Preparation and laboratory tests of low cost solid sorbents for post-combustion CO 2 capture Luigi Mazzocchi*, Maurizio Notaro*, M. Alvarez**, J.C. Ballesteros** S. Burgos**, J. M. Pardos-Gotor** S4FE 2009 – International Conference on "Sustainable Fossil Fuels for Future Energy" Roma, 6-10 Luglio 2009 PRESENTAZIONE POWER POINT * ERSE SPA ** ENDESA S.A Currently, all commercial post-combustion CO 2 capture plants use processes based on chemical absorption with an aqueous alkanolamine solvent; monoethanolamine (MEA) and diethanolamine (DEA) are the most popular solvents. However, they have a number of shortcomings for treating flue gases, including the large amount of heat required to regenerate the solvent and operational problems caused by corrosion and chemical degradation, which result in high capital and running costs (the typical energy penalty incurred by the operation of the MEA and DEA capture processes is an estimated 15–37 % of the net power output of a plant). Chemical absorption on solid sorbents is considered to be one of the most promising technologies for capturing CO 2 from flue gases because it could allow an energy consumption reduction: as a matter of fact, both heat capacity and regeneration temperature of the solid sorbent would be lower. Further, because the amine is anchored on the surface material and not dissolved into the water, absence of corrosion phenomena is expected. The success of such an approach is however dependent on the development of a low cost absorbent that has both a high CO 2 selectivity and absorption capacity, besides a restricted ∆T between the adsorption and desorption temperature. In the literature solid sorbents have been proposed based on amines or polyamine grafted onto a nano- sized solid support or “molecular basket”, in which nano-structured mesoporous silica is employed in order to obtain high adsorption capacity. However, these materials present critical aspect about sorbent preparation: the synthesis of nano-structured mesoporous results complex, very expensive owing to reagents cost and with very low yield. Moreover, the surface functionalisation with amino groups needs many reactions steps and presents difficulties due to the grafted amines thermal stability. In this work solid sorbents preparation through low cost and simple methodologies for TSA application in fixed bed has been carried out. For this purpose, amines have been deposited by impregnation on solid substrates. Low volatility amines, to minimize their losses by evaporation, and supports already supplied in pellets shape have been employed. Commercial supports of organic resins, activated carbons, alumina and silica-alumina with different shape (granular, cylindrical and spherical) and morphological characteristics (specific area surface and porosity) have been employed. Primary amines (CPAHC, chloropropylamine hydrochloride, and 2-aminoethanesulfonate, TAU), secondary amines (DEA, diethanolamine), tertiary amines (TEA, triethanolamine), sterically hindered amines (AMP, 2-amino-2-methyl-1-propano) and polymeric amines (PEI, polyethylenimine) have been investigated as sorption-active phase. The screening tests have been carried out in a fixed-bed absorber/desorber unit consisting of a 400 mm long stainless steel tube (I.D. 60mm), supplied with a porous metallic disk supporting the sorbent pellets. The gas composition that approximately simulates flue gas was created by blending gases using mass flow controllers, pressure regulator and gas cylinders. In a typical CO 2 adsorption experiment, the reactor was loaded with 135 grams of sorbent pellets and the gas inlet composition (v/v) was 10% CO 2 , 3% O 2 , 10% H 2 O vapour, N 2 to balance. Total gas flow rate was set at 60 Nlh -1 corresponding to a space velocity (SV) respectively of 0,44 Nlh -1 g -1 . Tests with low SO 2 concentration (100 ppm) have been also carried out to investigated the effect of this flue gases contaminant on the sorbents degradation.

