mineralogie Monatshefte, 181-191 (in German).
• Lafuente, B., Downs, R.T., Yang, H. and Stone, N. (2015) The power of databases: the RRUFF project. In Highlights in Mineralogical Crystallography (Armbruster, T. and Danisi, R.M. Eds.). W. De Gruyter, Berlin, Germany, 1-30. https://doi.org/10.1515/9783110417104-003
•  Le Corre,  L.,  Reddy,  V.,  Sanchez,  J.A.,  Dunn,  T., et al (2015) Exploring exogenic sources for the olivine on asteroid (4) Vesta. Icarus, 258, 483-499.
•  Lilova,  K.,  Deangelis,  M.T.,  Anovitz,  L.M. and  Navrotsky,  A. (2018) Surface energy of fayalite and its effect on Fe-Si-O oxygen buffers and the olivine-spinel transition. American Mineralogist, 103, 1599-1603.
•  Matthews,  S.,  Wong,  K.,  Shorttle,  O.,  Edmonds,  M. and  Maclennan,  J. (2021) Do olivine crystallization temperatures faithfully record mantle temperature variability?. Geochemistry, Geophysics, Geosystems, 22, e2020GC009157.
•  Menzies,  O.N.,  Bland,  P.A.,  Berry,  F.J. and  Cressey,  G. (2005) A mossbauer spectroscopy and X-ray diffraction study of ordinary chondrites: quantification of modal mineralogy and implications for redox conditions during metamorphism. Meteoritics and Planetary Science, 40, 1023-1042.
•  Mitchell,  M.B.D.,  Jackson,  D. and  James,  P.F. (1998) Preparation and characterisation of forsterite (Mg₂SiO₄) aerogels. Journal of Non-Crystaline Solids, 225, 125-129.
•  Nakato,  A.,  Nakamura,  T.,  Kitajima,  F. and  Noguchi,  T. (2008) Evaluation of dehydration mechanism during heating of hydrous asteroids based on mineralogical and chemical analysis of naturally and experimentally heated CM chondrites. Earth and Planetary Science Letters, 60, 855-864.
•  Nash,  W.,  Smythe,  D. and  Wood,  B.J. (2016) Synthetic olivine capsules for use in experiments. American Mineralogist, 101, 2260-2263.
•  Ndjock,  B.I.D.L.,  Baenla,  J.,  Mbah,  J.B.B.,  Elimbi,  A. and  Cyr,  M. (2020) Amorphous phase of volcanic ash and microstructure of cement product obtained from phosphoric acid activation. SN Applied Sciences, 2, 720.
•  Ni,  S.,  Chou,  L. and  Chang,  J. (2007) Preparation and characterization of forsterite (Mg₂SiO₄) bioceramics. Ceramics International, 33, 83-88.
•  Niida,  K. (1977) Olivine clinopyroxenite in serpentinized dunite - harzburgite masses of the Kamuikotan Zone, Hokkaido. Journal of the Faculty of Science, Hokkaido University, Series IV, Geology and Mineralogy, 17, 517-525.
•  Nitsan,  U. (1974) Stability field of olivine with respect to oxidation and reduction. Journal of Geophysical Research, 79, 706-711.
•  Pinto,  R.G.,  Yaremchenko,  A.A.,  Baptista,  M.F.,  Tarelho,  L.A.C. and  Frade,  J.R. (2019) Synthetic fayalite Fe₂SiO₄ by kinetically controlled reaction between hematite and silicon carbide. Journal of the American Ceramic Society, 102, 5090-5102.
•  Rani,  A.B.,  Annamalai,  A.R.,  Majhi,  M.R. and  Kumar,  A.H. (2014) Synthesis and characterization of forsterite refractory by doping with kaolin. International Journal of ChemTech Research, 6, 1390-1397.
•  Raza,  N.,  Raza,  W.,  Madeddu,  S.,  Agbe,  H., et al (2018) Synthesis and characterization of amorphous precipitated silica from alkaline dissolution of olivine. RSC Advances, 8, 32651-32658.
•  Rowe,  M.C. and  Brewer,  B.J. (2018) AMORPH: A statistical program for characterizing amorphous materials by X-ray diffraction. Computers & Geosciences, 120, 21-31.
•  Sekine,  Y.,  Shibuya,  T.,  Postberg,  F.,  Hsu,  H., et al (2015) High-temperature water-rock interactions and hydrothermal environments in the chondrite-like core of Enceladus. Nature Communications, 6, 8604.
• Takatori, K., Tomeoka, K., Tsukimura, K. and Takeda, H. (1993) Hydrothermal alteration experiments of olivine with varying Fe contents: an attempt to simulate aqueous alteration of the carbonaceous chondrites. In Lunar and Planetary Science Conference XXIV Abstract #1389.
•  Tomeoka,  K.,  Mcsween Jr,  H.Y. and  Buseck,  P.R. (1989) Mineralogical alteration of CM carbonaceous chondrites: A review. Proceedings of the NIPR Symposium on Antarctic Meteorites, 2, 221-234.
•  Tsubokawa,  Y. and  Ishikawa,  Y. (2017) Sintering polycrystalline olivine and polycrystalline clinopyroxene containing trace amount of graphite from natural crystals. Earth, Planets and Space, 69, 128.
•  Ueki,  K.,  Inui,  M.,  Matsunaga,  K.,  Okamoto,  N. and  Oshio,  K. (2020) Oxidation during magma mixing recorded by symplectites at Kusatsu-Shirane Volcano, Central Japan. Earth, Planets and Space, 72, 68.
•  Vacher,  L.G.,  Truche,  L.,  Faure,  F.,  Tissandier,  L., et al (2019) Deciphering the conditions of tochilinite and cronstedtite formation in CM chondrites from low temperature hydrothermal experiments. Meteoritics & Planetary Science, 54, 1-20.
•  Wang,  W.B.,  Shi,  Z.M.,  Wang,  X.G.,  Wang,  Z.X., et al (2017) The synthesis and properties of high-quality forsterite ceramics using desert drift sands to replace traditional raw materials. Journal of the Ceramic Society of Japan, 125, 88-94.
•  Wheeler,  E.P. (1965) Fayalitic olivine in Northern Newfoundland-Labrador. The Canadian Mineralogist, 8, 339-346.
•  Yoder,  H.S. and  Sahama,  T.H.G. (1965) Olivine X-ray determinative curve. American Mineralogist, 42, 475-491.
•  Yokoyama,  K.,  Tiba,  T. and  Chihara,  K. (1992) Picrite still in the Sado, Japan. Bulletin of the National Science Museum, Tokyo, Series C, 18, 1-11.
•  Zhang,  X.,  Wu,  G.,  Yao,  T.,  Zhang,  C. and  Yue,  Y. (2011) Characterization of individual fly ash particles in surface snow at Urumqi Glacier No. 1, Eastern Tianshan. Chinese Science Bulletin, 56, 3464-3473.