3560
Energy & Fuels 2007, 21, 3560–3562
Comparison of CaO-Based Synthetic CO2 Sorbents under Realistic Calcination Conditions Gemma Grasa,*,† Belén González,‡ Mónica Alonso,‡ and J. Carlos Abanades‡ Instituto de Carboquímica (CSIC), C/Miguel Luesma Castán No.4, 50015, Zaragoza Spain, and Instituto Nacional del Carbón (CSIC), C/Francisco Pintado Fe, No.26, 33011 OViedo, Spain ReceiVed April 4, 2007. ReVised Manuscript ReceiVed June 25, 2007
Several concepts to capture CO2 in power plants and hydrogen generation plants are under development using CaO as regenerable sorbent. The drastic decay in sorbent capture capacity of CaO obtained through calcination of natural sources of CaCO3 (limestones or dolomites) justifies the search of synthetic sorbents that aim to overcome this decay in capture capacity. We have reviewed some of the recent literature on the subject and tested some of the proposed sorbents under comparable conditions. Our results confirm the good performance of some of these synthetic sorbents under mild conditions and/or long carbonation times used in the original references. However, we show that these sorbents deactivate also very quickly when realistic regeneration conditions (high temperatures for calcination at high partial pressures of CO2) are used in the laboratory test. We conclude that none of the reviewed sorbents have a chance to compete with the performance of natural limestones, of much lower cost.
Introduction The carbonation reaction of CaO and CO2 to form CaCO3 and the subsequent calcination to regenerate the CaO are at the core of several emerging CO2 capture processes—both to separate CO2 from a flue gas in combustion processes1–3 or in gasification environments.4,5 The ideal sorbent should be able to react fast (in a few minutes under typical carbonation conditions) toward a maximum capture capacity and maintain a high CO2 capture capacity (in terms of g CO2/g sorbent) with the number of cycles. Low-price natural limestones and dolomites would seem to be the best CaO precursor for largescale CO2 capture in power plants.6 However, the capture capacity of these CaO sorbents decays rapidly with the number of carbonation calcination cycles (see for example Barker7 or Grasa and Abanades8). It has been reported from earlier works7 that a grain sintering mechanism is mainly responsible for the decay of sorbent capture capacity. It is therefore expected that high calcination temperatures (over 950 °C) and/or extended calcination times will accelerate sorbent degradation.8 At the same time, regeneration of the sorbent in a rich CO2 atmosphere is a prerequisite for any process aiming CO2 capture (for * Corresponding author: e-mail
[email protected]; Tel +34 976733977; Fax +34 976733318. † Instituto de Carboquímica (CSIC). ‡ Instituto Nacional del Carbón (CSIC). (1) Shimizu, T.; Hirama, T.; Hosoda, H.; Kitano, K.; Inagaki, M.; Tejima, K. Trans. IChemE 1999, 77A, 62–68. (2) Abanades, J. C.; Anthony, E. J.; Lu, D. Y.; Salvador, C.; Alvarez, D. AIChE J. 2004, 50, 1614–1622. (3) Abanades, J. C.; Anthony, E. J.; Wang, J.; Oakey, J. E. EnViron. Sci. Technol. 2005, 39, 2861–66. (4) Curran, G. P.; Fink, C. E.; Gorin, E. AdV. Chem. Ser. 1967, 69, 141–165. (5) Silaban, A.; Harrison, D. P. Chem. Eng. Commun. 1995, 137, 177– 190. (6) Abanades, J. C.; Rubin, E. S.; Anthony, E. J. Ind. Eng. Chem. Res. 2004, 43, 3462–3466. (7) Barker, R. J. Appl. Chem. Biotechnol. 1973, 23, 733–742. (8) Grasa, G.; Abanades, J. C. Ind. Eng. Chem. Res. 2006, 45, 8846– 8851.
