Mechanism of SO2 Removal by Carbon - Energy & Fuels (ACS

Mar 19, 1997 - The reaction of SO2 with carbon (C) in the presence of O2 and H2O involves a series of reactions that leads to the formation of sulfuri...
3 downloads 19 Views 209KB Size
Download PDF
284

Energy & Fuels 1997, 11, 284-291

Mechanism of SO2 Removal by Carbon§ Anthony A. Lizzio*,† and Joseph A. DeBarr†,‡ Illinois State Geological Survey, 615 East Peabody Drive, Champaign, Illinois 61820, and Department of Environmental Engineering and Science, University of Illinois, Urbana, Illinois 61801 Received November 8, 1996. Revised Manuscript Received January 20, 1997X

The reaction of SO2 with carbon (C) in the presence of O2 and H2O involves a series of reactions that leads to the formation of sulfuric acid as the final product. The rate-determining step in the overall process is the oxidation of SO2 to SO3. Three SO2 oxidation reactions are possible. Adsorbed SO2 (C-SO2) can react either with gas phase O2 or with adsorbed oxygen (C-O complex) to form sulfur trioxide (SO3), or gas phase SO2 can react directly with the C-O complex. In optimizing the SO2 removal capabilities of carbon, most studies only assume a given mechanism for SO2 adsorption and conversion to H2SO4 to be operable. The appropriate SO2 oxidation step and role of the C-O complex in this mechanism remain to be determined. The ultimate goal of this study was to prepare activated char from Illinois coal with optimal properties for low-temperature (80-150 °C) removal of sulfur dioxide from coal combustion flue gas. The SO2 adsorption capacity of activated char was found to be inversely proportional to the amount of oxygen adsorbed on its surface. A temperature-programmed desorption technique was developed to titrate those sites responsible for adsorption of SO2 and conversion to H2SO4. On the basis of these results, a mechanism for SO2 removal by carbon was proposed. The derived rate expression showed SO2 adsorption to be dependent only on the fundamental rate constant and concentration of carbon atoms designated as free sites. Recent studies indicate a similar relationship exists between the rate of carbon gasification (in CO2 or H2O) and the number of reactive sites as determined by transient kinetics experiments. Utilizing the concept of active or free sites, it was possible to produce a char from Illinois coal having an SO2 adsorption capacity surpassing that of a commercial catalytic activated carbon.

Introduction There are a number of research groups currently involved in the development of carbon-based processes and materials for removal of SO2 from coal combustion flue gas.1-16 The type of carbon used more often than not dictates the economic viability of a given process. A †

Illinois State Geological Survey. University of Illinois. Paper presented at the National Meeting of the American Chemical Society, New Orleans, recipient of R. A. Glenn Award from the Division of Fuel Chemistry. X Abstract published in Advance ACS Abstracts, February 15, 1997. (1) Richter, E.; Knoblauch, K.; Jungten, H. In Proceedings of 1st International Conference on Processing and Utilization of High Sulfur Coals, Columbus, OH, 1985; p 563. (2) Richter, E.; Knoblauch, K.; Jungten, H. Gas Sep. Purif. 1987, 1, 35. (3) Jungten, H.; Kuhl, H. Chem. Phys. Carbon 1989, 22, 145. (4) Richter, E. Catal. Today 1990, 7, 93. (5) Tsuji, K.; Shiraishi, I. In Proceedings of Electric Power Research Institute SO2 Control Symposium; Washington, DC, 1991; p 307. (6) Lu, G. Q.; Do, D. D. Carbon 1991, 29, 207. (7) Mochida, I.; Hirayama, T.; Kisamori, S.; Kawano, S.; Fujitsu, H. Langmuir 1992, 8, 2290. (8) Gangwal, S. K.; Howe, G. B; McMichael, W. J.; Spivey, J. J. A Novel Carbon-Based Process for Flue-Gas Cleanup. Final Report to U.S. Department of Energy, 1993. (9) Gangwal, S. K.; Howe, G. B.; Spivey, J. J.; Silveston, P. L.; Hudgins, R. R.; Metzinger, J. G. Environ. Prog. 1993, 12, 128. (10) Lu, G. Q.; Do, D. D. Gas Sep. Purif. 1994, 8, 17. (11) Kisamori, S.; Mochida, I.; Fujitsu, H. Langmuir 1994, 10, 1241. (12) Kisamori, S.; Kurado, K.; Kawano, S.; Mochida, I.; Matsumura, Y.; Yoshikawa, M. Energy Fuels 1994, 8, 1337. (13) Kim, J.-Y.; Hong, I.; Lee, J. G. In Proceedings of 22nd Biennial Conference on Carbon, San Diego, CA; 1995; p 534. (14) Fei, Y.; Sun, Y. N.; Givens, E.; Derbyshire, F. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1995, 40 (4), 1051. (15) Vladea, R. V.; Hinrichs, N.; Hudgins, R. R.; Suppiah, S.; Silveston, P. L. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1996, 41 (1), 432. ‡ §

