J. Phys. Chem. B 2008, 112, 4323-4328
4323
Influence of Additives Including Amine and Hydroxyl Groups on Aqueous Ammonia Absorbent for CO2 Capture Jong Kyun You,† HoSeok Park,† Seong Ho Yang,‡ Won Hi Hong,*,† Weonho Shin,‡ Jeung Ku Kang,‡ Kwang Bok Yi,§ and Jong-Nam Kim*,§ Department of Chemical and Biomolecular Engineering (BK21 Program) and Department of Materials Science and Engineering, KAIST, 373-1 Guseong-dong, Yuseong-gu, Daejeon, Republic of Korea, and Chemical Process Research Center, KIER 71-2 Jang-dong, Yuseong-gu, Daejeon, Republic of Korea ReceiVed: NoVember 22, 2007; In Final Form: January 23, 2008
Aqueous ammonia absorbent (10 wt %) was modified with four kinds of additives (1 wt %) including amine and hydroxyl groups, i.e., 2-amino-2-methyl-1-propanol (AMP), 2-amino-2-methyl-1,3-propandiol (AMPD), 2-amino-2-ethyl-1,3-propandiol (AEPD), and tri(hydroxymethyl) aminomethane (THAM), for CO2 capture. The loss of ammonia by vaporization was reduced by additives, whereas the removal efficiency of CO2 was slightly improved. These results were attributed to the interactions between ammonia and additives or absorbents and CO2 via hydrogen bonding, as verified by FT-IR spectra and computational calculation. Molecular structures as well as binding energies were obtained from the geometries of (ammonia + additives) and (ammonia + additives + CO2) at the optimized state. These experimental and theoretical findings demonstrate that additives including amine and hydroxyl group are suitable for modifying aqueous ammonia absorbent for CO2 removal.
Introduction CO2 absorption technology is of prime importance for the prevention of global warming and worldwide climate change. Recently, aqueous ammonia solution was suggested as an alternative absorbent of alkanolamine solutions for CO2 capture,1-4 due to its lower cost, higher CO2 absorption capacity, lower decomposition temperatures of ammonium bicarbonate, and a less corrosive environment for the absorber material.5 From an applicative point of view, however, the application of aqueous ammonia to the CO2 absorption process faces the following obstacles: The operation of the aqueous ammonia process requires special consideration of the vaporization of NH3 due to its highly vaporizing nature.2,6 Although a high NH3 concentration solution enables a high CO2 absorption efficiency, it causes ammonium ions to be lost as ammonia vapor, resulting in a reduction of CO2 absorption by means of a lower concentration of ammonia absorbent.7 In addition, crystal precipitates formed by the reaction between the vaporized NH3 and CO2 gas could cause scales on the wall or plugging of the pipes and valve around the top part of the absorber, which should be considered in the design of the CO2 absorber. The formulation of ammonia absorbent with functional additives feasibly could reduce the loss of NH3 from the absorbent solution. Additives such as primary, secondary, tertiary, and sterically hindered amines, piperazine, KOH, NaOH, ionic liquids, etc., have been extensively used to modify the properties of absorbents.8-11 Adding salts can reduce the loss of NH3 by means of lowering the partial pressure of NH3 in the absorbent. However, salts were excluded as potential additives, because they can strongly affect the solubility of CO2 * Corresponding authors. Tel.: +82-42-869-3959. Fax: +82-42-8693959. E-mail:
[email protected] (W.H.H.);
[email protected] (J.-N.K.). † Department of Chemical and Biomolecular Engineering, KAIST. ‡ Department of Materials Science and Engineering, KAIST. § Chemical Process Research Center.
