(620f) Modelling the Phase Behaviour of the CO2+H2O+Amine Mixtures Using Transferable Parameters with SAFT-VR – towards Solvent Design

The reduction in CO₂ emissions from anthropogenic sources has become a topic of widespread interest in recent years. As the power generation sector is by far the largest stationary-point-source of CO₂, being responsible for approximately 35% of total global CO₂ emissions¹ this issue has special relevance for the energy sector. The current method of choice for large-scale CO₂ capture is amine-based chemisorption; typically in packed columns, with the solvent of choice being a primary alkanolamine: monoethanolamine (MEA). However, MEA-based processes suffer from a number of significant disadvantages associated with the regeneration of the MEA solvent. Recently, blends of 2-amino-2-methyl-1-propanol (AMP) and ammonia (NH3) have been shown to be particularly promising novel solvent blends for CO₂ capture applications² owing to a greater capacity to absorb CO₂, lower energy of regeneration and a significantly improved resistance to degradation problems when compared to MEA solvents. However, in order to include this mixture in solvent and process design activities, it is necessary to develop accurate and reliable physical models with which to describe their thermophysical properties and fluid phase behaviour. To this end, we use the statistical associating fluid theory (SAFT)³. This is a molecular approach, specifically suited to hydrogen-bonding, chain-like fluids. In this contribution we use the SAFT approach for potentials of variable range (SAFT-VR⁴) to calculate the fluid phase behaviour of amine + H₂O + CO₂ mixtures. The molecules are modelled as homonuclear chains of attractive segments with a variable dispersion range, and a number of short-ranged off-centre attractive square-well sites are used to mediate the strong anisotropic interactions in the fluids. Following previous work on MEA⁵, we propose an asymmetric model of AMP. By asymmetric we mean that we explicitly discriminate between the primary amine functional group and the hydroxyl functional group on the AMP molecule. Six distinct association sites are required to mediate the hydrogen bonding interactions exhibited by this molecule. The SAFT-VR intermolecular parameters required to describe the site-site interactions are transferred from previous work on MEA⁵. Transferring the association parameters in this way allows us to preserve the maximum amount of molecular detail in build our model, despite the paucity of available experimental data. We model NH₃ as a single spherical segment with four association sites⁶. We select our final amine models, by consideration of its ability to represent vapour-liquid equilibrium (VLE) data as well as enthalpy of vaporisation and interfacial surface tension^7-10 where these data are available. Models for H₂O and CO₂ are taken from previous work^11-12 with one effective ?reaction? site incorporated in the CO₂ model to mediate the chemical interaction between the amines and CO₂. An added advantage of using a physically based theory for modelling reactive mixtures such as this is that reaction products are modelled implicitly in the thermodynamic model of the equilibrated fluid as associated aggregates, thus it is not necessary to obtain information detailing the physico-kinetic properties of reaction products a priori. In modelling binary mixtures of AMP+H₂O, unlike interaction parameters are transferred from previous work on a binary mixture of MEA+ H₂O⁵, and the unlike parameters for the mixtures of NH₃+H₂O are obtained as described in reference 6. There are, to the best of our knowledge, no experimental data for the NH₃+AMP mixture. Thus, in order to represent this mixture, we develop models for binary mixtures of NH₃+ethylamine and NH₃+ethanol in order to develop an understanding of the unlike association-interactions that dominate the phase behaviour of the NH₃+AMP mixture. We then transfer the unlike association-interactions from the molecular models of NH₃+ethylamine and NH₃+ethanol to the NH₃+AMP in order to describe the interactions of NH₃ the primary amine group and hydroxyl group on AMP, respectively. Unlike interaction parameters to describe the amine+ CO₂ interactions are determined by comparison with ternary experimental data. This allows us to describe the thermophysical properties and fluid phase behaviour of the NH₂+AMP mixture and that of aqueous mixtures of NH₂+AMP as CO₂ is absorbed. In conclusion, we exploit the physical basis of the SAFT approach and move towards the design of novel solvent blends, in the absence of experimental data.

1. Steeneveldt, R. et al., ChERD, 2006, 84(A9): 739 2. Choi, W. B. et al., J. Ind. Eng. Chem., 2009, 15(5) 635 3. Chapman, W.G. et al., Ind. Eng. Chem. Res., 1990. 29, 1709 4. Gil-Villegas, A. et al., J. Chem. Phys., 1997, 106 (10), 4168 5. Mac Dowell, N. et al., Ind. Eng. Chem. Res., 2010, 49(4), 1883 6. Mac Dowell, N. et al., Submitted to J. Phys. Chem. B., 2010 7. Blas, F. J. Et al., Molec. Phys. 2001, 99, 1851 8. Gloor, G. J. et al. Fluid Phase Equilib., 2002, 194, 521 9. Gloor, G. J. et al. J. Chem. Phys., 2004, 121,12740 10. Gloor, G. J. et al. J. Phys. Chem. C., 2007, 111, 15513 11. Clark, G. N. I. et al., Mol. Phys., 2006, 104(22-24):3561 12. Galindo, A. et al., J. Phys. Chem. B., 2002, 106: 4503