Figure 1. Thermochemical equilibrium of carbonation reaction (ref 9).
subsequent permanent storage). Calcination of CaCO3 in a pure CO2 atmosphere requires temperatures above 900 °C according to the equilibrium (see Figure 1). Steam could be used in the calciner to reduce the regeneration temperature,3 but even in these cases, the need for a fast calcination rate in continuous capture systems will require calcination temperatures well over 850 °C. Synthetic CaO precursors have been proposed to overcome the degradation of CaO from natural sources with the number of cycles and maintain a better capture capacity along the cycles. A common objective is to obtain materials with high surface area and more stable pore structure. Gorin et al.10 first proposed the use of alumina porous particles as carriers of a certain amount of CaO. They obtained good results with this sorbent, 12–20 g CO2 captured/100 g of calcined material for a regenerated particle, even under the demanding operating conditions of the acceptor gasification process (calcination conditions of T > 1000 °C and partial pressures of CO2 and (9) Baker, E. H. J. Chem. Soc. 1962, 70, 464. (10) Gorin, E.; Lancet, M. S.; Curran, G. P. US Patent US4330430; pub 1982-05-18.
10.1021/ef0701687 CCC: $37.00 2007 American Chemical Society Published on Web 09/15/2007
CaO-Based Synthetic CO2 Sorbents
steam over 4 atm). Feng et al.11 have recently proposed a similar technique. They supported very fine CaO powder on inert porous particles of alumina and achieved Ca conversion of 0.90 after nine reaction cycles. These good results are consistent with an earlier work from Barker,7,12 who observed that carbonation–calcination of CaO particles of ∼10 nm in diameter was reversible. The main weakness of these sorbents is the low CaO content that these particles present (4 wt % after two coatings for Feng et al.11). Even when CaO conversion with the number of cycles is high, it makes the total absorption capacity of CO2 by the particles very low, even below natural limestones absorption capacity for a comparable number of cycles (5–6% in weight for residual Ca molar conversion of 7–8 % in Grasa and Abanades8). These low capture capacities may not be acceptable in the process designs of these CO2 capture systems because the heat demand in the calciner will have to increase to heat up (from carbonation to calcination temperatures) the inert material accompanying the CaO circulating between the carbonator and the calciner. Therefore, sorbents with high Ca conversion as well as high Ca mass fraction are intrinsically more attractive for any CO2 capture system using CaO. A different approach to improve the stability of CaO sorbents is to use additives to slow down the sintering process.13,14 Reddy and Smirniotis13 doped the sorbents with alkali metals (Li, Na, K, Rb, and Cs) with different precursors and the solids prepared (with two batches of solids with 10 and 20 wt % alkali/CaO content) were tested under a wide range of carbonation temperatures. Promising results, 50 wt % CO2/wt sorbent for the 20% Cs/CaO sorbent at 600 °C and extended reaction time during the first calcination/carbonation cycle, were found (with calcination temperatures around 700 °C, in N2). Calcium titanate (CaTiO3) has also been used as an additive to CaCO3 to design suitable reversible sorbents for thermal energy storage via carbonation/calcination reaction.14 Reversibility of carbonation reaction was proved along 10 cycles (750 °C and calcination atmosphere of pure N2), improving the results with respect to the solids without titanate. Finally, several authors have recently evaluated the performance of different synthetic CaO precursors (some of them already tested for flue gas desulphurisation15,16 in earlier works) like calcium acetate,17,18 Ca(NO3)2 · 4H2O, CaO, and Ca(OH)218 or precipitated calcium carbonate from calcium hydroxide.19 The aim in all cases is to generate materials with a rich micropore structure. Calcium acetate performed the best among the precursors tested in a recent work by Lu et al.18 However, the high capture capacity of the sorbent and its reversibility are only proved when regeneration takes place at mild conditions (700 °C in He) and carbonation reaction is extended long enough (on the order of 300 min per cycle). This last group of promising sorbents has been the subject of experimental study in this work. The previous paragraphs already highlight a gap between the mild sorbent regeneration conditions used during the testing of (11) Feng, B.; Liu, W.; Li, X.; An, H. Energy Fuels 2006, 20, 2417– 2420. (12) Barker, R. J. Appl. Chem. Biotechnol. 1974, 24, 221–227. (13) Reddy, E. P.; Smirniotis, P. G. J. Phys. Chem. B 2004, 108, 7794– 7780. (14) Aihara, M.; Nagai, T.; Matsushita, J.; Negishi, Y.; Ohya, H. Appl. Energy 2001, 69, 225–238. (15) Adánez, J.; García-Labiano, F.; de Diego, L. F.; Fierro, V. Energy Fuels 1999, 13, 440–448. (16) Patsias, A. A.; Nimmo, W.; Gibbs, B. M.; Williams, P. T. Fuel 2005, 84, 1864–1873. (17) Silaban, A.; Narcida, M.; Harrison, D. P. Resour. ConserV. Recycl. 1992, 7, 139–153. (18) Lu, H.; Reddy, E. P.; Smirniotis, P. G. Ind. Eng. Chem. Res. 2006, 45, 3944–3949. (19) Gupta, H.; Fan, L. S. Ind. Eng. Chem. Res. 2002, 41, 4035–4042.
