Ind. Eng. Chem. Res. 2002, 41, 5455-5458
5455
SEPARATIONS Effect of the Pore-Size Distribution of Lime on the Reactivity for the Removal of SO2 in the Presence of High-Concentration CO2 at High Temperature Shengji Wu,† Md. Azhar Uddin,‡ Caili Su,† Shinsuke Nagamine,† and Eiji Sasaoka*,† Faculty of Environmental Science and Technology and of Engineering, Okayama University, Tsushima-naka 3-1-1, Okayama 700-8530, Japan
To develop a SO2 sorbent that is highly reactive in the presence of high concentrations of CO2 at high temperature, the effect of the pore-size distribution on the reactivity was investigated at 800 °C using a natural limestone, natural lime, modified macroporous lime, and limestone. The modified lime samples were prepared from a kind of natural lime by water-acetic acid swelling and water swelling methods. The modified limestone was prepared from the modified lime by carbonation. Pores smaller than ca. 200 nm in the modified lime and natural lime virtually disappeared, and pores larger than ca. 200 nm also considerably decreased in pore size after carbonization. The reactivity of the modified limestone depended on the degree of development of pores larger than 200 nm, because the macropores provided a diffusion route for SO2 in the sorbent during sulfation. CaO + CO2 f CaCO3
Introduction Limestone is an important high-temperature desulfurization sorbent that can be used for SO2 capture in fluidized-bed coal combustors. In a coal combustor at atmospheric pressure, limestone usually decomposes into CaO and CO2, and then CaO reacts with SO2 to form CaSO4 ,1 as is shown in the following reactions:
CaCO3 f CaO + CO2
(1)
CaO + SO2 + 1/2O2 f CaSO4
(2)
However, in a pressurized fluidized-bed coal combustor, limestone can remain when the partial pressure is high enough to prevent the decomposition of calcium carbonate. Under these conditions, the direct reaction of CaCO3 and SO2 is considered to be the main reaction for SO2 removal.2
CaCO3 + SO2 + 1/2O2 f CaSO4 + CO2
(3)
Furthermore, it was reported that a part of the limestone can decompose to lime in the low CO2 concentration zone in the combustor.3 In this case, the lime formed may be simultaneously converted to CaCO3 and CaSO4 in the zone in which CO2 concentration CO2 is high. * To whom correspondence should be addressed. Telephone: 81-86-251-8900. Fax: 81-86-251-8900. E-mail:
[email protected]. † Faculty of Environmental Science and Technology. ‡ Faculty of Engineering.
(4)
The internal and external structures of the limestone and lime are changed when CaSO4 is produced during this process, because the molecular volumes of CaO, CaCO3, and CaSO4 are 16.9, 36.9, and 46.0 cm3/mol, respectively.4 Particularly, the volume change from CaO to CaSO4 is large enough to shrink the pore size and plug pores in the lime particles (the molar volume ratio CaSO4/CaO ) 2.7).5 However, the volume change during to the conversion from CaCO3 to CaSO4 is considerably smaller than that of the conversion from CaO to CaSO4 (CaSO4/CaCO3 ) 1.2): in the case of the conversion of CaCO3 to CaSO4, it may be expected that the pores may become a little smaller in size and are not plugged by the product. However, because natural limestone usually is a nonporous material,6,7 the diffusion through the layer of CaSO4 on the CaCO3 particle affects the rate of SO2 removal. Hydration and carbonation methods have already been proposed for reactivation of the spent natural lime for the removal of SO2.8,9 We have found a method to prepare a highly active macroporous CaO from limestone and lime particles using acetic acid vapor,10,11 a water-acetic acid mixture (liquid), and also water (liquid).12 In our present work, modified lime was prepared from a kind of natural lime by a mixture of water-acetic acid swelling and water swelling methods. These samples were used to clarify the following questions: Which is more reactive for SO2 removal, CaO or CaCO3? Which is the predominant reaction, carbonization or sulfation? To what degree does carbonization change the pore size of the CaO sorbent? What is the suitable range of pore size for SO2 removal?
10.1021/ie020362a CCC: $22.00 © 2002 American Chemical Society Published on Web 10/04/2002
5456
Ind. Eng. Chem. Res., Vol. 41, No. 22, 2002
Figure 2. Simultaneous carbonation and sulfation in system II. Figure 1. Pore-size distribution of the samples.
