Studies on the Surface Chemistry Based on Competitive Adsorption of

Environ. Sci. Technol. 2002, 36, 4928-4935

Studies on the Surface Chemistry Based on Competitive Adsorption of NOx-SO2 onto a KOH Impregnated Activated Carbon in Excess O2 YOUNG-WHAN LEE* AND JIN-WON PARK Department of Chemical Engineering, Yonsei University, Sinchon-dong, Seodaemun-gu, Seoul 120-749, Korea JEONG-HO YUN Aspen Engineering Suite, AspenTech Ltd., Sheraton House, Castle Park, Cambridge CB3 0AX, United Kingdom JUNG-HYE LEE AND DAE-KI CHOI Korea Institute of Science & Technology, P.O. Box 131, Cheongryang, Seoul 130-650, Korea

This study identifies surface chemistry characteristics based on competitive behavior in the simultaneous adsorption behavior of NOx (NO rich) and SO2 using KOH impregnated activated carbon (K-IAC) in excess O2. The NOx and SO2 adsorption on K-IAC occurred mainly through the acidbase reaction. The high surface area with many pores of activated carbon acted as storage places of oxide crystal produced from NOx and SO2 adsorption. KOH, an impregnant, provided the selective adsorption sites to NOx and SO2, enabling simultaneous adsorption. However, larger amounts of SO2, with higher adsorption affinity to K-IAC compared to NOx, were adsorbed in a NOx/SO2 coexistent atmosphere. Oxygen was chemisorbed to K-IAC, which enhanced the selective adsorptivity for NO. In binarycomponent adsorption of NOx and SO2 on K-IAC, oxide crystals such as KNOx (x ) 2,3) and K2SOx (x ) 3,4) were dominantly formed through two different adsorption mechanisms by chemical reaction. Depending on the extent that oxide crystals blocked pores, compositions of oxide crystals were distributed differently according to depth.

Introduction Nitrogen oxides (NOx) and sulfur dioxide (SO2) are classified, internationally, as substances for strict regulation as they are pollutants lethal to nature and mankind. For some time there has been much research on the NOx and SO2 removal efficiency of numerous materials. Examples are activated carbon, activated carbon fiber, alumina, metal catalyst, etc. (1-6). Commonly, combustion flue gas is temperature of 373-423 K with about 100-1000 ppm of NOx and SO2, each, and significantly O2-rich. NOx has a 7:1-10:1 ratio of NO to NO2. An advantages of controlling combustion flue gas with carbonaceous materials is that not only are NOx and SO2 removed but also other noxious substances, such as hydrogen fluoride (HF), hydrogen chloride (HCl), mercury (Hg) (7, 8). And, because industrial water is not used, secondary pollution is not the outcome and useful byproducts can be obtained * Corresponding author phone: (822)958-5885; fax: (822)9585809; e-mail: [email protected]. 4928

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 22, 2002

during this process. To date reports relate that noxious substance removal using general activated carbon (GAC) based on the physical adsorption, results in low selective adsorptivity with the exception of volatile organic compounds. Therefore this method is not implemented despite its many advantages; high surface area, unique pore structure, economy, etc. (9, 10). We have proposed a mechanism in our previous studies (11-14) using K-IAC to NOx and SO2 by conducting a research on surface chemistry along with the adsorption characteristics. ToF-SIMS and AES/SAM were used as surface analysis methods for the first time in our previous study to define surface chemistry with NOx and SO2 adsorbed onto impregnated activated carbon, and we were able to collect a great deal of information. To date, related studies (15-18) until recently have reported that in pure-component adsorption conducted for NOx and SO2 respectively using K-IAC, both showed higher selective adsorptivity, up to multiples of tens, than that of GAC. However, there has been scant research on simultaneous adsorption of the binarycomponent with NOx-SO2 using K-IAC, and no study at all on related surface chemistry (5). In recent years, studies on the surface group of the activated carbon have been conducted with great interest, but surface chemistry has been very difficult to define. In particular, surface pores are of irregular size and roughness making the analysis all the more difficult. To understand the use of impregnated activated carbon appropriately, it is most important for surface chemical characteristics as well as the selective adsorption behavior of adsorbate to be accurately identified (18). This study, therefore, has been aimed at identifying the simultaneous adsorption characteristics of NOx and SO2 on K-IAC in an excess O2 and defining the surface chemistry under the condition.