S0887-0624(96)00197-1 CCC: $14.00

high-quality carbon adsorbent for SO2 removal should have a high adsorption capacity, rapid adsorption kinetics, low reactivity with oxygen, minimal loss of activity after regeneration, low-pressure drop, high mechanical strength, and low cost. The overall objective of this study17-23 has been to produce activated char from Illinois coal with optimal SO2 removal properties and to gain a better understanding of SO2 removal by carbon. The reaction of SO2 with carbon in the presence of O2 and H2O at relatively low temperatures (20-150 °C) involves a series of reactions that leads to the formation of sulfuric acid as the final product. The overall reaction is SO2 + 1/2O2 + H2O + C f C-H2SO4. Most studies that have sought to maximize the SO2 removal capabilities of a carbon only assume a certain mechanism for SO2 adsorption and conversion to H2SO4 is operable. In (16) Mochida, I.; Kuroda, K.; Yasutake, A.; Yoshikawa, M.; Matsumura, Y. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1996, 41 (1), 335. (17) Lizzio, A. A.; DeBarr, J. A.; Kruse, C. W.; Donnals, G. L.; Rood, M. J. Production and Use of Activated Char for Combined SO2/NOx Removal. Final Technical Reports to the Illinois Clean Coal Institute, 1994, 1995. (18) DeBarr, J. A.; Lizzio, A. A. In Proceedings of International Conference on Carbon, Granada, Spain; 1994; p 268. (19) Lizzio, A. A.; DeBarr, J. A. Fuel 1996, 75, 1515. (20) DeBarr, J. A. The Role of Free Sites in the Removal of SO2 from Simulated Flue Gases by Activated Char. M.S. Thesis, University of Illinois at Urbana-Champaign, 1995. (21) DeBarr, J. A.; Lizzio, A. A.; Rood, M. J. In Proceedings of 88th Annual Air and Waste Management Association Meeting, San Antonio, TX; 1995. (22) DeBarr, J. A.; Lizzio, A. A. In Proceedings of 22nd Biennial Conference on Carbon, San Diego, CA; 1995; p 562. (23) Lizzio, A. A.; DeBarr, J. A. In Proceedings of 12th International Annual Pittsburgh Coal Conference, Pittsburgh, PA; 1995; p 652.

© 1997 American Chemical Society

Mechanism of SO2 Removal

the literature,1-5,8,10,24 the following reaction sequence has usually been presented before any discussion on SO2 removal by carbon is begun.

C + SO2 f C-SO2 C + 1/2O2 f C-O C + H2O f C-H2O C-SO2 + C-O + C-H2O f C-H2SO4 It implies that SO2, O2, and H2O are all adsorbed on the surface of the carbon in close enough proximity and in the proper steric configuration to react and form H2SO4. A clearer understanding of this sequence of reactions may lead to the development of activated carbons better suited for adsorption of SO2 and its conversion to H2SO4. Experimental parameters that may affect SO2 removal by carbon include flue gas temperature, concentration of other pollutants in the flue gas, particle size of the sorbent, flow rate, amount of O2 and H2O in the flue gas, and the physical and chemical properties of the carbon such as surface area and functional groups. In this paper, we focus on the effect of surface area and chemisorbed oxygen on the SO2 adsorption capacities of chars prepared from a bituminous coal. Using temperature-programmed desorption (TPD), we titrate the carbon sites responsible for adsorption of SO2 and its conversion to H2SO4, and on the basis of the experimental results, we propose a more detailed mechanism for SO2 removal by carbon. Experimental Section Char Preparation. Activated chars were prepared from a sample of Illinois Colchester (No. 2) hvC bituminous coal (IBC-102) obtained from the Illinois Basin Coal Sample Program. A 5 cm ID batch, fluidized bed reactor was used to pyrolyze 200 g of 48 × 100 mesh coal (N2, 900 °C, 0.5 h) and activate the resultant char (H2O, 860 °C, 30% conversion). The steam-activated char was treated with nitric acid (10 M HNO3, 80 °C, 2 h) to modify its pore structure and surface chemistry. The HNO3-treated char was heated in nitrogen to various temperatures (200-925 °C) and held there for 1 h to desorb carbon-oxygen (C-O) complexes formed during the HNO3 treatment. A commercial activated carbon (Calgon F400) was also studied. A chemical activation method was employed to increase the surface area of IBC-102 char. Potassium hydroxide (KOH) was physically mixed with IBC-102 coal (48 × 100 mesh) to form a coal/KOH mixture. A blender was used to mix one part coal with two parts KOH. The mixture was air-dried overnight. Two hundred grams of dried coal/KOH mixture was added to the 5 cm FBR and pyrolyzed in N2 at 800 °C for 1 h. The char sample was cooled in N2 and washed repeatedly with distilled water to remove leftover potassium. Char samples (3 g) were removed periodically during the washing process to obtain samples having different surface areas and amounts of potassium. SO2 Adsorption Capacity. The SO2 adsorption capacities of prepared chars were determined by thermogravimetric analysis (Cahn TG-131). In a typical run, a 30-50 mg char sample was placed in a platinum pan and heated at 20 °C/ min in flowing N2 to 120 °C to remove moisture and impurities; once the temperature and weight stabilized, the flow of N2 was (24) Gail, E.; Kast, W. Chem. Eng. Sci. 1990, 45, 403.

