Limitations to Quantitative Analysis of Fluid Inclusions in Geological Samples by Laser Raman Microprobe Spectroscopy

Abstract

The theoretical parameters underlying quantitative Raman spectroscopy [relative normalized differential Raman scattering (RNDRS) cross sections] are discussed. A method is described [based on Placzek's polarizability theory (see text for reference)] by which relative abundances of species of geological interest (e.g., CO2, CH4, SO2) in individual fluid inclusions are analyzed in practice. Whereas the precision of the analysis is determined by “practical limitations” (which include counting statistics and instrumental reproducibility), its accuracy depends on “practical” as well as “theoretical limitations.” The latter are due to uncertainties in the available RNDRS cross sections. These uncertainties are so large (or even unknown to date, e.g., for liquid or high-pressure gas phases) that it becomes very difficult to estimate the accuracy of the final result. By making the (not very realistic) assumption that the RNDRS cross sections available in the literature apply to the individual fluid inclusion under analysis as well as to the instrumental configuration, we determined the mean composition of a set of synthetic fluid inclusions to be 14.8 ± 2.0 mole % CH4 and 85.2 ± 2.0 mole % CO2.