Experimental Section Adsorbent Preparation. The adsorbent was prepared by impregnating a KOH (Junsei Chemical Co.) solution into GAC obtained from coconut shell (Dongyang Carbon Co.). GAC was sieved through a 8/16 mesh and treated with N2 flowing for 4 h at 413 K. Next the treated GAC was dried at 383 K. KOH was impregnated in an aqueous solution state in the GAC via incipient wet impregnation and was used following dehydration at 403 K. Manufactured K-IAC was stored in a desiccator from common airborne moisture and contaminants to prevent adsorbent function reduction. At this time, cautions were taken for impregnation, drying, and storage of adsorbent because the processes have a great influence on the adsorption capacity. Through atomic absorption spectroscopy (AAS) analysis, potassium loading of K-IAC was confirmed. Table 1 shows detailed properties of K-IAC used. Fix-Bed Adsorption Experiments. The fixed-bed adsorption system used in this experiment is shown in Figure 1. Inside the columns, steel mesh was placed in the upper and lower extremities of the adsorbent to support the samples and minimize channeling phenomenon. The temperature of the column was maintained with an electric furnace located at its outer wall. For the system line, the temperature was maintained by using a heat band and heat insulating material and was regulated with a proportional-integral-differential (PID) temperature controller. The temperature was measured by connecting a K-type thermocouple (Omega Engineering Inc.) located inside the line and connected to a recorder. Each certified 2% NO/N2, 5000 ppm NO2/N2, and 2% SO2/N2, high purity O2 was diluted to the desired 10.1021/es011510b CCC: $22.00

 2002 American Chemical Society Published on Web 10/17/2002

TABLE 1. Properties of K-IAC Used K content (wt %) bulk density (kg m-3) BET surface area (m2 g-1) mesopore area (m2 g-1) micropore volume (cm3 kg-1) average pore diameter (Å)

9.96 570 719 15 232 11.5

FIGURE 2. NOx and SO2 concentration profiles for competitive adsorption of NOx-SO2 binary-component. Gas composition: 891 ppm NO, 116 ppm NO2, 1004 ppm SO2, 18.1% O2. Temperature ) 403 K. Linear velocity ) 29.971 cm/sec.

FIGURE 1. Experimental setup.

TABLE 2. Range of the Experimental Conditions K-IAC (g) bed height (cm) column diameter (cm) reaction time (h) contact time (sec) total flowrate (l/min) linear velocity (cm/sec) temperature (K) concentration NO (ppm) NO2 (ppm) SO2 (ppm) O2 (%)

10.673 20 1.09 5 0.667 1.678 29.971 403 891 116 1004 18.1

concentration range via a mass flow controller (Brooks Co., model 5280E). In the fore-end of the adsorption column, an inline static mixer was installed to facilitate mixing. Concentration of NO, NO2, and SO2 that exhausted from the bypass line and adsorption column were analyzed by using a chemiluminescent NOx analyzer (Thermo Environmental Instruments Inc., model 42C) and a flue gas analyzer (Eurotron Instruments S.p.A., GreenLine Mk II) for SO2, CO, and O2 analysis. Detailed conditions of each experiment are shown in Table 2. Surface Characterization. The surface chemical analysis was performed for samples by using SEM (Hitachi, s-4100), XRD (Rigaku, D/MAX-III A), AES/SAM (Perkin-Elmer, PHI Model 670), and static type ToF-SIMS (Perkin-Elmer, PHI Model 700 ToF-SIMS/SALI). The X-Ray Diffractometer (XRD) was used to investigate the inorganic components of the K-IAC. X-ray patterns were recorded in the scan range 2θ ) 5-90°, at a scan rate of 0.1 o per minute. The Auger Electron Spectroscopy (AES) survey scan and sputter depth profile were recorded using the following analytical conditions: primary beam energy Ep ) 10 keV, primary beam current Ip ) 0.0099 µA, and beam diameter ∼0.4 µm. The resolution of the cylindrical mirror analyzer was set to 0.6%. The argon ion beam, with an ion energy of 1.5 keV and a current density of 0.6 µAm2, was produced by a differentially pumped ion gun. The sputter profiles were analyzed using the software package PC PHI-MATLAB. In this analysis, AES was used