Energy & Fuels, Vol. 11, No. 2, 1997 285 switched to a mixture of gases containing 5% O2, 7% H2O, and the balance N2. Once there was no further weight gain due to adsorption of O2 and H2O, SO2 was added in concentrations representative of a typical flue gas for combustion of highsulfur coal (2500 ppm of SO2). The weight gain versus time was recorded by a computerized data acquisition system. TPD. TPD experiments (N2, 25-1000 °C, 5 °C/min, 1 h at 1000 °C) were performed to determine the amount of oxygen adsorbed on the char surface. TPD experiments were carried out in a flow-through, 2.5 cm i.d. stainless steel fixed bed reactor system. In a typical run, 0.5 g of sample was heated at 5 °C/min in flowing nitrogen (0.5 L/min) to a final temperature of 1000 °C and held there for 1 h to desorb CO and CO2 from the char surface. Rosemount Model 880 CO and CO2 nondispersive infrared analyzers were used to continuously monitor the concentrations of CO and CO2 in the effluent gas. Surface Area. Surface areas were determined from the amount of N2 and CO2 adsorbed at 77 and 195 K, respectively, using a dynamic sorption method in conjunction with a singlepoint BET adsorption equation. Single-point N2 BET surface areas were determined from N2 (77 K) adsorption data obtained at a relative pressure (P/P0) of 0.30 using a Monosorb flow apparatus (Quantachrome Corp.). Single-point CO2 BET surface areas were determined from CO2 (195 K, dry iceethanol) adsorption data obtained in a custom-made U-tube apparatus at a P/P0 of 0.15 (220 Torr of CO2 in helium, P0 ) 1450 Torr). Further details of the experimental equipment and procedures used in this study are provided elsewhere.19-21

Results and Discussion Effect of Surface Area and Chemisorbed Oxygen. Table 1 lists the SO2 adsorption capacities and N2 BET and CO2 BET surface areas of two series of carbons prepared from IBC-102 coal and a commercial activated carbon (Calgon F400). The SO2 capacities normalized with respect to N2 BET surface area (SO2/ N2) are seen to vary by more than 2 orders of magnitude (5.8/0.05) for the IBC-102 char and by a factor of 5 (0.46/ 0.08) for the commercial carbon, thus indicating a poor correlation. It is interesting to note that the steamactivated IBC-102 char (H2O, 860 °C) and untreated Calgon F400 carbon have comparable SO2 adsorption capacities despite the rather large difference in their N2 BET surface areas (about 800 m2/g). Also, Table 1 shows that the SO2 adsorption capacities of these carbons seem to correlate better with their CO2 BET surface areas. Due to its smaller size (3.3 versus 3.46 Å) and higher adsorption temperature (195 versus 77 K), the CO2 molecule is able to diffuse more readily into the micropores than N2, thus achieving equilibrium more quickly and better representing the surface area of the carbon accessible to SO2 at 120 °C. Table 1 also shows an inverse correlation exists between the SO2 adsorption capaciy and amount of oxygen chemisorbed on both series of nitric acid treated, thermally desorbed carbons. The SO2 capacity normalized with respect to chemisorbed oxygen (SO2/O2) is seen to increase monotonically with increasing thermal desorption temperature (from 200 to 925 °C). A similar effect of heat treatment on the SO2 adsorption capacity of polyacrylonitrile-based activated carbon fibers was recently reported by Mochida and his research group7,11,12 and attributed to an increase in the number of active sites generated by the evolution of CO and CO2 during decomposition of C-O functional groups. Davini25,26 (25) Davini, P. Fuel 1989, 68, 145. (26) Davini, P. Carbon 1990, 28, 565.

286 Energy & Fuels, Vol. 11, No. 2, 1997

Lizzio and DeBarr

Table 1. Correlation of SO2 Adsorption Capacity with Surface Area and Chemisorbed Oxygen sample

SO2 capacitya (mg of SO2/g of char)

N2 BET (m2/g)

CO2 BET (m2/g)

IBC-102, 900 °C, 0.5 h IBC-102, 900 °C; H2O, 860 °C IBC-102, 900 °C; H2O, 860 °C; HNO3 IBC-102, 900 °C; H2O, 860 °C; HNO3, 525 °C IBC-102, 900 °C; H2O, 860 °C; HNO3, 725 °C IBC-102, 900 °C; H2O, 860 °C; HNO3, 925 °C Calgon F400 Calgon F400, HNO3 Calgon F400, HNO3, 200 °C Calgon F400, HNO3, 525 °C Calgon F400, HNO3, 725 °C Calgon F400, HNO3, 925 °C

7 176 24 91 241 287 206 -b 46 117 156 214

1.2 220 400 460 500 550 1000 600 456 533 463

98 613 585 693 727 726 1000 -

a

SO2/N2 (mg/m2)

SO2/CO2 (mg/m2)

5.8 0.80 0.06 0.20 0.48 0.05 0.21 0.08 0.26 0.29 0.46

0.07 0.29 0.04 0.13 0.33 0.39 0.21 -

O2 (wt %)

SO2/O2

0.5 1.1 16.4 5.9 1.6 0.5 0.5 15.7 14.3 5.6 3.4 1.7

1.40 16.0 0.15 1.54 15.0 57.4 41.2 0.32 2.09 4.59 12.6

SO2 capacity determined after 6 h. b Not determined.

