Pahlavanzadeh and Fakouri Baygi employed PC SAFT EoS to pred
Pahlavanzadeh and Fakouri-Baygi  employed PC-SAFT EoS to predict the glycine receptors of carbon dioxide by aqueous MEA solutions. They used the ideal form of Smith–Missen algorithm to calculate liquid phase concentrations. The results illustrate acceptable accuracies so that the AAD was achieved equal to 36%. Moreover, Najafloo et al.  calculated CO2 solubility in MEA solutions using eSAFT-HR EoS. They obtained the overall AAD equal to 34.71% so that show better accuracy in comparison with PC-SAFT EoS and NRTL. In another similar work, Button and Gubbins  predicted CO2 absorption in MEA and DEA aqueous solutions using SAFT EoS. Their results show deviation equal to 0.01 from experimental data in mole fraction unit.
Nasrifar and Tafazzol  calculate the saturated liquid density, vapor pressure and vaporization heat for some alkanolamines including MEA, DEA, and MDEA. Moreover, H2O + DEA, MeOH + DEA and MeOH+H2O binaries were investigated and their interaction parameters were obtained. Subsequently, VLE prediction for H2O + DEA + MeOH was performed. Eventually, H2S and CO2 solubility prediction in MEA aqueous solutions were carried out utilizing PC-SAFT EoS. It should be mentioned that they considered the activity coefficients of the species equal to one.
Despite the fact that SAFT variant EoSs present good accuracies in modeling of systems containing association components, they are complex to use and to implement. To overcome the difficulty, Kontogeorgis et al.  developed SRK plus association equation of state (SRK-CPA EoS) to incorporate simplicity and accuracy. Consequently, various types of CPA EoS were developed using different cubic EoSs. The main purpose of development of a newer version is to arrive at a simpler and more accurate one. So far, some CPA variants were employed to correlate acid gas solubility in alkanolamine solutions. Some works, utilized CPA variants for acid gas solubility modeling, have been investigated below.
The electrolyte SRK-CPA EoS was utilized by Afsharpour  to model the solubility of H2S in aqueous MDEA solutions. This model contains four parts including SRK EoS, Wertheim association term, MSA theory together with the Born term. The obtained results demonstrate AAD = 17.02%.
VLE calculations of the H2O + MDEA + CO2 ternary system using mPR-CPA EoS were performed in the work of Zoghi et al. . A modification of Peng−Robinson EoS in addition to association term was considered as the molecular part of the model. Furthermore, a combination of MSA theory and the Born term contributes as electrolyte sections of the model. The results demonstrate an overall AAD equivalent to 17.3%. In a similar work , mPR-CPA EoS was employed to model H2O + MDEA + AEEA + CO2 quaternary system through the same procedure. An overall AAD = 24.3% was acquired in this study. In addition, Afsharpour et al.  use the same model to calculate H2S solubility in MDEA aqueous solutions with the overall AAD equal to 20.85%. As one can observe in these mentioned studies, the applied mPR-CPA EoS has 14 adjustable parameters for each pure component making it difficult to use the EoS.
According to the aforesaid, the use of a model with fewer optimizable parameters with good accuracy sounds to be more applicable. In simpler words, the main challenge of choosing a suitable model is to balance between simplicity and accuracy. Up to now, the SRK-CPA showed good potential as a reliable model. Therefore, in this work, we extend the electrolyte version of the classic SRK-CPA EoS to apply for VLE calculations of CO2 solubility in aqueous MDEA solutions in a broad range of operational conditions.
Results and Discussions
Conclusion In this study, an electrolyte version of SRK-CPA EoS was employed to correlate the solubility of CO2 in aqueous MDEA solutions at various solvent compositions, temperatures, pressures and acid gas loadings. Since electrolyte parts of the model including MSA and Born terms are responsible for long-range interactions, therefore, to model pure components and binary subsystems, non-electrolyte version of the model was applied.