Raman quantification factor calibration for CO–CO 2 gas mixture in synthetic fluid inclusions: Application to oxygen fugacity calculation in magmatic systems

Abstract

With a combined approach using Solid-State 13C-MAS NMR and Laser Raman Microspectroscopy, we investigated the CO 2–CO gas composition ( X(CO 2)) in fluid inclusions synthesized under high pressure (200–300 MPa), high temperature (1225–1250 °C) and reducing conditions (17 < P(H 2) < 62 bars). Fluid inclusions are entrapped in a volatile-bearing basaltic glass which was characterized by FTIR for determining the water solubility (H 2O m). 13C-MAS NMR is used as a standard analysis for determining the X(CO 2). The Raman quantification factors between 13CO 2 and 13CO are determined from peak area ( F-factor), peak height ( G-factor) and according to the Placzek's polarizability theory. The calibration is derived for both CO 2 Fermi diad resonances: 2 ν 2 and ν 1. We obtain similar values for the main CO 2 resonance (2 ν 2) with 1.956 and 1.809 for F and G respectively. Results are consistent with the fact that peak height and area will measure the same quantity. For ν 1, multiple calibration trends are observed. The different trends are explained by the different 13C/ 12C ratio observed in between the samples. However, we suggest that such resonance is not suitable for determining the fluid inclusion compositions. We extended the 13C results for calibrating the F- and G-factors for 12CO 2– 12CO gas mixture in the fluid inclusions and for the main CO 2 resonance. For 12CO 2– 12CO mixture, F and G values are 1.856 and 1.756 which is in the same order as the derived values for 13C species. Thus, we propose that no significant 13C/ 12C fractionation occurs in the fluid phase and both isotopes will behave in a similar way. Using the derived calibration for 12C and 13C species, the X(CO 2) in the fluid phase was recalculated. Results are similar for both isotopes witnessing the similar behaviour of 12C and 13C fluid species during the experiments. The log f(O 2) experienced by each sample has been calculated through a thermodynamic approach using 2 independent methods. The log f(O 2) calculated from the H 2O m in the glass and the X(CO 2) in the fluid phase are in good agreement. Large discrepancy is observed for low H 2O m content which gives lower log f(O 2) value than expected from experimental conditions. Large uncertainties on the H 2O m measurements will induce a very approximate value for the log f(O 2). This method may not be accurate enough at low H 2O m and using the X(CO 2) in the fluid phase would therefore provide a better estimate of the log f(O 2).