into the modified absorbent. Functional groups of additives play a crucial role in improving the performance of a pristine absorbent, as the physicochemical properties of blended absorbents are strongly influenced by intermolecular interactions between pristine absorbents and additives.12 In this work, we selected additives including both amine and hydroxyl groups for an aqueous ammonia absorbent, utilizing the following strategies: (1) decreasing the loss of ammonia and (2) maintaining or enhancing CO2 removal efficiency of ammonia. Considering the heat required for regeneration of the modified absorbent, we chose sterically hindered amines such as AMP, AMPD, AEPD, and THAM. The regeneration energy of a sterically hindered amine is lower than other conventional amine absorbents such as monoethanolamine (MEA), diethanolamine (DEA), and so on. Since sterically hindered amines have a small heat of CO2 absorption and low CO2 loading under the same partial pressure and regeneration temperature, their regeneration can be achieved more easily than that of conventional amine absorbents. Herein, we examined the NH3-additive-CO2 system experimentally and theoretically in order to correlate interactions between ammonia and additives with the performance, i.e., the loss of absorbent and CO2 capture efficiency, of the blended absorbent. The performances of the modified absorbent were compared to those of aqueous ammonia without additives. Furthermore, the geometry and interaction between the ammonia and additives were simulated by density functional theory (DFT), as supported by FT-IR data. Experimental and Calculation Section Chemicals. Ammonia solution was obtained from Junsei (28 wt %), which was diluted with the deionized water. 2-Amino2-methyl-1-propanol (AMP, MW: 89.14), 2-amino-2-methyl1,3-propandiol (AMPD, MW: 105.14), 2-amino-2-ethyl-1,3propandiol (AEPD, MW: 119.16), and tri(hydroxymethyl)
10.1021/jp711113q CCC: $40.75 © 2008 American Chemical Society Published on Web 03/19/2008
4324 J. Phys. Chem. B, Vol. 112, No. 14, 2008
Figure 1. Removal efficiency of aqueous and modified ammonia absorbent (298 K, 1 atm, 10 wt % aqueous ammonia solution, 1 wt % additive).
SCHEME 1: Schematic Diagram of Reactor System for CO2 Scrubbing
You et al. the absorbent was measured at 298 K. The bubble blowing apparatus, which has an inlet port attached to the glass filter for CO2 bubbling and mixing, is composed of a glass bottle with a 45 mm i.d., containing 200 mL of absorbent aqueous solution. To remove entrapped ammonia and precipitates, sulfuric acid and ice traps were set up successively. The mixed gas of 15 vol % CO2 and 85 vol % N2 was used as feed gas. The flow of feed gas was controlled at a rate of 970 mL/min by mass flow controllers (MFC, MKS 2259C). The concentration of CO2 was measured by a CO2 analyzer. Loss of Absorbent. The loss of absorbent was measured by two methods under controlled conditions, i.e., 313 K, 1 atm, and 60% relative humidity. First, the changes in the weights of the absorbents by vaporization were measured as a function of time. The weight loss percent of absorbent was then calculated by eq 1. Weight of absorbent at the initial and measured time was obtained by subtracting the weight of water from that of the aqueous absorbent solution. We assumed that the content of vaporized water (0.073 atm at 313 K) was negligible due to the considerably lower vapor pressure compared to that of ammonia (15.337 atm at 313 K).
weight loss % ) weight of absorbent at the measured time - weight of absorbent at the initial time × 100 (1) weight of absorbent at the initial time Second, the gas stream after CO2 absorption was analyzed as a function of reaction time by gas chromatography (GC), which was performed using an HP 5890 with porapak N (50/80 mesh) as a packed column and a thermal conductivity detector (TCD). The vapor of the sample was collected into a Tedlar bag. All samples with a constant volume of 100 µL were injected into the GC port. The loss of ammonia from the absorbent was obtained from eq 2.