Figure 1. TPD profiles of nitric acid/thermally desorbed IBC102 chars.

ascribed enhanced SO2 adsorption on oxidized activated carbon to the presence of basic C-O groups on the carbon surface. Most recently, Kim et al.13 and Fei et al.14 proposed that the inherent nitrogen in polyacronitrile and shale oil derived activated carbon fibers, respectively, increases their catalytic activity for SO2 adsorption and conversion to H2SO4. Concept of Free Sites. The TPD profiles of thermally desorbed, nitric acid treated IBC-102 chars shown in Figure 1 indicate that the absence of oxygen on the carbon surface results in a char with highest SO2 adsorption capacity (e.g., HNO3-925 °C char). Apparently, carbon atoms, which are not tied up by an adsorbed oxygen atom, have valence electrons that are more available and are more reactive toward SO2 adsorption. We, therefore, postulate that the unoccupied or free sites control the uptake of SO2. It is wellknown that once carbon is evolved as CO or CO2 during carbon gasification, a new carbon atom of equal or greater reactivity will be exposed. These newly exposed carbon atoms, coined nascent sites by Phillips et al.,27 in many cases define the surface chemistry of a carbon.28,29 It is these nascent sites, made active by the (27) Phillips, R.; Vastola, F. J.; Walker, P. L., Jr. Carbon 1970, 8, 197. (28) Boehm, H. P.; Bewer, G. In Proceedings of 4th London International Conference on Carbon and Graphite; 1974; p 344.

thermal desorption treatment at 925 °C, that we believe are responsible for enhanced SO2 adsorption on the HNO3-925 °C char. Nitric acid, being a very strong oxidant, oxidizes the surface of the carbon so that even carbon atoms that would otherwise not react with the oxygen in air are oxidized. Upon thermal desorption treatment, a once relatively unreactive carbon atom is released as CO or CO2 leaving behind a carbon atom which is then primed to react with SO2. The evolution of CO or CO2 may also serve to activate the pore structure, thereby widening the pores and making them more accessible to reactant gases; this could also increase SO2 adsorption. A parallel study with activated carbon fibers (ACF) in our laboratory,30 however, has shown that ACF with smaller pores (9 Å) and lower surface areas (600 m2/g) actually adsorb more SO2 than higher surface area ACF (1900 m2/g) having larger pores (18 Å). This increase in SO2 adsorption for low surface area fibers is presumably due to enhanced adsorbentadsorbate interactions caused by the smaller average pore size.31 Measurement of Free Sites. To test our hypothesis that free sites are primarily responsible for SO2 adsorption, the TPD profiles of the nitric acid treated, thermally desorbed IBC-102 chars of Figure 1 were utilized. We defined the free sites for this series of chars as being those sites that were once occupied by oxygen because of nitric acid treatment, but now, because of the thermal desorption treatment, have become unoccupied or free. To quantify the CO free sites for a given char, say, the HNO3-725 °C char, we subtracted its CO evolution profile from the one of the original nitric acid treated char (Figure 2). The unshaded region in Figure 2 represents the number of free sites that are created by CO evolution during the thermal desorption treatment at 725 °C. Table 2 lists the number of both the CO and CO2 free sites calculated for each of the nitric acid treated thermally desorbed IBC-102 chars. The last two columns show the SO2 adsorption capacities normalized with respect to the CO and CO2 free sites. The SO2 adsorption capacity normalized with respect to CO and CO2 free sites is seen to vary by factors of 1.1 (1.15/ 1.04) and 2.7 (1.49/0.55), respectively. This is an excellent correlation compared to those found for surface area or adsorbed oxygen (Table 1). (29) Leon y Leon, C. A.; Radovic, L. R. Chem. Phys. Carbon 1994, 24, 213. (30) DeBarr, J. A.; Lizzio, A. A.; Daley, M. A. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1996, 41 (1), 339. (31) Foster, K. L.; Fuerman, R. G.; Economy, J.; Larson, S. M.; Rood, M. J. Chem. Mater. 1992, 4, 1068.

Mechanism of SO2 Removal

Energy & Fuels, Vol. 11, No. 2, 1997 287

adsorption sites was also applied to this carbon. Again, a good correlation was found between SO2 capacity and number of free sites (Table 2). The SO2 capacity normalized with respect to the number of CO and CO2 free sites varies by factors of only 1.7 (1.43/0.84) and 2.9 (2.12/0.73), respectively. It is interesting to note that there is nearly a one to one ratio of SO2 capacity to CO free sites for both the IBC-102 and Calgon F400 carbons. Mechanism of SO2 Removal by Carbon. On the basis of these experimental results, we propose the following sequence of reactions to explain the mechanism of SO2 removal by carbon.

