https://doi.org/10.1021/acs.iecr.0c00940
“Let us imagine that a researcher wants to choose a suitable candidate to prepare a water-lean solvent containing 30 wt % MEA, possessing only a table with the viscosity of each diluent and its CO2 physical solubility at typical absorber temperatures. One is interested mainly in obtaining an enhancement of mass transfer rates with the new solvent, i.e., the water-lean solvent resulting from this mixture should deliver higher mass transfer rates than aqueous MEA; otherwise, there is no good reason for the shift. As a researcher, one knows that there is a trade-off between viscosity and CO2 solubility. But how strong is the nature of this trade-off?
Figure 7 addresses this issue with the results of the model developed throughout Section 2. The lines in Figure 7 designate values of η and HA for which a solvent of fixed ε, varying between ε = 20 and 60, can deliver the same mass transfer coefficient as aqueous MEA. For these calculations, the diffusivities DA had to be evaluated with the Wilke–Chang expression (eq 13), which is a weak function of the diluent molar weight M. Therefore, we have fixed M = 100 g·mol–1, as this value is close to the average of that for most interesting diluent candidates (ethylene glycol, N-methyl-2-pyrrolidone, sulfolane), and χ = 1.0.
To read Figure 7, one must pay notice to the fact that, below each line, the proposed diluent is able to deliver mass transfer coefficients above those of aqueous 30 wt % MEA. In this sense, it becomes evident that the “room for maneuver” is very limited.
In case-A-type analyses, the curves for different dielectric permittivities ε fall very closely by, this being the most optimistic scenario. Still, for diluents above the viscosity of η = 5 mPa·s, there are meager chances of finding corresponding Henry’s coefficients low enough so as to make a feasible water-lean solvent. In fact, the corresponding HA for a valid trade-off decreases exponentially with η. In other words, the viscosity of a diluent is a lot more impactful than its CO2 solubility in terms of enhancing mass transfer rates. Meanwhile, case-B-type analyses are even more pessimistic. Due to the viscosity increase with loading, and now unaided by carbamate equilibrium shifts, organic diluents with low ε are barely interesting for solvent formulation. This result reinforces how important the equilibrium shift is to elevate the mass transfer coefficients of water-lean solvents.
Perhaps a better way of making our point is considering that what is understood as CO2 solubility is the inverse of Henry’s coefficient as defined in this study. Taking case A as an example, Figure 8 shows the trade-off directly in terms of viscosity and CO2 solubility. This time, the value referring to water is plotted as a black star in the image. Additionally, the data is shown in a semilog plot.
As can be seen in Figure 8, in solvents a couple of times more viscous than water, the required CO2 solubility for a valid trade-off is a hundred-fold higher. Although this is not impossible by any means, and while there are solvents that can be found above this theoretical line (N-methyl-2-pyrrolidone and propylene carbonate being some of them), this puts into perspective the whole enterprise of looking for good physical solvents as primal matter for water-lean solvent formulation. It might be the case that looking for chemicals with lower viscosity could be more interesting than chemicals with high CO2 solubility.”