normalized area ) peak area of ammonia at measured time (2) peak area of ammonia in 10 wt % AM at 10 min
aminomethane (THAM, MW: 121.14) were used as additives. They were supplied by Sigma-Aldrich (99%) and used with no further purification. In the case of the aqueous ammonia absorbent, the concentration of ammonia was 10 wt %, as this is the known optimum concentration of an ammonia aqueous solution for CO2 absorption.6 In the case of the blended ammonia absorbent, the concentration of ammonia in solution was 10 wt % and the concentration of additive was 1 wt %. Therefore, the final concentration of the blended ammonia absorbent was 11 wt % including ammonia of 10 wt % (AM) and additive of 1 wt % in aqueous solution. CO2 Capture Efficiency of Absorbent. A schematic diagram of the experimental system for investigating the CO2 absorption capacity of the absorbents is presented in Scheme 1. The temperature of the solution was kept within (0.5 K from the set-point during experiments by circulating water through a water jacket around the reactor. The CO2 capture efficiency of
FT-IR Spectra. A spectroscopic instrument was used to analyze the reaction as follows. FT-IR spectra were collected on a JASCO FT-IR 4100. The pressure was set as equal for all samples to avoid differences caused by the pressure and penetrating depth. Each spectrum, which was recorded as the average of 13 scans with a resolution of 4 cm-1, was collected from 4000 to 650 cm-1. Then, the spectrum data in the regions other than 1900-900 cm-1 were omitted due to the independence of peak from reaction. Calculation Section. Interaction of ammonia with four additives was examined through DFT calculations.13,14 After geometry optimizations of ammonia and four additives, interactions between ammonia and four additives were calculated. For all calculations, full geometry optimizations were carried out at the B3LYP1,2/6-311G* level, and the binding energies of ammonia with four additives in the optimized structures at standard state were obtained from the following expression:
Eb ) E[ammonia + additive] E[ammonia] - E[additive] (3) where E[ammonia + additive] is the total energy of ammonia with the additive, E[ammonia] is the total energy of ammonia, and E[additive] is the total energy of the additive. In the case of ammonia and CO2 with four additives, full geometry optimizations were also carried out at the B3LYP1, 2/6-311G*
Absorbent for CO2 Capture
Figure 2. (a) Weight loss of aqueous ammonia solution by evaporation (313 K, 60% relative humidity, 10 wt % aqueous ammonia solution, 1 wt % additive). (b) Weight loss of aqueous ammonia solution during CO2 absorption reaction by evaporation (313 K, 10 wt % aqueous ammonia solution, 1 wt % additive, normalized area ) measured peak area/peak area of 10 wt % aqueous ammonia solution at 10, 30, and 60 min).
level, and their binding energies in the optimized structures at standard state were obtained from the following expression:
Eb ) E[ammonia + CO2 + additive] E[ammonia] - E[CO2] - E[Additive] (4) where E[ammonia + CO2 + additive] is the total energy of ammonia with the additive and E[CO2] is the total energy of carbon dioxide. Results and Discussion Removal Efficiency and Loss of Ammonia-Based Absorbents. In order to evaluate the CO2 absorption capacity of the modified absorbents, the CO2 removal efficiency was compared to that obtained using aqueous ammonia without additives. Figure 1 shows the CO2 removal efficiencies of the aqueous and blended ammonia absorbents during a CO2 absorption reaction of 120 min. All results for CO2 removal efficiency were averaged in an error range of (2%. Adding 1 wt % additives to the absorbent only slightly affected the CO2 absorption behavior, and thus it was determined that four candidates could serve as successful additives with this concentration. In comparison to the aqueous ammonia without any additive, all
J. Phys. Chem. B, Vol. 112, No. 14, 2008 4325
Figure 3. IR spectra of ammonia absorbents blended by additives (a) before and (b) after CO2 absorption.