O2 chem

O2 + 2C f 2C-O

SO2 ads

SO2 + C f C-SO2

SO2 oxid

C-SO2 + O2 + C f C-SO3 + C-O

H2O ads

C + H2O f C-H2O

H2SO4 form. C-SO3 + C-H2O f C-H2SO4 + C Figure 2. Subtraction of CO evolution profiles to determine CO free sites.

To the best of our knowledge, this type of quantitative approach has never before been used to explain SO2 removal by carbon. The concept of free or unoccupied sites was first postulated and utilized by Laine et al.32 to explain the gasification reactivity of carbon in oxygen. Our new approach shows that it may be possible to titrate directly the free sites responsible for SO2 adsorption and conversion to H2SO4. A slight variation in this same approach could be used to explain results in other studies that have examined SO2 removal by carbon. For example, Davini25,26 found that the SO2 adsorption capacity of a commercial grade carbon increased with an increase in activation temperature. The activation conditions he used ranged from 300 °C in air to 800 °C in 2% O2. Char gasification reactivity studies have shown that less oxygen will adsorb on the carbon surface at higher temperatures and in lower pressures of an oxidizing gas.28,33,34 The higher temperature used by Davini during activation is likely to have deposited less oxygen on the carbon surface, thus preserving more free sites for reaction with SO2. In fact, further analysis of his SO2 and O2 adsorption data26 using our concept of free sites shows that the SO2 adsorption capacities of two groups of carbons, which each varied by more than a factor of 3, could be made to vary by less than a factor of 1.2 when normalized to the total number of free sites. Table 2 shows that there was little improvement in the SO2 capacity of the commercial activated carbon (also a steam-activated bituminous coal char) with the nitric acid, thermal desorption treatment, perhaps because its pore structure was optimized for adsorption of contaminants in liquid phase applications. Any further increase in pore size resulting from the oxidation/thermal desorption treatment may have offset any increase in free sites. Our procedure for assessing free (32) Laine, N. R.; Vastola, F. J.; Walker, P. L., Jr. J. Phys. Chem. 1963, 67, 2030. (33) Lizzio, A. A.; Jiang, H.; Radovic, L. R. Carbon 1990, 28, 7. (34) Radovic, L. R.; Jiang, H.; Lizzio, A. A. Energy Fuels 1991, 5, 68.

regeneration (water wash) C-H2SO4 f H2SO4 + C regeneration (thermal) C-H2SO4 f SO2 + H2O + CO or C-O Adsorption of SO2 and O2 occurs in parallel on the free carbon active sites (denoted by C). Molecular oxygen dissociates on two free sites to form a pair of C-O complexes. SO2 essentially competes with O2 for free sites. The reaction of oxygen with carbon produces a stable C-O complex, which we have found inhibits SO2 adsorption. The next step is the oxidation of SO2 to SO3. Conceivably, three reactions are possible. The C-SO2 complex can react directly with molecular oxygen in the vicinity of a free site to form a C-SO3 intermediate and stable C-O complex. SO2 oxidation could also occur by the reaction of either the C-SO2 complex or SO2 with the C-O complex, both of which form C-SO3 and one free site:

C-SO2 + C-O f C-SO3 + C SO2 + C-O f C-SO3 + C However, the occurrence of these two reactions is less likely since we observed an inverse correlation between SO2 adsorption and stable C-O complex. The fact that O2 and SO2 compete for surface sites does not preclude the possibility that only a small percentage of C-O surface sites is responsible for oxidizing the C-SO2 complex. The conversion of a stable C-O complex into a reactive C(O) intermediate, as proposed in recent carbon gasification studies,33-36 could also play a role in SO2 adsorption and conversion to H2SO4.19 According to the generic SO2 removal mechanism most often presented in the literature (see Introduction), one could assume that the more oxygen adsorbed on the carbon surface, the more SO2 will be (35) Back, M. H. Carbon 1991, 29, 1290. (36) Radovic, L. R.; Karra, M.; Skokova, K. In Proceedings of 22nd Biennial Conference on Carbon, San Diego, CA; 1995; p 636.

288 Energy & Fuels, Vol. 11, No. 2, 1997

Lizzio and DeBarr

Table 2. Correlation of SO2 Adsorption Capacity with Free Adsorption Sites sample

SO2 capacitya

adsorbed oxygena

CO/CO2

CO free sitesa,b

CO2 free sitesa,b

SO2 capacity/ CO free sites

SO2 capacity/ CO2 free sites

IBC-102, HNO3 IBC-102, HNO3, 525 °C IBC-102, HNO3, 725 °C IBC-102, HNO3, 925 °C Calgon F400, HNO3 Calgon F400, HNO3, 200 °C Calgon F400, HNO3, 525 °C Calgon F400, HNO3, 725 °C Calgon F400, HNO3, 925 °C

