| dc.description.abstract | This work gives contributions to kinetics and equilibrium of CO2 absorption into N,Ndiethylethanolamine (DEEA), N-Methyl-1,3-propane-diamine (MAPA) and their blends. The main emphasis of this thesis, based on the measured experimental data, was on developing mass transfer and kinetic models for the CO2 absorption by aqueous blends of DEEA and MAPA. Additionally, chemical and phase equilibrium models are presented for the respective subsystems.
Chapter 4 presents a literature review of models for calculating the solubility of CO2 and N2O in water, and points at the lack of precise measurements for the solubilities of both CO2 and N2O, especially in the high temperature, low pressure range – which is relevant for the CO2 capture processes in the post-combustion scenario. A new set of experimental data for the N2O solubility in water is given. Literature correlations for the Henry constant of CO2 and N2O are recommended. Alternative, and equally good, correlations are proposed in this work, which have the advantage of using the same modelling premises for both CO2 and N2O.
Chapter 5 presents the VLE models for DEEA-H2O, MAPA-H2O, DEEA-MAPA-H2O and DEEA-H2O-CO2 subsystems, developed in the framework described in Chapter 2. In aqueous solutions, the activity coefficients of MAPA are generally very low, and those of DEEA are very high. The activity of DEEA predicted by the NRTL model shows inflection points in the low-concentration, high-temperature range, indicating that a second liquid phase is formed.
Section 6.1 presents the CO2 absorption rate measurements into unloaded MAPA aqueous solutions. The experimental results were interpreted using the two-film masstransfer model (described in Chapter 3) and invoking the pseudo first order assumption. Needed experimental values for density, viscosity and Henry’s law constant for CO2 were measured and are given. The results indicate that MAPA is almost twice as fast as piperazine, 8 times faster than 2-(2-aminoethyl-amino)ethanol, and 15 times faster than monoethanolamine, when comparing unloaded 1M solutions at 25ºC.
The observed kinetic constant was modelled using the direct mechanism, considering both water and MAPA as possible proton receivers. Very good agreement was obtained between the experimental data and the model. The estimated pre-exponential parameter values for representing the reaction when water is the proton receiver were very high. This is believed to be due to the strong bonds between MAPA and water, evidenced by the NRTL model results.
The mass transfer and kinetics of CO2 reactions in aqueous DEEA solutions are discussed in section 6.2. A literature review reveals that there is little agreement between the sources reporting experimental data and models for the kinetics of the DEEA-catalyzed CO2 hydration reaction. The data were carefully analyzed in this thesis, but the reason behind the scatter in literature data for the mass transfer of CO2 into DEEA solutions still remains unclear.
Two experimental apparatuses were used to generate new data sets: a stirred cell reactor (SCR) and a string of discs contactor (SDC). The data produced in the SCR show an unexpected tendency: the pseudo-first order kinetic constant increases linearly with DEEA concentration only up to approximately 3M, and then starts decreasing.
It is interesting to notice that the scatter in the data produced in the SDC covers the scatter in literature data. Hence, it is likely that the factors behind the uncertainty in the data produced in this work also, at least in part, explain the disagreement between the consulted literature sources.
Two models are used to describe the absorption flux: the two-film model, invoking the pseudo-first order assumption; and the penetration model. It is shown that the models predict very different kinetic constants and that this difference is a function of loading.
Some of the data obtained in the SDC support the fact that the pseudo-first kinetic constant has a maximum close to 3M at 25ºC. Having this result in mind, and analysing the models obtained in Chapter 5, the DEEA clustering hypothesis was formulated, as presented in section 6.3.
It is assumed that, in dilute aqueous DEEA solutions (including aqueous blends of MAPA and DEEA), the DEEA molecules are mostly either in the form of an internal salt-like species or distributed in small DEEA clusters, which do not internally contain CO2. The gas is dissolved in the water between clusters, much as in pure water. Because of the high water coordination number, these small clusters are still reactive toward CO2. With increasing concentration, the clusters become bigger, with intermolecular bonding being favoured. At some concentration (higher than 3M to 4M for aqueous DEEA solutions), the clusters are big enough to have space for CO2 and an increase in the total physical uptake is observed. When the CO2 molecules are dissolved within large DEEA clusters, they have very little access to water molecules. Therefore a decrease in the CO2 hydration reaction rate is observed.
Section 6.4 presents experimental data on the CO2 transfer into aqueous blends of DEEA and MAPA. Additionally, data on density, viscosity and N2O solubility into the studied blends are given. The data are interpreted using the two-film theory and invoking the pseudo-first order assumption. The clustering hypothesis allows for qualitatively explaining the complex behavior that is obtained in the experiments. Blends of the 2M MAPA family, in special D5M2, are shown to have high mass transfer rates and fast kinetics.
The observed kinetic constant was modelled using the direct mechanism, considering water, MAPA and DEEA as possible proton receivers. The contributions of water and MAPA were calculated using the values obtained from the experiments with aqueous MAPA solutions. Based on this, the contribution of DEEA is shown to represent 52 to 98% of the observed kinetic constant. This is not an unreasonable result, since DEEA is a stronger base than either MAPA or water.
Because the behaviour of the 1M MAPA family is distinctly different from that of the 2M MAPA family, one model for the observed kinetic constant for each family had to be provided. Good agreement was obtained between the experimental data and the model, indicating that the direct mechanism can explain the MAPA carbamate formation reaction when DEEA is present. | nb_NO |
