Histogram-reweighting grand canonical Monte Carlo simulations were used to obtain the phase behavior of CO2-H2O mixtures over a broad temperature and pressure range (50 °C ≤ T ≤ 350 °, 0 ≤ P ≤ 1000 bar). We performed a comprehensive test of several existing water (SPC, TIP4P, TIP4P2005, and exponential-6) and carbon dioxide (EPM2, TraPPE, and exponential-6) models using conventional Lorentz-Berthelot combining rules for the unlike-pair parameters. None of the models we studied reproduce adequately experimental data over the entire temperature and pressure range, but critical assessments were made on the range of T and P where particular model pairs perform better. Away from the critical region (T≤ 250 °), the exponential-6 model combination yields the best predictions for the CO 2-rich phase, whereas the TraPPE/TIP4P2005 model combination provides the most accurate coexistence composition and pressure for the H 2O-rich phase. Near the critical region (250 ° ≤ T ≤ 350 °), the critical points are not accurately estimated by any of the models studied, but the exponential-6 models are able to qualitatively capture the critical loci and the shape of the phase envelopes. Local improvements can be achieved at specific temperatures by introducing modification factors to the Lorentz-Berthelot combining rules, but the modified combining rule is still not able to achieve global improvements over the entire temperature and pressure range. Our work points to the challenge and importance of improving current atomistic models so as to accurately predict the phase behavior of this important binary mixture.
All Science Journal Classification (ASJC) codes
- Physical and Theoretical Chemistry
- Surfaces, Coatings and Films
- Materials Chemistry