Abstract
Accurate phase equilibrium data for CO2 mixtures are essential for safe and cost-effective design of carbon capture, transport and storage chains. We report measurements of the saturation water content of the carbon dioxide-rich phase for both the CO2+H2O and the CO2+H2O+NaCl system. The experiments span the temperatures 35–120 °C, pressures 1–70 MPa, and NaCl concentrations of 0, 78 and 150 g NaCl/kg water (NaCl molalities 0, 1.3 and 2.6). Total uncertainties in the reported water mole fractions are mostly between 100 and 300 ppm. The measurements are compared to previous literature data, and two equations of state: the Spycher model and EOS-CG. The EOS-CG model agrees well with our CO2+H2O data at high pressures, but generally overpredicts the water content at pressures below
- 15 MPa. The Spycher model agrees well with our data for CO2+H2O at pressures below
- 15 MPa, but exhibits some systematic deviations that can be traced to simplifying assumptions made in the model development. The Spycher model predicts the influence of NaCl on water content fairly well, but surprisingly we find that better results are obtained simply by applying Raoult’s law for the influence of ions on water activity. We derive theoretically-based bounds on the influence of salt on water content, and use them to investigate systematic errors in the measurements presented herein and in the literature. The present work indicates a clear potential to improve both models and that some previous water content
- 15 MPa. The Spycher model agrees well with our data for CO2+H2O at pressures below
- 15 MPa, but exhibits some systematic deviations that can be traced to simplifying assumptions made in the model development. The Spycher model predicts the influence of NaCl on water content fairly well, but surprisingly we find that better results are obtained simply by applying Raoult’s law for the influence of ions on water activity. We derive theoretically-based bounds on the influence of salt on water content, and use them to investigate systematic errors in the measurements presented herein and in the literature. The present work indicates a clear potential to improve both models and that some previous water content