(720c) Kinetic Study of Carbon Dioxide Absorption in High CO2-Loaded, Aqueous Diethanolamine Solutions as a Function of Pressure and Liquid Phase in-Situ Speciation

For both energy and environmental purposes, the removal of acid gases such as carbon dioxide (CO₂) is involved in numerous processes such as natural gas treatment, as well as in petroleum refining, coal gasification, and hydrogen production. For such purification, reactive absorption with aqueous solutions of alkanolamines remains the most mature and widely implemented industrial technology.

Accurate understanding and modeling of the transfer and reaction mechanism occurring during the carbon dioxide absorption in alkanolamine solutions is of vital importance for optimizing the design of absorption units and reducing the absorber size.

Most of the kinetic experimental studies devoted to the absorption of CO₂ in aqueous amine solutions have been restricted to the determination of the acid gas absorption rate by following the partial pressure drop of carbon dioxide during absorption by alkanolamine solvent (Blauwhoff et al. (1984), Versteeg and van Swaaij (1988), Rinker et al. (1996); Aboudheir et al. (2003), Jamal et al. (2006)?). Nevertheless, as the nature and concentrations of the different components in the liquid phase depend on various chemical reactions and their equilibria, obtaining speciation data during the absorption process may be very useful in the development of more realistic absorption models. In addition, all authors limited their experimental kinetic studies to the absorption of carbon dioxide into unloaded amine solutions except Aboudheir et al. (2003) who studied the carbon dioxide absorption into CO₂-loaded methanolamine solutions.

The objective of the study is to acquire original experimental data providing information about the reaction product distribution of the different species in the liquid phase in order to investigate the reaction mechanism occurring during the absorption of CO₂ by loaded amine solutions. Therefore, experimental kinetic measurements on CO₂-loaded DEA aqueous solution were conducted in a new developed equipment that combines pressure measurements in the gas phase with an in-situ spectroscopic analysis of the liquid phase. This original experimental device consist in a stirred-cell reactor equipped with two four-bladed impellers to allow efficient mixing of gas and liquid phases. An attenuated total reflection (ATR) cell with a diamond crystal is fixed at the bottom of the reactor. The whole equipment is located within a Fourier transform infrared spectrophotometer enabling the online recording of the infrared spectrum of the solution. This new apparatus is designed to monitor continuously the CO₂ pressure in the vapor phase, as well as the liquid bulk speciation during the absorption of the CO₂ in amine aqueous solutions.

First, the hydrodynamics of the reactor have been characterized through the physical absorption of N₂O into DEA aqueous solutions. This allowed to establish an expression of the liquid-phase mass-transfer coefficient as a function of the physical properties of CO₂ loaded, DEA solutions and stirring speed. Then a method has been developed in order to determine the concentrations of the different species present in the liquid bulk from the ATR-IR spectra. In a first step, reference spectra and response coefficients were determined for each species thanks to standard aqueous solutions or pure DEA CO₂-loaded solutions for the carbamate. Then, a set of experiments has been carried out in a 40 wt%-DEA aqueous solution at 313.15 K and 333.15 K at various CO₂ loadings. The IR intensity was monitored at 1530 cm^-1 (CO₂^- antisymetric stretching band of the carbamate species) and 2340 cm^-1 (vibration band of dissolved carbon dioxide). By considering ATR-IR intensity as a sum of the contributions of each compound and mass balance equations (carbonate species were neglected), concentrations of the different species present in the liquid phase (carbamates, bicarbonates, DEA, protonated DEA and molecular CO₂) as a function of CO₂ loading have been experimentally determined.  It has to be noticed that the determination of the concentration of the molecular form of the dissolved CO₂ requires flow of dry CO₂-free air within the analysis chamber of the spectrophotometer, in order to remove the ?background' from ambient atmospheric CO₂. The concentrations of the different species calculated from the ATR-IR spectrum at different CO₂ loading were in good agreement with the general evolution tendencies of NMR Spectroscopy results published by Böttinger W. et al. (2008).

Furthermore, experimental kinetic measurements were carried out using this equipment considering loaded 40wt%-DEA aqueous solution at 313.15 K and 333.15 K, the CO₂ loading ranging between 0 and 0.8 mol_CO2/mol_DEA. Figure 1 shows the ability of the new apparatus to monitor the pressure drop (figure 1a) as well as the evolution of the liquid phase composition (figures 1b-1 and 1b-2) during the kinetic study of the CO₂ absorption into a loaded 40wt%-DEA aqueous solution at 313.15 K.