0.037 0.142 0.376 0.448 -c 0.072 0.183 0.244 0.334

5.12 1.84 0.50 0.16 4.91 4.47 1.75 1.06 0.53

1.4 6.7 12.8 -d 1.4 0.8 7.6 5.5 5.4

0 1.37 3.35 3.90 0 0.86 1.28 2.26 3.28

0 2.59 2.94 3.01 0 0.34 2.52 2.56 2.75

-c 1.04 1.12 1.15 -c 0.84 1.43 1.08 1.02

-c 0.55 1.28 1.49 -c 2.12 0.73 0.95 1.22

a Moles per kiloggram of char. b Calculated assuming that 1 chemisorbed O evolved as 2 CO during TPD is equivalent to 2 CO free 2 sites, and 1 chemisorbed O2 evolved as 1 CO2 is equivalent to 1 CO2 free site. c Not determined. d CO2 concentration below detectable limits.

adsorption is enhanced for both chars when oxygen is present in the flue gas (compare 0 and 5% O2 plots in Figure 3). The adsorption of SO2 on a free site is seen to account for nearly 50% of the weight gain for both chars. These data also seem to support the SO2 oxidation mechanism, i.e., the one involving the reaction of adsorbed SO2 with O2 and a free site to form C-SO3. Hartman and Coughlin38 found the kinetic rate constant for the catalytic oxidation of SO2 to SO3 to be an important factor in the removal of SO2 by carbon. If one assumes that the rate-determining step in the overall reaction is, indeed, catalytic oxidation of SO2 to SO3, and assuming that the first SO2 oxidation reaction is operable, the overall rate of reaction could be expressed as

rate ) k2[C][C-SO2][O2]

Figure 3. Effect of oxygen on SO2 adsorption.

adsorbed. In this study, however, we have seen that the formation of stable C-O complex during char preparation serves only to occupy otherwise reactive free adsorption sites, suggesting that this type of C-O complex is not an essential reaction intermediate in the conversion of SO2 to H2SO4. It may also be that our TPD method did not titrate those small amounts of C-O complexes responsible for catalytic oxidation of SO2 to SO3 and that a fleeting C(O) complex, formed contemporaneously with SO2 adsorption and/or by transformation of stable complex into reactive intermediate, acts as the “catalyst”. Isotope labeling studies could be useful in determining if such an intermediate exists.37 Another question is whether SO2 removal would be possible if there was no oxygen in the flue gas and/or adsorbed on the carbon surface. This was resolved by using two of the IBC-102 chars, one with some adsorbed oxygen (HNO3-525 °C) and one with essentially no adsorbed oxygen (HNO3-925 °C). Figure 3 shows that, with 5% O2 in the flue gas, the HNO3-925 °C char adsorbs the most SO2. With no oxygen in the flue gas, this char still adsorbs SO2, but in this case, it is not converting C-SO2 to H2SO4 due to a lack of oxygen. SO2 (37) Kapteijn, F.; Meijer, R.; Moulijn, J. A. Energy Fuels 1992, 6, 494. (38) Hartman, M.; Coughlin, R. W. Chem. Eng. Sci. 1972, 27, 867.

where k2 is a fundamental rate constant. The unknowns in this expression are k2 and the concentration of C-SO2 complex. At steady state, the rates of each of the reactions in the mechanism presented above are equal to that of the rate-determining step, and by definition, the concentration of each of the intermediate species does not change with time. Thus

d[C-SO2]/dt ) 0 ) k1[C][SO2] - k2[C][C-SO2][O2] Solving for [C-SO2]

[C-SO2] )

k1[SO2]

k1[C][SO2] k2[C][O2]

)

k2[O2]

and substituting this into the rate expression

rate ) k1[C][SO2]n where n is the order of reaction with respect to SO2. The rate is then only a function of the concentration of free sites and the partial pressure of SO2, a departure from the more elaborate rate expressions recently proposed in the literature (see, for example, refs 2 and 24). Another possibility is that SO2 chemisorbs dissociatively on two free sites. The C-SO complex could then react with molecular oxygen to form C-SO3.

SO2 + 2C f C-SO + C-O C-SO + O2 f C-SO3 The rate expression derived assuming dissociative

Mechanism of SO2 Removal

Figure 4. Effect of SO2 concentration.

chemisorption of SO2, however, reduces to a secondorder dependence on free sites, which was not observed experimentally (Table 2).

rate ) k1[C]2[SO2]n Since SO2 capacity was directly proportional to the number of free sites as determined by TPD, we conclude that nondissociative chemisorption of SO2 is most likely occurring under these conditions. For concentrations of SO2 1500 ppm, the value of n approached 0 (Figure 4). Nevertheless, at a constant partial pressure of SO2, the rate of SO2 adsorption and conversion to H2SO4 should be directly proportional to the number of free adsorption sites as confirmed by our TPD experiments. It remains to be determined whether the concept of free sites can be used to obtain a better understanding of adsorption of other contaminants in flue gas. If so, it could facilitate preparation of activated char optimized for removal of SO2 as well as other air toxics (e.g., nitrogen oxides, mercury) from coal combustion flue gas. Recent studies suggest that adsorption of NOx and reduction to N2 is more or less controlled by free sites on the carbon surface.39-41 In the above mechanism, the C-SO3 intermediate reacts with adsorbed water to form sulfuric acid and the free site is restored. Another free site is restored to its original state after regeneration. Methods of char regeneration commonly used include thermal desorption to produce a concentrated stream of SO226,42 or washing (39) Teng, H.; Suuberg, E. M.; Calo, J. M. Energy Fuels 1992, 6, 398. (40) Yamashita, H.; Tomita, A.; Yamada, H.; Kyotani, T.; Radovic, L. R. Energy Fuels 1993, 7, 85. (41) Mochida, I.; Kisamori, S.; Hironaka, M.; Kawano, S.; Matsumura, Y.; Yoshikawa, M. Energy Fuels 1994, 8, 1341. (42) Cha, C. Y.; Vaillancourt, M. B.; Kim, S. S. Prepr.sAm. Chem. Soc., Div. Pet. Chem. 1992, 1453.