Energy & Fuels, Vol. 21, No. 6, 2007 3561
Figure 2. Capture capacity vs cycle number for the CaO precursors tested in this work. Calcination temperature 700 °C in pure N2; carbonation 30 vol % CO2 in N2 at 700 °C for 100 min.
most of these sorbents and precursors and the requirements to regenerate the sorbent in a real CO2 capture system (high calcination temperatures imposed by the equilibrium of Figure 1). In an attempt to close this gap, we have reproduced in this work similar operation conditions to those reported in some of the original references as well as more demanding calcination conditions that are considered to be more realistic for CO2 capture applications (temperature over 900 °C and high partial pressures of CO2). Experimental Section Three different synthetic CaO precursors (calcium acetate, CaAc2; calcium oxalate, CaC2O4; and calcium hydroxide, Ca(OH)2) were submitted to repetitive carbonation/calcination cycles to study their behavior with respect to CO2 capture in a TG apparatus. Different calcination temperatures have been tested, from very mild conditions (700 °C in N2) up to 950 °C with pure CO2. Experimental conditions for all the tests carried out are specified in the Figure captions. The cyclic tests were carried out in a thermogravimetric analyzer (TGA) especially designed for long multicycle carbonation–calcination experiments. The TGA consisted of a quartz tube (1.8 × 10-2 m i.d.) placed inside a two zone furnaces capable of working at temperatures up to 1000 °C. The sample holder was a platinum basket 8 × 10-3 m in diameter and 2 × 10-3 m in height. Temperature and sample weight were continuously recorded in a computer. The reacting gas mixture (CO2, O2/air, and/or N2) was set by mass flow controllers and fed to the bottom of the quartz tube. A special characteristic in the design of this TGA was the existence of two zones in the furnace capable of working at different temperatures. This furnace can be displaced (by means of a pneumatic piston) up and down. The position respect to the platinum basket alternates between calcination conditions (>700 °C) or carbonation conditions (around 650 °C). For each run in the TGA around 20 mg of sorbent was introduced in the sample holder. Initial experiments were carried out to determine the total gas flow needed to eliminate external diffusion effects around the sample basket. This flow was set at a value higher than 4 × 10-6 m3/s (STP), which corresponds to a spatial gas velocity of 0.08 m/s at 950 °C.
Results Two sets of experiments for each CaO precursor were carried out under mild regeneration conditions but varying carbonation reaction times. Figure 2 shows the results in terms of maximum capture capacity for sorbents regenerated at 700 °C and extended carbonation times (100 min). As can be seen, carbonation reaction is very close to total CaO conversion. This is especially so for CaAc2, which still presents XCaO 0.95 after five calcination/carbonation cycles. CaOx and Ca(OH)2 show a slight sorbent degradation already under these conditions, although
3562 Energy & Fuels, Vol. 21, No. 6, 2007
Figure 3. Capture capacity vs cycle number for CaO precursors tested in this work. Calcination temperature 700 °C in pure N2; carbonation 30 % CO2 in N2 at 700 °C for 5 min.