modified absorbents maintained or slightly enhanced the CO2 removal efficiency. In this measurement range, total CO2 removal capacities of aqueous and blended ammonia calculated from respective areas of curves are as follows: AM < (AM + AMPD) < (AM + AEPD) < (AM + AMP) < (AM + THAM). The enhanced CO2 removal efficiencies of blended ammonia absorbents were attributed to intermolecular interactions between the additives and CO2. The loss of ammonia was measured with and without CO2, as shown in Figure 2. All results for the loss of ammonia were averaged in an error range of (1.5%. At the initial vaporizing time, all of the aqueous and blended ammonia absorbents showed dramatic loss without the reaction with CO2. The loss rate of ammonia-based absorbents decreased as a function of the reaction time during the CO2 absorption reaction. These results indicate that the loss of ammonia was dominant at the initial state due to fast evaporation and gradually decreased owing to the reduction of the concentration of ammonia. In particular, four additives played an important role in decreasing the loss of ammonia irrespective of the existence of CO2. Their performance is likely due to the interactions between the hydroxyl groups of the additives and ammonia via hydrogen bonding. The loss of ammonia decreased according to the following sequence: AM > (AM + AMPD) > (AM + AEPD) > (AM + AMP) > (AM + THAM).
4326 J. Phys. Chem. B, Vol. 112, No. 14, 2008
You et al.
Figure 4. Geometry of ammonia absorbent and (a) AMP, (b) AMPD, (c) AEPD, and (d) TMAM (298 K, 1 atm) at the optimized state calculated by the Gaussian-type orbital 6-311G*.
Interactions between Ammonia and Additives. FT-IR spectra were measured to elucidate the interactions between blended ammonia and CO2 as well as between aqueous ammonia and additives. Figure 3a shows IR spectra of aqueous and blended ammonia. The existence of additives in the blended absorbents was confirmed by the appearance of additional peaks related to a C-H stretching band in a range of 2700-3000 cm-1. Furthermore, a shoulder peak around 1100 and new peaks at 1250-1400 cm-1 were also related to the presence of additives (see Supporting Information Figure S1). The broadening and shift of three peaks corresponding to ammonia solution, around 3400, 1630, and 1100 cm-1, were attributed to hydrogen bonding between ammonia and amine/or hydroxyl groups of additives. After CO2 absorption, bands attributed to the amine/ or hydroxyl groups of the additives and aqueous ammonia were broadened and shifted due to intermolecular interactions between CO2 and functional groups of the aqueous and blended ammonia absorbents. In particular, the changes in two peaks at 1100 and 1050 cm-1 indicate that both ammonia and additives played a role in capturing CO2 during absorption. Figure 4 presents the geometry of ammonia and four additives at the optimized state. Hydrogen bonding between hydroxyl groups of additives and ammonia determined the geometry at the optimized state. Taking into consideration that hydroxyl groups of respective additives interacted with ammonia molecules via hydrogen bonding consisting of (hydroxyl group of additives donating hydrogen) O-H‚‚‚N (nitrogen of ammonia accepting hydrogen), the geometries of ammonia molecules that interacted with individual additives were strongly influenced by these intermolecular interactions. Information about the molecular structures of blended ammonia absorbents such as bond angles, bond distances, and dihedral angles was obtained from the geometries at the optimized state (Supporting Information Tables S1, S2, S3, and S4). In the geometries of the blended ammonia absorbents, the bond distances between ammonia and additives via hydrogen bonding were d ) 1.88 Å for AM with AMP, d ) 1.89 Å for AM with AMPD, d ) 1.88 Å for AM with AEPD, and d ) 1.86 Å for AM with THAM, respectively. Table 1 shows the binding energies of ammonia and the four
TABLE 1: Binding Energies of Ammonia with AMP, AMPD, AEPD, and THAM, and Ammonia and CO2 with AMP, AMPD, AEPD, and THAM at the Optimized State Calculated by the Gaussian-Type Orbital 6-311G* (298 K, 1 Atm) binding energy (kJ/mol) AM-AMP AM-AMPD AM-AEPD AM-THAM