Experimental Section Preparation of Macroporous Limes. The raw limestone used in this study was composed of 55.6 wt % CaO, 0.29 wt % SiO2, 0.18 wt % Al2O3, 0.08 wt % Fe2O3, and 0.04 wt % MgO. The limestone was obtained from Yabashi Col. Raw lime was obtained by heating the limestone at 800 °C for 1 h in an air atmosphere. Pre-experiments confirmed that the lime was uniform, and reproducible data could be obtained. For both the water-acetic acid swelling method and the water swelling method, crushed particles with diameter of less than 0.3 mm were used. A total of 1.0 g of the crushed lime particles was mixed with 3 mL of a mixture of water-acetic acid or water, held at 80 °C for 2 h and then dried at room temperature for 5 days. Finally a cylindrical dried sample (ca. 20 mm in diameter and height varying with the composition of the sample) was obtained. Because the reaction between CaO and the mixture of acetic acid and water was very fast in the liquid phase, all of the acetic acid used in the swelling process reacted with CaO completely during the 2 h. The molar content of calcium acetate in the precursor of the swelled lime was calculated as the following: molar percentage of calcium acetate ) 1/2(molar amount of acetic acid in the water-acetic acid mixture)/ (molar amount of raw lime) × 100. All dried samples prepared by the various swelling methods were converted to lime by calcination at 800 °C for 1 h and then crushed and sieved to the particle size of range (average diameter: 2.0 mm). The raw limestone and lime were also crushed and sieved to the same particle size of range (2.0 mm). The pore-size distributions of the fresh and precarbonated samples were mainly measured using a mercury penetration porosimeter (Micromeritics, Auto Pore III) and partially measured using a nitrogen adsorption apparatus (Micromeritics, Gemini 2375). Pore-size distributions of the samples used in this study are shown in Figure 1. Because the pore-size distribution of the samples differed greatly from each other, we could study the effect of the pore-size distribution on reactivity. From the pore-size distributions shown in Figure 1, it was also concluded that we could roughly control the pore-size distribution of the sorbent by the water-acetic acid method. Apparatus and Procedure. The reactivities of the modified lime and precarbonated lime with SO2 were measured at 800 °C using a flow-type thermogravimet-
ric apparatus equipped with a quartz tubular reactor (1.5 cm inside diameter).8 About 25 mg of the sample was placed in a platinum wire net sample holder (1.3 cm diameter). Three mixtures of gases [system I of SO2 (1500 ppm), O2 (3%), CO2 (10%), H2O (10%), and the remainder N2; system II of SO2 (1500 ppm), O2 (3%), CO2 (50%), H2O (10%), and the remainder N2; and system III of CO2 (50%) and the remainder N2] were used as the inlet gases for the reactor in this study. The reactivities of modified lime and the raw lime were measured in system I for 2 h. The reactivities of the modified and precarbonated limes and the raw lime were measured in system II for 2 h just after in situ precarbonation with system III (for 4 h). The reactivities of the modified lime and the raw lime in the copresence of SO2 and 50% CO2 were measured in system II for 2 h. The pore-size distributions of the fresh and used samples were measured using a mercury penetration porosimeter (Micromeritics, Auto Pore III) and a nitrogen adsorption apparatus (Micromeritics, Gemini 2375). Results and Discussion Reactivities of Modified Lime and Raw Lime in the Presence of 50% CO2. Weight gains of the typical lime samples during the simultaneous carbonation and sulfation in system II are shown in Figure 2. Because the rate of weight gain of the sample was induced by both the carbonation and sulfation of the sample in this system, the sulfation rates of the sample during the sulfation could not be evaluated from the data of Figure 2. Therefore, to know the weight of the sulfated sample, the samples used in system II were in situ heated to 1000 °C in the absence of SO2 and CO2 and held at this temperature until the weight change of the sample stopped. Because it was supposed that a part of the sulfated sample decomposed to SOx and CaO, the evolution of SOx from the sample during the heating was measured by Arsenado III (a wet method). However, the amount of evolution of SOx was negligible. From this experiment and the other X-ray diffraction analysis (not shown), it was confirmed that the sample was composed of CaSO4 and CaCO3 was completely converted to a mixture of CaSO4 and CaO after heating. The value of conversion of the sample from CaO to CaSO4 was calculated from the initial sample weight (CaO) and the final weight after heating. Consecutive Carbonization and Sulfation. Weight gains of the typical lime samples during the carbonation in system III and the sulfation in system II are shown in Figure 3. The initial rates of carbonation of the
Ind. Eng. Chem. Res., Vol. 41, No. 22, 2002 5457
Figure 5. Change of the pore-size distribution of the samples by carbonization. Figure 3. Consecutive carbonation and sulfation.