Among the tested amines, the most effective for CO 2 capture is DEA (on all the investigated supports) and PEI (only on alumina). The other investigated amines have not suitable characteristics for the following reasons: TEA and CPAHC show a very low “net CO 2 capture capacity”, AMP, probably because of its higher vapour pressure, is poorly anchored to the support and tends to be expelled from the solid matrix during the regeneration stage, TAU has a “net CO 2 capture capacity” comparable with DEA in the first absorption/desorption cycle, but it drastically falls in the following ones. The sorbents based on DEA supported on alumina and silica-alumina have generally 1,5÷2 times higher performance and better oxidation resistance than sorbents based on activated carbon. In fact, although the activated carbon sorbents did not show significant degradation during the absorption/desorption cycles; after the ageing in room air over about three months these sorbents show a drastic decreasing of the CO 2 capture capacity (probably due to a catalytic effect of the activated carbon in amine oxidation). On the contrary, in the sorbents based on alumina and silica- alumina supports no degradation has been detected after two month in room air. Except for DEA supported on macroporous alumina, that presents poor sorption efficiency, all the DEA sorbents based on alumina and silica-alumina supports show a good CO 2 capture capacity. The support specific area surface increase does not influence the sorbents CO 2 capture efficiency, whereas larger mesopores diameter and the presence of macroporosity in the range 400-600 nm are able to improve the sorbents performance. For the same range porosity, silica-alumina seems a better DEA support than alumina. The sorbents characterized by the best performance showed a “net CO 2 capture capacity” (the CO 2 amount absorbed per gram of sorbent) of about 4,5 % wt. The sorbents showed also high stability and complete regenerability, as confirmed by the good matching between the amount of the absorbed and desorbed total CO 2 measured on five subsequent absorption/desorption cycles. In the next figure a typical absorption trend for a sorbent based on DEA supported on alumina is reported. DEA/Alumina 0,00 2,00 4,00 6,00 8,00 10,00 12,00 0 500 1000 1500 2000 2500 t (sec) [ C O2 ] out ( % v / v ) Figure1 – Absorption trend for a DEA/Alumina sorbent ([CO 2 ]=10%, [O 2 ]=3%, [H 2 O]=10% (v/v), T=40 °C and total flow rate of 60 Nl/h) The tests carried out in presence of SO 2 (100 ppm) as flue gases contaminant, indicated essentially a good sorbent stability following the interaction with SO 2 contaminant. Although in the absorption step the SO 2 concentration went completely down to zero and it did not release out from the sorbent during the regeneration, the sorbent CO 2 capture doesn’t seem to show any decrease. Nevertheless, the trend of the SO 2 concentration sloped down much more slowly than that one of CO 2 concentration, indicating an absorption competition between CO 2 and SO 2 that is much more favourable to the former in consequence of the considerable concentration difference. For this reason, it is possible to say that the SO 2 amount absorbed from the sorbent is surely more less than that fed to the absorber as a whole. A preliminary energy analysis comparing the CO 2 capture process based on solid sorbents with the industrial method employing a 30% wt. DEA aqueous solution has been carried out. In this evaluation only the CO 2 desorption has been considered since it represents the more energetically expensive stage. The required heat for the sorbent regeneration has been estimated by the following energy balance Q= −∆H reaction + mC p ∆T, where the enthalpy of the carbamate dissociation and the sorbent

3 thermal capacity by mass unity of CO 2 captured are considered. For DEA aqueous system the data concerning an optimized industrial process have been used, while the experimental (Cp and net CO 2 capture capacity) have been used for the DEA supported on solid material. The results of the energy analysis are reported in the following histogram. 1500 5300 1500 4300 1500 2300 1500 950 0 1000 2000 3000 4000 5000 6000 7000 Q r ege ner a t i on ( kJ/ Kg CO2) 17 DEA 30% wt. solution 76 DEA/Organic resin 58 DEA/Activated carbon 23 DEA/Alumina sorbent specific mass (kg/kg CO2) Regeneration energetic consumption thermal capacity (kJ/kg CO2) ∆H reaction (kJ/kg CO2) According to the obtained experimental data, the best sorbent allows in the regeneration step a 65% energy saving in comparison with DEA amine solution process.