-57.470 -54.380 -54.722 -62.247
binding energy (kJ/mol) AM-AMP-CO2 AM-AMPD-CO2 AM-AEPD-CO2 AM-THAM-CO2
-21.449 -29.817 -17.091 -17.518
additives. Binding energies of ammonia with AMP, AMPD, AEPD, and THAM were -57.470, -54.380, -54.722, and -62.247 eV, respectively. When the strength of interaction between ammonia and the additives was increased, the loss of ammonia by vaporization was reduced. These theoretical results were consistent with those of the experiment. Figure 5 shows the geometry of ammonia, CO2, and the four additives at the optimized state. The geometry of ammonia, CO2, and the four additives at the optimized state was strongly influenced by hydrogen bonding between CO2 and the additive as well as between CO2 and ammonia. Bond distances between the additives and CO2 via hydrogen bonding were d ) 3.08 Å for AMP (H9)-CO2 (O22), d ) 2.76 Å for AMPD (H4)-CO2 (O25), d ) 2.63 Å for AEPD (H26)-CO2 (O10), and d ) 2.68 Å for THAM (H10)-CO2 (O24), respectively, whereas bond distances between ammonia and CO2 via hydrogen bonding were d ) 2.53 Å for AM (H20)-CO2 (O24)-AMP, d ) 2.47 Å for AM (H22)-CO2 (O23)-AMPD, d ) 2.53 Å for AM (H25)-CO2 (O28)-AEPD, and d ) 3.21 Å for AM (H20)CO2 (O26)-THAM, respectively. Molecular structures of the blended ammonia absorbents with CO2 such as bond angles, bond distances, and dihedral angles were obtained from the geometries at the optimized state (Supporting Information Tables S5, S6, S7, and S8). In particular, the binding energies of ammonia and CO2 with individual additives were -21.449 eV for AMP, -29.817 eV for AMPD, -17.091 eV for AEPD, and -17.518 eV for THAM, respectively. Ammonia with THAM yielded the highest value among all CO2 removal efficiencies
Absorbent for CO2 Capture
J. Phys. Chem. B, Vol. 112, No. 14, 2008 4327
Figure 5. Geometry of ammonia, CO2, and (a) AMP, (b) AMPD, (c) AEPD, and (d) THAM (298.15 K, 1 atm) at the optimized state calculated by the Gaussian-type orbital 6-311G*.
of pristine and blended absorbents despite its lower binding energy with CO2 compared to that of ammonia, CO2, and AMP (or AMPD). As such, it appears that the superior functionality of THAM, which contains three hydroxyl groups and one amine group, is related to the total content of CO2 capture. In contrast, the maximum CO2 removal efficiencies of aqueous and blended ammonia are as follows: AM < (AM + THAM) < (AM + AEPD) < (AM + AMP) < (AM + AMPD), as shown in Figure 1. The strength of binding energy in ammonia, CO2, and the additive was fairly consistent with the maximum CO2 removal efficiency. Consequently, four additives contributed to a slight improvement of CO2 capture via intermolecular interactions with CO2. CO2 liberation from an absorbent as well as CO2 capture is an important step for CO2 management by absorption in the power generation sector. The dissociation energies of CO2 from blended absorbents should be considered for solvent regeneration in terms of energy consumption. Regeneration energies of blended absorbents are attributed to regeneration energies of their parent absorbents and mixing ratio.15 In the blended absorbent system, the amount of additive in the total absorbent was 9.13 wt %. AMP was 1.88 mol %, AMPD was 1.60 mol %, AEPD was 1.42 mol %, and THAM was 1.39 mol %. Therefore, the regeneration energies of ammonia could dominantly contribute to those of the blended absorbents. The binding energies of the blended absorbents with CO2 except for AM + AMPD + CO2 were lower than the lowest dissociation energy of ammonia for CO2 release (reaction from ammonium bicarbonate into ammonium carbonate and CO2: 26.79 kJ/mol).7,16 Further studies on the dissociation energy including chemical reactions and physical absorption are necessary to estimate energy consumption theoretically. Taking into consideration that the heat of reaction for the MEA process has been reported to be 83.72 kJ/mol,17 blended absorbents can be more beneficial than MEA as a representative CO2 absorbent in terms of energy consumption.