still over 80% Ca conversion is obtained after eight cycles. These results corroborate the data published by Lu et al.18 and Barker12 for very fine particles and extended carbonation times. A second set of experiments were carried out also at 700 °C but reducing the carbonation time, trying to accommodate it to typical residence times in fluidized bed carbonator reactors. These are expected to be on the order of a few minutes for reasonable values of bed inventories. Results show a substantial decrease in sorbent capture capacity along cycling (see Figure 3). As an example, for cycle 8, CaOx and Ca(OH)2 present XCaO close to 0.6 and 0.83 for CaAc2. As in the previous series, CaAc2 shows the best performance among the CaO precursors tested. Calcium acetate calcines forming highly cenospheric particles with thin and porous walls of CaO resembling “popcorn”-like structures. The high porosities exhibited by these particles greatly increase sorbent utilization for SO2 capture15,16,20,21 and are likely to be responsible also for the improved performance toward carbonation. The final series of experiments with these precursors of CaO were conducted at more realistic calcination temperatures and calcination times (Figure 4). Two different calcination temperatures (850 and 950 °C) have been tested for CaAc2, and 950 °C has been the calcination temperature tested for the other two precursors. As can be seen from Figure 4, carbonation capacity decreases dramatically during the first 20 cycles. Results are better than those obtained with a natural limestone (also included in Figure 4) extensively studied in previous works.8,22 The deactivation mechanism that degrades the initially rich texture of a calcine obtained from a natural source of CaO (limestones and dolomites) must be also present in these precursors.23 This degradation suffered by synthetic precursors was already found by Silaban et al.17 during experiments involving calcination temperatures in the range 750–900 °C, even thought they only tested two carbonation/calcination cycles with CaAc2 at these temperatures. In view of these results, we are pessimistic about the viability of these synthetic CaO precursors for CO2 capture systems. Their improved performance for mild calcination conditions (and especially for long carbonation times) must be associated with (20) Levendis, Y. A.; Zhu, W.; Wise, D. L.; Simons, G. A. AIChE J. 1993, 39, 761–73. (21) Adánez, J.; de Diego, L. F.; García-Labiano, F. Fuel 1999, 78, 583– 592. (22) Abanades, J. C.; Alvarez, D. Energy Fuels 2003, 17, 308–315. (23) Alvarez, D.; Abanades, J. C. Energy Fuels 2005, 19, 270–278. (24) Alvarez, D.; Abanades, J. C. Ind. Eng. Chem. Res. 2005, 44, 5608– 5615.
Grasa et al.
Figure 4. Capture capacity vs cycle number for CaO precursors tested in this work; also results from a natural limestone are included. Calcination temperature 850 °C, 30 vol % CO2 in air for the first series; 950 °C in pure CO2 for the rest.
the very small size and/or special shapes (cenospheres in the case of CaAc2) of the particles or grains of CaO formed during the first calcination. This initial advantage is somehow maintained for a few cycles because the surface area is high and the product layer of CaCO3 that marks the end of the fast carbonation period (about 50 nm as measured in a previous work24) allows high CaO conversions for solids with high surface area. But, attending to the results of Figure 4, the sintering process that is present on every calcination (grain growth of the CaO resulting after every calcination step of CaCO3 and reduction in surface area) must be equally present in these samples. The CO2 capture capacity of CaO from these synthetic precursors is better than natural limestones even at high calcination temperatures, but it is very unlikely that the difference would justify the choice of these materials with respect to much cheaper natural precursors. Also, what emerges from this work is the need for testing any new sorbent at conditions (temperatures, atmospheres, and reaction time) close to the real system before making any claims about improved sorbent performance. Conclusions The rapid decay of CO2 capture capacity of CaO derived from natural limestones and dolomites may justify the search for new improved CaO sorbents for CO2 capture applications. To assess the performance of these new sorbents, the experimental reaction conditions should be as close as possible to operation conditions in the real CO2 capture system, which are fixed by the equilibrium (calcination temperatures around 900 °C in atmosphere of CO2) and reasonable reaction times (on the order of minutes). Synthetic CaO precursors proposed in the literature that claimed to perform well under mild reaction conditions and extended calcinations times have been shown to degrade rapidly when tested under more realistic conditions. Calcium acetate seems to perform the best, showing still XCaO over 0.2 after 100 carbonation/calcination cycles. However other considerations (like price, availability, and handling) need to be taken into account to justify their use instead of CaO from natural limestone. Acknowledgment. This work is partially funded by the European Commission (C3-Capture) and the Spanish Ministry of Education (”Juan de la Cierva” program). The help from E. Fernández during the experimental work is also acknowledged. EF0701687