Energy & Fuels, Vol. 11, No. 2, 1997 289

with H2O or dilute H2SO4 to produce medium-strength H2SO4.8,9,12,14 The water adsorbed in the pores may also act as an inherent regeneration medium whereby the sulfuric acid formed on the carbon surface is continuously dissolved into a “reservoir” of water adsorbed in the pores. As the acid goes into solution, the free site once again becomes available to react with SO2 (or O2). Thus, the site may undergo numerous cycles of adsorption/desorption without an external means of regeneration. The production of H2SO4 proceeds indefinitely until water adsorbed in the pores becomes saturated with H2SO4 and/or the free sites become occupied with oxygen, at which point the spent carbon needs to be regenerated, e.g., thermal treatment or flushing with water or dilute acid. (The latter treatment will not remove any of the C-O complex). Thus, it can be expected that the amount of water retained in the pore volume of the carbon should determine its equilibrium SO2 adsorption capacity, which may require more than 40 h to attain (Figure 4), whereas the rate (or kinetics) of SO2 adsorption, say, in the first 6 h of adsorption, will be controlled by the number of free sites. In an earlier paper, Jungten and Kuhl3 hypothesized that the active sites would control the rate of SO2 adsorption, but this was never verified experimentally. The observed increase in the SO2 adsorption capacity of nitric acid treated chars with increasing thermal desorption temperature is probably due to an increase both in the concentration of free active sites and in accessible pore volume of the char. The latter leads to a greater reservoir for storage of dissolved H2SO4 and an increased ability to regenerate the active sites for additional SO2 adsorption. Char Optimization. Chemical activation of IBC102 coal with potassium hydroxide was performed in an attempt to increase the SO2 adsorption capacity of the char. The objective was twofold: (1) to increase surface area above and beyond that achieved by steam activation and (2) to deposit a base metal in the pore structure that could readily attract an acid gas such as sulfur dioxide. Table 3 shows that the first objective was met. Figure 5 presents the SO2 adsorption profiles for the KOH-activated chars after the first, third, and final washes. The SO2 capacity of the first wash char sample after 6 h is about 35% less than that of the final wash sample. This may be due to some pore blocking by potassium compounds not removed by the first wash. By the third wash most of these compounds apparently have been rinsed from the char. It is interesting to note that the KOH-activated char became noticably warm when first removed from the 5 cm FBR after activation. The exothermic reaction can be attributed to rapid rehydration of the KOH and/or large amounts of oxygen chemisorbing on the char surface. To further examine the oxygen functional groups formed on the char surface, TPD profiles were obtained for this char (Figure 6). There are two distinct CO2 desorption peaks compared to the one peak observed for most chars (see Figure 2 or ref 19). Integration of the CO and CO2 desorption profiles shows that 9 wt % oxygen was deposited on the char surface. It was of interest to remove this oxygen using the thermal desorption treatment we used to increase the SO2 capacity of the nitric acid treated char. Figure 5 shows that the SO2 adsorption profile for the KOH-activated

290 Energy & Fuels, Vol. 11, No. 2, 1997

Lizzio and DeBarr

Table 3. N2 BET Surface Areas and SO2 Adsorption Capacities of Selected IBC-102 Chars and Commercial Activated Carbons

a

sample

N2 BET (m2/g)

SO2 capacity, 6 h (mg of SO2/g of char)

SO2 capacity, 15 h (mg of SO2/g of char)

Calgon F400 (D) Calgon Centaur (B) IBC-102, H2O, 860 °C (F) IBC-102, H2O, 860 °C (E) F, HNO3, 80 °C E, HNO3, 80 °C F, HNO3, 80 °C, N2, 925 °C (C) E, HNO3, 80 °C, N2, 925 °C (A) IBC-102, KOH, 800 °C (first wash) IBC-102, KOH, 800 °C (third wash) IBC-102, KOH, 800 °C (final wash) IBC-102, KOH, 800 °C, N2, 925 °C (final wash)

1000 360 220 545 400 395 463 560 765 1087 1133 -

206 327 176 190 -b 275 250 105 121 185 186

330a 350 220a 290 330 420 150 172 275 310

Extrapolated value. b Not determined.

Figure 6. TPD profile of the KOH-activated char (final wash).