Conclusion We have demonstrated that four additives (AMP, AMPD, AEPD, and THAM) containing amine and hydroxyl groups are effective modifiers for a blended aqueous ammonia absorbent for CO2 removal. The loss of ammonia by vaporization was decreased according to the following sequence: AM < (AM + AMPD) < (AM + AEPD) < (AM + AMP) < (AM + THAM). In comparison to aqueous ammonia without any additive, all modified absorbents exhibited slightly improved CO2 removal efficiency. These experimental results are consistent with computational calculations. The geometry and binding energies derived by DFT revealed that the reduction of the loss of ammonia and the enhancement of CO2 capture were related to intermolecular interactions between the additives and ammonia/or CO2, as verified by FT-IR spectra. In particular, molecular structures of ammonia with additives and blended ammonia absorbents with CO2 at an optimized state, such as bond angles, bond distances, and dihedral angles, were strongly influenced by hydrogen bonding. Taking into consideration that the interactions between functional groups of additives and ammonia/or CO2 influenced the performance of the absorbents, it is suggested that functional blending agents including amine and hydroxyl groups can serve as useful additives for CO2 capture. Acknowledgment. This research was supported by a Grant (DC2-101-1-0-0) from the Carbon Dioxide Reduction and Sequestration Research Center, one of the 21st Century Frontier Programs funded by the Ministry of Science and Technology of the Korea Government. Supporting Information Available: IR spectra, and tables of bond distances, bond angles, and dihedral angles. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Bai, H.; Yeh, A. C. Ind. Eng. Chem. Res. 1997, 36, 2490-2493.
4328 J. Phys. Chem. B, Vol. 112, No. 14, 2008 (2) Li, X.; Hagaman, E.; Tsouris, C.; Lee, J. W. Energy Fuels 2003, 17, 69-74. (3) Yeh, A. C.; Bai, H. Sci. Total EnViron. 1999, 228, 121-133. (4) Diao, Y. F.; Zheng, X. Y.; He, B. S.; Chen, C. H.; Xu, X. C. Energy ConVers. Manage. 2004, 45, 2283-2296. (5) Meng, L.; Burris, S.; Bui, H.; Pan, W.-P. Anal. Chem. 2005, 77, 5947-5952. (6) You, J. K.; Park, H. S.; Hong, W. H.; Park, J.; Kim, J. N. Korean Chem. Eng. Res. 2007, 45, 258-263. (7) Yeh, J. T.; Resnik, K. P.; Rygle, K.; Pennline, H. W. Fuel Process. Technol. 2005, 86, 1533-1546. (8) Ali, S. H. Int. J. Chem. Kinet. 2005, 37, 391-405. (9) Ali, S. H.; Merchant, S. Q.; Fahim, M. A. Sep. Purif. Technol. 2002, 27, 121-136.
You et al. (10) Paul, S.; Ghoshal, A. K.; Mandal, B. Ind. Eng. Chem. Res. 2007, 46, 2576-2588. (11) Seo, D. J.; Hong, W. H. Ind. Eng. Chem. Res. 2000, 39, 20622067. (12) Yoon, S. J.; Lee, H. Chem. Lett. 2003, 32, 344-345. (13) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (14) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (15) Sakwattanapong, R.; Aroonwilas, A.; Veawab, A. Ind. Eng. Chem. Res. 2005, 44, 4465-4473. (16) Yeh, J. T.; Pennline, H. W.; Resnik, K. P.; Rygle, K. Presented at the 3rd Annual Conference on Carbon Capture and Sequestration, Alexandria, VA, May 3-6, 2004. (17) Kohl, A.; Nielsen, R. Gas Purification, 5th ed.; Gulf Publishing Co.