In order to interpret the experimental absorption data and estimate kinetic rate coefficients, a numerical mass-transfer model based on the film theory (Whitman 1923) in which all chemical reactions are considered to be reversible was developed. The numerical scheme of differential-algebraic equations was discretized utilizing the finite volume method. The set of obtained equations were solved with the appropriate initial and boundary conditions and were integrated through time using the code DDASSL (Petzold, 1983). Then the variations of the CO₂ pressure, and bulk concentrations, of the various species as a function of time as predicted by this model were compared to the experimental measurements.

The validation of the model have been done on unloaded solutions by considering the zwitterions mechanism (Claplow, 1968; Danckwerts 1979) which has become one of the most accepted mechanism for primary and secondary amine (Versteeg and Oyevaar (1989), Glasscock et al. (1991), Littel et al. (1992), Rinker et al. (1996)?). This mechanism involves the reaction of CO₂ with a primary or a secondary amine to form a zwitterion intermediate that is subsequently deprotonated by a base to produce carbamate and protonated base. The reactions between CO₂and diethanolamine solution have been also described in the literature by the termolecular mechanism introduced by Crooks and Donnellan (1989) where it is assumed that the reaction is a single-step between CO₂and DEA and the initial product is not a zwitterion but a loosely bound encounter complex. These two mechanisms were tested to represent the experimental pressure and speciation data obtained during absorption on loaded CO₂-diethanolamine solutions. Since CO₂-loaded solutions were used in the experimental work, the concentrations of bicarbonates, carbonates and carbamate in the aqueous solutions were considered significant. Hence, the contributions of these species to the reaction mechanism can no longer be neglected and therefore were investigated.

Such additional information on the reactions and the evolution of the component distributions during the CO₂ absorption leads to more accurate understanding and modeling of this commonly used technology for CO₂ capture over a large range of experimental conditions.

Figure 1a:Vapor Phase Evolution: Carbon dioxide partial pressure evolution in the vapor phase during CO₂ absorption

Figure 1b-1: Liquid Phase Evolution: Concentrations of molecular carbon dioxide (CO₂) and protonated amine (DEAH⁺) in the liquid phase during CO₂ absorption

Figure 1b-2: Liquid Phase Evolution: Concentrations of bicarbonate (HCO₃^-), carbamate (DEACOO^-) and diethanolamine (DEA) in the liquid phase during CO₂ absorption

Figure 1: Absorption of CO₂ into a loaded aqueous solution of DEA.. Mass fraction=40%, T= 313.15 K, CO₂ loading = 0.596-0.626 mol_CO2/mol_DEA, P_CO2=320-80 kPa.

Literature Cited

Aboudheir, A., Tontiwachwuthikul, P., Chakma, A., Idem, R. Chemical Engineering Science, 2003, 58, 5195-5210.

Blauwhoff, P. M. M., Versteeg, G. F. and Van Swaaij, W. P. M. Chemical Engineering Science 1984, 39, 207-225.

Böttinger, W., Maiwald, M., Hasse, H. Fluid phase Equilibria., 2008, 263, 131-143.

Crooks, J. E., Donnellan, J. P. Journal of Chemical Society of Perkin Transactions. 2 1989, 331-333.

Glasscock , D. A., Crichfield, J. E., Rochelle, G. T. Chemical Engineering Science, 1991, 46,2829.

Jamal, A., Meisen, A. and Lim, C. J. Journal of Chemical Science 2006, 61, 6571-6589.

Little, R. J., Versteeg, G. F. and Van Swaaij, W. P. M. Chemical Engineering Science 1992, 47, 2027-2035.

Petzold, L. R. Scientific Computing, IMACS/North- Holland Publishing Co., R. Stepleman et al.(eds), 1983, 65-68.

Rinker E. B., Ashour, S. S., Sandall, O. C. Industrial and Engineering Chemistry Research 1996, 35, 1107-1114.

Versteeg, G. F. and Van Swaaij, W. P. M. Chemical Engineering Science 1988 , 43, 573-585.

Versteeg, G. F., Oyevar M. H. Chemical Engineering Science 1989, 44,1264-1268.

Whitman, W.G.Chem & Met. Engng. ,1923, 29, 146-148.