https://doi.org/10.1016/j.seppur.2020.118193
“This Section 6.3 will concern itself on literature data for mass transfer parameters that are somehow applicable to industrial operations. This means either liquid phase mass transfer coefficients, overall mass transfer coefficients or CO2 molar fluxes in water-lean solvents. These parameters are often measured in equipment such as string of discs, wetted wall columns, stirred cell reactors or even bench scale absorber columns. Another sort of rate information, obtained in more loosely controlled environments and typically just comparative in nature, will be dealt with in a different part of this review (Section 7).
Mass transfer rates of water-lean solvents have been measured by multiple authors. Of all chemical combinations, by far the most investigated one was that of MEA + methanol + water. Studies carried by Sema et al. [9], Fu et al. [10], Gao et al. [11], [12], [13] and Rashidi et al. [14] have all exhaustively observed the increase in absorption rates when shifting from aqueous MEA to mixtures containing methanol. Tamajón et al. [167] also obtain enhanced absorption rates in MDEA + methanol solutions. Conversely, Chen et al. [46] saw the absorption rates dropping in MDEA + ethanol solutions, probably because the kinetics of this tertiary amine in ethanol are so depressed that chemical absorption is almost precluded.
As implied before, methanol is a special organic diluent since its viscosity is fairly lower than that of water. Therefore, even if its dielectric permittivity is also lower, ε ≈ 29 according to Wohlfarth [165], viscosity shall not be a big issue when operating with methanol mixtures. When adding this up to the high CO2 solubility in pure methanol, mass transfer rates are understandably higher in hybrid formulations with this alcohol.
Of course, methanol has the issue of its high volatility, and both solvent losses and high latent heat expenses are consequences of its employment. Gao et al. [168] have attempted to address this problem by adding glycerol into a methanol–MEA solvent. The problem was addressed, but absorption rates were evidently reduced (though remaining higher than those of aqueous MEA).
For the cases involving chemicals with low volatility, several researchers have found similar behavior among distinct solvents. Initially, absorption rates are clearly higher in water-lean solvents containing diluents with good physical CO2 absorption capacities. The investigation of Garcia et al. [144] on unloaded water-lean solvents containing blends of MEA or DEEA/MAPA with glycols and carbitol is quite illustrative of this phenomenon. However, as solvents are loaded and their viscosities increase, this advantage starts disappearing. This has been observed by Yuan and Rochelle [18], [169] and by Wanderley et al. [66] in clear terms, and can be also interpreted from the data of Zou et al. [170] regarding CO2 absorption in a blend of MEA, sulfolane and water. An example of this behavior can be seen on Fig. 13, which was plotted based on data from Dugas and Rochelle [61] and from Yuan and Rochelle [169].”
“Fig. 13. Decrease in liquid phase mass transfer coefficient with loading observed for mixtures of 5 molal piperazine with water, water + sulfolane and water + imidazole at 40 °C. The organic diluent/water ratio is given in mass basis. Data obtained from Dugas and Rochelle [61] and Yuan and Rochelle [169].”
”
In terms of performance, N-methyl-2-pyrrolidone is perhaps the best diluent found so far for enhancing absorption rates [18], [66], [169], [170]. Its low Henry’s coefficient and viscosity, which can be seen on display in Fig. 10, plus its dielectric permittivity at 25 °C of ε ≈ 32.6 [165] (lower than that of water, but higher than that of any alcohol) would make this a great candidate for water-lean solvent formulation, were it not reproductively toxic [32]. As the industry is slowly backing away from using N-methyl-2-pyrrolidone, it is hard to make a case for solvents relying on this chemical. Sulfolane has also been seen to deliver enhancements in mass transfer rates [66], [169], [170], which is somewhat unexpected due to the high viscosity of this diluent (see Fig. 10, Fig. 11). However, since sulfolane has also been identified at eventually promoting phase separation upon loading in all most water-lean solvents, one should wonder what exactly is being measured when mass transfer rates are identified in sulfolane–MEA or sulfolane–piperazine mixtures.
Liu et al. [171] have tried developing water-lean solvents targeted at having low viscosity when absorbing CO2. Their case is quite special, as the proposed water-lean solvent is a pure amine with no diluent. Functionalized variations of ethylenediamine were proposed by the authors, with the functional groups being added with the goal of reducing viscosity and sustaining high absorption capacity. In the end, the authors were able to obtain an amine which did not become excessively viscous when loaded, though still quite so when compared to traditional aqueous amines.
In regards to phase separating solvents, identifying the effect of adding organic diluents is particularly difficult, since one cannot be sure if the absorption rates change due to phase separation or due to the new organic diluent itself. It is also not clear whether these rates change for the better or for the worse. Karlsson et al. [172] verified higher mass transfer rates in precipitating NMP–AMP systems, while Ye et al. [120] saw that substituting water by NMP in TETA/DEEA demixing systems steadily decreased the absorption rates. Investigating precipitating potassium prolinate + ethanol mixtures, Shen et al. [173], Bian and Shen [174] and Bian et al. [158] have identified that, though mass transfer rates initially increase by opting for alcohol instead of water in their mixtures, those quickly decrease as precipitation starts. Demixing drastically decreases absorption rates in MEA + 2,4,7,10-tetraoxaundecane solvents [175] (a diluent whose structure is strikingly similar to that of TEGDME). Clearly, a better understanding on phase separation effects on CO2 absorption rates is required for those studying biphasic systems.”