Characterization of diethylenetriamine (DETA) as absorbent for CO2
MetadataVis full innførsel
Absorption of CO2 with amine-based absorbents is an established and proven technology. Unfortunately, it is still very energy intensive and has high capital costs. The overall challenge when aiming at using this technology for world wide CO2 capture, is to bring these two factors down with new and environmentally acceptable solvents. The search forward can be carried out by process design improvements or by finding new and better solvents. An ideal solvent should have a high capacity, high absorption rates, low enthalpy of absorption, result in low steam consumption, give negligible problems with degradation, corrosion and foaming, be non-volatile, as well as being environmentally benign. The main objective of this work is to gather further fundamental information (chemical, kinetic and equilibrium) related to the characterization of DETA as a possible absorbent for CO2. The work was started by measuring physiochemical properties, density and viscosity for different concentrations, temperatures, loaded with CO2 and partially neutralized with H2SO4 in an Anton Paar Stabinger Viscometer SVM 3000. The obtained results for density and viscosity were compared to available sources and modeled with a Redlich-Kister Model in order to determine the thermodynamic properties (partial molar volume and partial molar volume at infinite dilution). The variations in shape of the density and viscosity curves for aqueous amine solution were found very interesting when compared to the different functionalities in the molecular structures. N2O solubilities in the various amine solutions were measured in a solubility cell. Validation of the apparatus and experimental procedure was done by measuring solubility for N2O and CO2 in water and for N2O in pure amines (MEA and MDEA). The results agreed very well with the literature. The solubility measurements were performed for different DETA concentrations, temperatures, or DETA loaded with CO2, and partially neutralized with H2SO4. The commonly used ‘N2O Analogy’ has been validated further for amines (MEA and DETA) fully neutralized with H2SO4 as various concentrations. The solubility model of Wang et al. (1992) was implemented in aqueous DETA solution but more parameters were suggested in this system, thus the extended Wang model was proposed. The model based on a Redlich-Kister approach was developed and gave a satisfactory fit. The effect of CO2 on the N2O solubility into aqueous 30 wt % MEA and DETA were also performed to see the salting-out effect in the solution. Through NMR studies, both qualitative and quantitative measurements were performed in order to elucidate the speciation in the DETA-H2O-CO2 system. 22 species were found to exist in the system and most of them could be identified apart from tricarbamate and free CO2. Quantitative 13C NMR was performed to measure the concentrations of species. The results suggested that this method is reliable in measuring the concentrations, however it was very time consuming experimentally, caused by the insensitivity and low natural abundance of carbon nuclei. Two important observations regarding this technique are the long relaxation time and the need for suppressing enhanced signals from neighboring nuclei (NOE effect). In VLE modeling, a Deshmukh-Mater model was tried implemented to the DETA-H2O-CO2 system at 2.5 M for predicting CO2 partial pressures against loading at different temperatures. Two steps are needed to fit the parameters in DM model. The first step was to determine the temperature dependency of the equilibrium constants by setting short-range interactions to zero. Secondly, to determine the ij b parameters, which showed significant sensitivity, while keeping the equilibrium constants obtained from the first step. However, in this work only the first step was executed because, with the complexity of the DETA system, the available experimental data material was not large and varied enough to merit going into this procedure. The obtained speciations from the VLE model were compared to NMR results to check the quality of the model. The model fit to the CO2 partial pressure data became acceptable only after fitting also the DETA protonation constants, indication the importance of proper values for these equilibrium constants. The kinetics of the reactions of carbon dioxide in aqueous amine systems were measured for unloaded solutions (MEA and DETA) and for DETA partially neutralized with H2SO4. The measurements were performed in a string of discs contactor for different temperatures and amine concentrations. The pseudo-first order approximation was used to obtain the observed reaction rates and two reaction mechanisms, i.e. the termolecular and zwitterion mechanisms were used to interpret the experimental data. Both mechanisms gave in principle identical results when the zwitterion deprotonation was assumed rate determining and the zwitterion formation rate constant set to infinity. However, the advantages of the termolecular mechanism are that it basically has less parameters and more robust in terms of determining the parameters statistically. The introduction of an additional parameter as in the zwitterion mechanism leads to noise in determining the parameters. The kinetic results suggest that DETA has significantly faster kinetics than AEEA, EDA and MEA, but slower than Piperazine. The kinetic rate constant of the secondary amine group in DETA was also determined and found to be about two orders of magnitude lower than for the primary amine group. However, it should be pointed out that the measured secondary amine group rate constant may not represent the true value for CO2 absorption as the neutralization of the two primary amine groups was done with H2SO4 and not CO2. In addition, a validation of the pseudo-first order approximation by using the penetration model for the fast amines MEA and DETA was performed. The reduction in MEA and DETA concentrations at the interface was found to be less than 1%. This indicated that the pseudo-first order approximation is acceptable.