In situ Raman spectroscopic quantification of aqueous sulfate: Experimental calibration and application to natural fluid inclusions

Sulfate is an important component of geological fluids, which is often encapsulated in fluid inclusions (FIs). Analyses of the composition of such FIs can provide important information on the origin and properties of the fluids inside. However, conventional microthermometric analyses are often hampered due to the complex composition of such fluids. Alternatively, in situ Raman spectroscopy is a powerful method to determine the aqueous sulfate concentration of FIs. Unfortunately, the effect of a common component, chloride, on the analytical results has not been well addressed yet. In addition, disagreements also exist in the current spectrum-processing method, which also prevents the application of this method in various geological cases. In this study, a series of fused silica capillary capsules (FSCCs) containing aqueous Na2SO4 (0.01-1.5 mol/kg) and Na2SO4-NaCl (mNa(2)SO(4) = 0.05 and 0.5 mol/kg, mNaCl = 0.1-4.0 mol/kg) solutions were prepared as standards. In situ Raman spectra for the stretching vibration of SO42- (v(1)-SO42-), the bending vibration (v(2)-H2O) and the stretching vibration (v(s)-H2O) of water were collected at room temperature. Results show that, for sulfate solutions, both v(2)-H2O and v(s)-H2O bands can be used as internal standards for the determination of aqueous sulfate concentration (mSO(4)(2-)). The peak area (A) ratios between v(1)-SO42- and v(2)-H2O (or v(s)-H2O) bands increased nearly linearly with mSO(4)(2-) in Na2SO4 solutions. Consequently, the relationships between mSO(4)(2-) and the spectral parameters (i.e., A(v(1)-SO42-)/A(v(2)-H2O) and A(v(1)-SO42-)/A(v(s)-H2O)) were established. However, the calculated mSO(4)(2-) will be underestimated when applying the above calibration curve to NaCl-Na2SO4 solutions, using A(v(2)-H2O) as an internal reference. In this case, the influence of chloride on the calculated mSO(4)(2-) will be largely eliminated when using A(v(s)-H2O) as an internal reference, or fitting the v(s)-H2O band with two Gaussian components at similar to 3230 cm(-1) and similar to 3450 cm(-1), and then using the total peak area as internal reference. It should be noted that our calibration results allow comparisons across different laboratories, because the spectral intensity was corrected to reduce the response function of the spectrometer, and the temperature and wavenumber dependence. Then, the A(v(1)-SO42-)/A(v(s)-H2O)-mSO(4)(2-) relationship was applied to natural FIs in fluorite from the Dalucao hydrothermal REE deposit (southwestern China), and in halite from the Paleogene evaporite in the Jianghan basin (South China). The mSO(4)(2-) in pre-REE stage fluorite-hosted FIs can reach 0.37 mol/kg, indicating that sulfate are responsible for the hydrothermal transport of REE. The mSO(4)(2-) in the halite-hosted FIs were found to be in close association with the mineral assemblage, serving as an indicator for the degree of evaporation or sedimentary facies.