The experimental incorporation of Fe into talc: a study using X-ray diffraction, Fourier transform infrared spectroscopy, and Mössbauer spectroscopy

Abstract

Talc is a common Mg-rich trioctahedral layer silicate that occurs both as a primary and as a secondary mineral in a wide range of rock types. Substitution of Fe²⁺ for Mg is fairly extensive in certain rock types, particularly banded iron formations, yet there is relatively limited fundamental crystal-chemical information on this substitution. This study is an experimental investigation of Fe²⁺ substitution for Mg using X-ray diffraction, infrared spectroscopy, and Mössbauer spectroscopy. Talc was synthesized in 0.5 Fe cation [0.17 X _Fe, X _Fe = Fe/(Fe + Mg)] increments along the join Mg₃Si₄O₁₀(OH)₂–Fe₃Si₄O₁₀(OH)₂ over the range of 350–700 °C, oxygen fugacities (fO₂) from ~Ni–NiO to 3.3 log(fO₂) units below Ni–NiO, and at a pressure of 0.2 GPa. High yields of talc without any coexisting Fe-bearing phases were obtained up to 0.33 X _Fe, beyond which talc coexisted with fayalitic olivine, magnetite, or both, indicating saturation in Fe for syntheses along the talc join. Infrared spectroscopy was used to determine independently the X _Fe of talc, showing a deviation from the observed and expected composition starting at X _Fe of 0.37 ± 0.03. Minor additional solid solution occurred beyond this to a maximum X _Fe solubility of 0.50. Mössbauer spectroscopy indicated the dominance of octahedral Fe²⁺ in talc with octahedral Fe³⁺ ranging from 2.9 to 21.5 at.%, depending on the ambient fO₂. X-ray diffraction analysis did not confirm the strong dependence of the interplanar spacing d ₀₀₃ on the oxygen fugacity as reported earlier in the literature. This study provides the first experimentally constrained unit-cell volume of 474.4 ± 2.2 Å³ (142.6 ± 0.7 cm³/mol) for the end-member Fe₃Si₄O₁₀(OH)₂. The observed upper limit of iron solubility in talc of about 0.5 X _Fe agrees with the majority of analyses reported for talc, and that values above this are attributed to intergrowths of talc with the structurally distinct minnesotaite.

Appendix

Accurate determination of the fugacity of O₂ (fO₂) is critically important in this study. In the past, experimentalists have used the double capsule technique, where an inner capsule permeable to hydrogen is sealed with the sample and then encapsulated within an outer gold capsule filled with an oxygen buffer, i.e., Ni–NiO, Co–CoO, and hematite–magnetite, plus water. Because H₂ cannot readily pass through the gold membrane, it was assumed that the conditions within the gold capsule “closed” system maintained the fO₂ of the buffer that was added, a reasonable assumption at temperatures below about 600 °C (e.g., Chou 1986). However, this technique is tedious, costly, and unnecessary. For cold-seal vessels, it has been shown (e.g., Matthews et al. 2003) that the high Ni content of the autoclave and filler rod (René 41) keeps the fO₂ conditions during experiments very close to the Ni–NiO buffer. The mass of the pressure medium in cold-seal vessels is much greater than the fluid in the sample capsule (>100:1), which provides an essentially infinite buffering reservoir.

For experiments done in internally heated gas vessels, the oxygen fugacity was controlled by using a given hydrogen pressure in the ambient argon. The gas vessel was first charged with a specific hydrogen pressure, after which the hydrogen source tank was closed off. The system then pumped with argon to a desired initial total pressure after which the argon source tank was closed. At this point, the ratio of hydrogen to argon, and therefore the partial pressure ($P_{{{\text{H}}_{2} }}$) and mole fraction ($X_{{{\text{H}}_{2} }}$) of hydrogen, is fixed. Any additional compression of this gas mixture, including the pressure rise from the thermal expansion of the gas mixture during heating, only serves to define the final pressure–temperature conditions of this gas mixture. For hydrogen in equilibrium with water, the fO₂ is controlled by the reaction:

$$ {\text{H}}_{2} {\text{O}} = H_{2} + 0.5{\text{O}}_{2} $$

(1)

