Best practices for the analyses of CO2 fluids by Raman Spectroscopy

Penny E Wieser , Charlotte L DeVitre, Isabelle Susman

Raman spectroscopy is a key method for determining CO₂ densities in geological fluids, yet acquisition, densimeter calibration, and spectral processing methodologies vary widely between laboratories. However, the precision of this method, and the broader applicability of densimeters generated in a single lab, are still debated. This study describes a series of tests to determine how different instrument and acquisition parameters affect CO2 densities and errors on any given Raman instrument, allowing users to determine analytical best practices for their specific analytical set up. First, we discuss the effect of spectral non-linearity, demonstrating that shifts in the selected acquisition window and the choice of Ne lines used for drift correction can generate a diversity of densimeters even on a single instrument. However, our repeat measurements of natural fluid inclusion (FI) standards over ~3 years yield a 1σ variation of <0.01 g/cm³, indicating that when consistent methodologies are applied to calibration data and unknowns, Raman analyses are highly reproducible over many years. We investigate the play off between peak fitting error and parameters such as acquisition time, laser power, and sample positioning, rebutting recent suggestions that Raman analyses are associated with very large peak-fitting errors. We show that although high laser powers greatly improve signal:noise ratios (and thus peak fitting errors), natural fluid inclusions hosted in olivine show changes in density outside analytical error with increasing powers (density changing by ~0 to 0.09 g/cm3 per 10 mW of increased power). We show that the amount of heating of the trapped fluid, and thus the change in density, varies drastically based on the absorption coefficient of the host (with melt inclusions and low Fo olivines showing significant more heating than higher Fo olivines). We recommend analyses of fluids withi high optical absorption phases at low laser powers (<3-4 mW), and offer practical strategies to optimize signal without raising laser power, noting that small changes in focus and X–Y position can greatly increase signal intensity.