Figure 5. Kinetics of SO2 adsorption on selected KOHactivated chars.

char (final wash) thermally desorbed at 925 °C for 1 h is similar to that of the original char containing 9% oxygen. According to the concept of free sites discussed in this paper, the removal of oxygen from the char surface should have enhanced the SO2 adsorption capacity of the char. One possible explanation for the lack of increase in the SO2 capacity of this char with thermal desorption is that the original C-O-K complexes in the char were transformed into relatively unreactive C-K complexes upon heat treatment at 925 °C in inert gas. The C-O complexes that are removed as CO and CO2 during heat treatment to 925 °C leave behind highly reactive nascent sites as previously discussed, but in this case, the mobile potassium atoms, which are still present in the char, find and react with these nascent sites forming stable C-K complexes that are seemingly unreactive toward SO2. The potassium concentration in the KOH-activated carbon (final wash) was determined to be 0.2 mol %, which may have been

enough to deactivate a good portion, if not not all, of the free sites created by the oxidation/thermal desorption treatment. Typical free site concentrations measured in this study were 0.01-0.05 mol % (Table 2). In another attempt to increase the SO2 adsorption capacity of IBC-102 char, the steam-activated IBC-102 char was activated to a higher level of conversion (55% instead of 30%), thereby increasing its surface area from 220 to 545 m2/g (Table 3). This char was then subjected to the nitric acid/thermal desorption treatment. Figure 7 presents SO2 adsorption profiles for the steamactivated chars (E and F), the two nitric acid treated, thermally desorbed, steam-activated chars (A and C), and two commercial activated carbons (B, Centaur; and D, F400) manufactured by the Calgon Carbon Corp.. Table 3 lists the N2 BET surface areas and SO2 adsorption capacities (after 6 and 15 h) for these carbons. A comparison of the SO2 adsorption profiles for the two steam-activated chars (E and F) indicates that the kinetics of adsorption are faster for the lower surface area char (F) but that the SO2 capacity of the higher surface area char (E) is greater than that of the lower surface area char after about 4 h. Although pore size measurements were not made on these two chars, the average pore size of the 220 m2/g char is presumed to be smaller than that of the 545 m2/g char since it is likely that the micropores were enlarged in going from 30% to 55% conversion. Thus, smaller pores seem to result in faster kinetics, while larger surface areas or

Mechanism of SO2 Removal

Energy & Fuels, Vol. 11, No. 2, 1997 291

be improved further by increasing the number of free sites and/or total pore volume, e.g., by performing a nitric acid thermal desorption treatment on the high surface area KOH-activated char. Conclusions

Figure 7. Kinetics of SO2 adsorption on selected IBC-102 chars and commercial activated carbons.

total pore volumes lead to greater equilibrium adsorption capacities. This same trend was recently observed for activated carbon fibers varying in surface area from 500 to 1900 m2/g.43 Figure 5 shows that the HNO3 thermal desorption treatment increased the SO2 capacity of both steam-activated chars. The SO2 capacity of the one derived from the higher surface area char (A) actually surpasses that of the Centaur carbon (B) after 10 h. The Centaur carbon is a relatively expensive carbon ($2.50/lb) recently developed by Calgon Carbon for both liquid and vapor phase catalytic reactions, including SO2/NOx removal. It remains to be determined whether the SO2 capacity of IBC-102 char can (43) DeBarr, J. A.; Lizzio, A. A.; Daley, M. A. Energy Fuels 1997, 11, 267-271.

In this study, we found the SO2 adsorption capacity of a coal char to be inversely proportional to the amount of oxygen adsorbed on its surface. TPD was used to titrate those sites responsible for adsorption of SO2 and conversion to H2SO4. On the basis of these results, a more detailed mechanism for SO2 removal by carbon has been proposed. The derived rate expression shows SO2 adsorption to be dependent only on a fundamental rate constant and concentration of carbon atoms we designate as free sites. The results obtained here are analogous to those of a recent study33 which found that a similar relationship exists between the specific rate (Rsp) of carbon gasification in carbon dioxide and the number of reactive sites measured initially by transient kinetics (TK) experiments and subsequently confirmed by TPD, i.e., Rsp ) k[C]. In that study, TK and TPD were used to titrate those occupied sites or C(O) intermediates responsible for controlling the ratedetermining desorption step in CO2 gasification of carbon. The results of both studies seem to support a unified approach to the reactions between carbon and oxygen-containing gases, as most recently proposed by Chen et al.44 and Moulijn and Kapteijn.45 On the basis of this concept of free sites, we were able to optimize the SO2 adsorption capacity of activated char prepared from Illinois coal. Acknowledgment. This work was supported by the Illinois Clean Coal Institute through the Illinois Coal Development Board and the U.S. Department of Energy. We gratefully acknowledge the technical assistance of Gwen Donnals. We appreciate discussions with Carl Kruse, Scott Chen, and John Lytle. EF960197+ (44) Chen, S. G.; Yang, R. T.; Kapteijn, F.; Moulijn, J. A. Ind. Eng. Chem. Res. 1993, 32, 2835. (45) Moulijn, J. A.; Kapteijn, F. Carbon 1995, 33, 1155.