Figure 4. Fractional conversion of the samples to CaSO4 under the different sulfation systems for 2 h.
samples were very fast, and the fractional conversion (molar ratio of CaCO3 formed/initial CaO) of the two samples prepared by water-acetic acid and water swelling methods was about the same value (ca. 93%). However, the conversion of the raw lime was ca. 82%. This value was considerably lower than that of the other samples. This result may suggest that some pores of raw lime were plugged by carbonation. As the sample weight increased because of the sulfation of the sample, the rate of the sulfation was calculated from the weight gain of the sample in system II. This was also confirmed using the same method as that in the case of the simultaneous carbonization and sulfation (the case of Figure 2). The rates of the sulfation of the precarbonated samples differed from each other: The reactivity of the lime modified by the water-acetic acidsand water swellingsmethod was considerably improved. Comparison of the Reactivity of the Modified Sample and the Pore-Size Distribution. Figure 4 shows the fractional conversion of the samples sulfated under different conditions for 2 h (the systems I, II, and III f II). Because the CaO samples could not be converted to carbonate in system I, the results in system I show the reactivity of the CaO sample with SO2. In the system of III f II, the samples were almost converted to CaCO3 by the precarbonation. Therefore, the reactivity could be regarded as the reactivity of the CaCO3 sample. In system II, simultaneous carbonization and sulfation of the sample may occur. The order
of the conversion of the each sample in the three systems was as follows: system I > system II g system III f II. From a comparison of Figures 1 and 4, it was clear that the reactivity of the sample depends on the volume of macropores in the sample. However, no relationship between the reactivity and the surface area of the sample was observed (data were not shown). The dependency of the reactivity of the samples on the volume of macropores may be explained by diffusion of SO2. That is, the macropores provided a diffusion route for SO2 without plugging of pores during sulfation.12 In the case of the samples in which macropores were well developed (for example, water-acetic acid of 28.0%, 37.4%, and 46.7%), the order of the fractional formation of CaSO4 in system II was obviously larger than that in system III f II. This result suggests that both the sulfation and carbonation of CaO occurred in system II. This means that the sulfation of CaO proceeded under the condition in which the concentration ratio of SO2 to CO2 was 15/1000. The reactivity order of the CaCO3 samples correctly coincided with the order of the lime samples (system I). This result suggests that the difference in the poresize distribution (pore size and pore volume) among the lime samples still remained after carbonization and affected the reactivities of the CaCO3 samples. To confirm this hypothesis, changes in the pore-size distribution of typical samples were measured. As shown in Figure 5, the modified lime prepared from the precursor containing 46.7% calcium acetate still possessed a number of macropores larger than ca. 200 nm and retained this after carbonation, but the pore size was considerably shrunk and the pore volume was decreased by carbonation. However, in raw lime the pores virtually disappeared after carbonation. In the case of the modified lime prepared by water swelling, pores smaller than ca. 200 nm also disappeared and a new peak located at ca. 350 nm appeared. It was supposed that this peak was formed because of shrinking of the pores by more than 350 nm, as shown in Figure 5. Figure 6 shows the sulfation rates of the three samples in systems I and III f II. It was clear that the pore-size distribution greatly affected the reaction rate: the reaction rate of the modified CaO sample using a mixture of water-acetic acid (46.7%) was considerably higher than that of the carbonized sample. To confirm the effect of the pore-size distribution of the sample on its reactivity, the change in the pore-size
5458
Ind. Eng. Chem. Res., Vol. 41, No. 22, 2002
enough for SO2 removal, the reactivity of CaCO3 may be on the same order as that of CaO. (2) Simultaneous carbonization and sulfation were observed. (3) Carbonization decreased the pore size of the CaO sample considerably. (4) Pores that are larger than about 300 nm in CaO, which was used as the precursor of CaCO3, may play a significant role in controlling the reactivity. Acknowledgment This work was financially supported by the Ministry of Education, Science, Sports and Culture of Japan through the Grant-in-Aid on Priority-Area Research (Region No. 737, Grant 11218207). Literature Cited
Figure 6. Comparison of the rates of sulfation of CaO and CaCO3.