At equilibrium and under the condition that the activity of water is essentially unity, there is the relationship:

$$ \Delta G = 0 = \Delta G_{1,T}^{\text{o}} + RT\ln (f_{{{\text{H}}2}}^{\text{o}} ) + (0.5)RT\ln (f_{{{\text{O}}2}}^{\text{o}} ) - RT\ln (f_{{{\text{H}}2{\text{O}}}}^{\text{o}} ) + RT\ln (a_{{{\text{H}}2}} a_{{{\text{O}}2}}^{0.5} ) $$

(2)

where ΔG is the Gibbs free energy change of reaction (1), $\Delta G_{1,T}^{\text{o}}$ is the Gibbs free energy change for all phases at 1 atm and the T of interest, $f_{i}^{\text{o}}$ is the fugacity of pure i at the P and T of interest, a _i is the activity (=$f_{i} /f_{i}^{\text{o}}$) of the species in the gas mixture, R is the gas constant, and T is temperature in Kelvins. Establishing the $f_{{{\text{O}}_{2} }}$ at a given P and T then becomes a matter of specifying the $f_{{{\text{H}}_{2} }}$, and therefore the $a_{{{\text{H}}_{2} }}$, and then solving Eq. (2) for the corresponding $a_{{{\text{O}}_{2} }}$ and therefore the $f_{{{\text{O}}_{2} }}$ ($= a_{{{\text{O}}_{2} }} \cdot f_{{{\text{O}}_{2} }}^{\text{o}}$). The $f_{{{\text{H}}_{2} }}$ in these experiments is determined by multiplying the $P_{{{\text{H}}_{2} }}$ of the hydrogen–argon mixture by the corresponding hydrogen fugacity coefficient from Shaw and Wones (1964) at the P and T of interest assuming the fugacity coefficient of hydrogen in an argon–hydrogen mixture is the same as that of pure hydrogen (i.e., the Lewis and Randall gas rule). Thermodynamic data for each gas phase at 1 atm and the T of interest are from Holland and Powell (1998), while the fugacities of H₂, O₂, and H₂O at P and T are taken from Shaw and Wones (1964), Byrne and Thodos (1961) (supplemented by the data of Belonoshko and Saxena 1991), and Holland and Powell (1998), respectively. It is possible to use any desired mixture of hydrogen and argon and therefore not be restricted to a specific oxygen buffer; however, the hydrogen pressures chosen in this study were mostly kept between that of the Co–CoO buffer and the iron–magnetite/wüstite–magnetite oxygen buffers, similar to those used earlier by Forbes (1969, 1971).

This method of establishing the $f_{{{\text{O}}_{2} }}$ was tested at 500 °C and 0.25 GPa using the CoO–MnO–Co variable oxygen sensor developed by Pownceby and O’Neill (2000). In brief, the variable oxygen sensor uses the activity of CoO in CoO–MnO solid solutions in equilibrium with metallic Co as an oxygen fugacity sensor. A mixture with initial molar ratios of CoO/MnO/Co of about 1:1:2 was placed into a Ag₅₀Pd₅₀ capsule and sealed with 10 wt% H₂O. This sensor capsule was loaded into the gas vessel and the entire intensifier—gas vessel assemblage was charged with an initial H₂/(H₂ + Ar) mixture of $X_{{{\text{H}}_{2} }}$ = 0.0017 prior to sealing off the gas supply. A gas mixture with this $X_{{{\text{H}}_{2} }}$ has an expected log(fO₂) of −23.6 at the pressure and temperature of interest. After treatment for 125 h, the assemblage consisted of a (Co, Mn)O solid solution, excess CoO and Co. The composition of the (Co, Mn)O was determined to have a mole fraction of CoO = 0.69(2) based on the calibrated change in unit-cell dimensions with composition. Using the relationship between $X_{\text{CoO}}$ and oxygen fugacity of Pownceby and O’Neill (2000), the corresponding log($f_{{{\text{O}}_{2} }}$) is −24.2, or 0.6 log units lower than expected, suggesting that the calculated $f_{{{\text{O}}_{2} }}$ values reported here may be slightly (0.6-log units) higher than the actual values. For reference, the log($f_{{{\text{O}}_{2} }}$) for the Co–CoO buffer is −24.0 at these conditions (Chou 1987). Table 1 lists the hydrogen mole fractions and calculated oxygen fugacities relative to the Ni–NiO buffer for experiments done in the internally heated gas vessels.