Figure 7. Pore-size distribution of the spent sorbents.
distribution of the sample during sulfation was measured using the spent CaO and CaCO3 samples (water47% acetic acid). As shown in Figure 7, it was confirmed that the spent CaO sample kept much more volume of the pores in the range from 200 to ca. 5000 nm than the CaCO3 sample did after sulfation. From these results, it is supposed that if the macropores in CaCO3 are developed enough for SO2 removal, the reactivity of the CaCO3 sample may be the same as that of CaO. Conclusion The effect of the pore-size distribution of a lime on its reactivity for the removal of SO2 in high concentrations of CO2 was studied using modified lime samples. The modified lime samples were prepared from a kind of raw natural lime using water-acetic acid swelling and water swelling methods. From this study, the following conclusions were obtained: (1) If the macropores in CaCO3 are developed
(1) Anderson, D. C.; Anderson, P.; Galwey, A. K. Surface Textural Change during Reaction of CaCO3 Crystals with SO2 and O2 (Air). Fuel 1995, 74, 1024. (2) Mohammad, R. H.; Longwell, J. P.; Sarofin, A. F. Analysis and Modeling of the Direct Sulfation of CaO3. Ind. Eng. Chem. Res. 1988, 27, 2203. (3) Abe, R.; Sasatsu, H.; Harada, T.; Misawa, N.; Tsuji, Y.; Goto, H. Emission of SO2 and Desulfurization Mechanism in 71Mwe PFBC Demonstration Plant. J. Jpn. Inst. Energy 2000, 79, 57. (4) Yrjas, K. P.; Cornelis, A. P.; Hupa, M. M. Hydrogen Sulfide Capture by Limestone and Dolomite at Elevated Pressure. 1. Sorbent Performance. Ind. Eng. Chem.. Res. 1996, 35, 176. (5) Hartman, M.; Coughlin, R. W. Reaction of sulfur Dioxide with Limestone and Influence of Pore Structure. Ind. Eng. Chem. Process Des. Dev. 1974, 13, 248-253. (6) Zevenhoven, R.; Yrjas, P.; Hupa, M. How Does Sorbent Particle Structure Influence Sulfur Capture under PFBC Conditions? 13th International Conference on Fluidized Bed Combustion, Orlando, FL, May 1995; p 1381. (7) Naruse, I.; Nishimura, K.; Otake, K. Characteristics of Desulfurization Reaction by Shells. Kagaku Kogaku Ronbunshu 1995, 21, 904. (8) Counturier, M. F.; Marquis, D. L.; Steward, F. R.; Volmerange, Y. Reaction of Partially-Sulfated Limestone Particles from a CFB Combustor by Hydration. Can. J. Chem. Eng. 1994, 72, 1. (9) Agnihotri, R.; Shriniwas, S. C.; Mahuli, S.K.; Fan, L.-S. Sorbent/Ash Reactivation for Enhanced SO2 Capture Using a Novel Carbonation Technique. Ind. Eng. Chem. Res. 1999, 38, 812. (10) Sasaoka, E.; Uddin, M. A.; Nojima, S. Novel Preparation Method of Mcroporous Lime from Limestone for High-Temperature Desulfurization. Ind. Eng. Chem. Res. 1997, 36, 3639. (11) Sasaoka, E.; Sada, N.; Uddin, M. A. Preparation Method of Mcroporous Lime from Natural Lime by Swelling Method with Acetic Acid for High-Temperature Desulfurization. Ind. Eng. Chem. Res. 1998, 37, 3943. (12) Wu, S.; Sumie, N.; Su, C.; Sasaoka, E.; Uddin, Md. A. Preparation of Macroporous Lime from Natural Lime by Swelling Method with Water and Acetic Acid Mixture for Removal of Sulfur Dioxide at High-Temperature. Ind. Eng. Chem. Res. 2002, 41, 1352.
Received for review May 16, 2002 Revised manuscript received August 21, 2002 Accepted August 21, 2002 IE020362A