quantitatively to help determine the chemical state of potassium (K), oxygen (O), carbon (C), sulfur (S), and nitrogen (N) on the K-IAC. Scanning Auger-electron Microscopy (SAM) and Scanning Electron Microscopy (SEM) was applied for the morphology analysis of the K-IAC. Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) analysis was carried out using a system equipped with a two-stage reflectron-type analyzer. A low dose and pulsed Cs+ primary ion beam, with an impact energy of 10 keV, was employed. The spectrometer was run at an operating pressure of 10-9 mbar. The primary ion beam was directed on a square area of 50 µm × 50 µm. The system was operated in high sensitivity mode with a pulse width of 50 ns and with a beam current of 0.5 nA, resulting in a primary ion dose of approximately 4 × 1011 ions cm-2 analysis-1. SIMS spectra were acquired over a mass range of m/z ) 1-100 in negative modes.

Results and Discussion Binary-Component Adsorption of NOx (NO Rich) and SO2 on K-IAC. Our previous studies (11-14) reported the following results common to the adsorption of the individual pure-components, NOx and SO2, on K-IAC: a. Basic characteristics were provided to the nonpolar surface of activated carbon using KOH, enhancing the adsorption efficiency for NOx and SO2. b. OH-, in KOH, acted as selective adsorption sites reacting NOx and SO2. K+ helped NOx and SO2 to be stably adsorbed to K-IAC. c. After pure-component adsorption of NOx and SO2, respectively, on K-IAC, the adsorbates had strong chemical interaction with KOH according to surface characteristics, producing oxide crystal on the surface and thereby changing the chemical structure of K-IAC surface. In this experiment, binary-component adsorption of NOx-SO2, based on the preceding pure-component adsorption results, was attempted. Figure 2 shows the concentration curve of NOx and SO2 emitted from the column outlet after adsorption in the conditions of Table 2. NOx breakthrough started at the onset of reaction time and because NO is produced at NO2 adsorption. It was confirmed in our previous papers (11, 12) that the molar ratio of NO2 adsorption versus NO production in low column inlet concentration, approximately 150 ppm, VOL. 36, NO. 22, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4929

was 3:1. This result coincided with data on NOx adsorption on γ-alumina presented by Lee et al. (23). But as concentration increased, the molar ratio of NO produced, per total NOx, became relatively low compared to NO2. The increase of NO production was in proportion to adsorptivity. However, since the amount of NO in reaction gases is much higher than that of NO2, NO breakthrough progress was due, not to NO2 adsorption, but to the fact that adsorption is impossible when only NO is present in K-IAC. NO can be adsorbed only when it is converted into NO2 in the presence of oxygen (24), so rapid NO breakthrough at onset was caused by failure to react with KOH in K-IAC. As reaction time continues, however, especially from 62 min, oxygen atoms are chemisorbed on the surface, making very slow NO breakthrough. NO was converted into NO2 due to surface oxygen, adsorbed onto the surface, and then adsorbed on other selective adsorption sites. As the relatively well-adsorbed SO2 occupied significant amounts of selective adsorption sites, NO produced increased, at 116 min, after a minimal decrease at 62 min. At 160 min NO2 increased to column outlet concentration (Cf) of 493 ppm and then decreased gradually. Column inlet concentration (Ci) of NO2 was 116 ppm. It was thus clear that NO which was produced in higher concentration, after conversion to NO2, due to surface oxygen, was the amount that could not be adsorbed to selective adsorption site, and emitted from the column. NOx reached saturation point (Cf/Ci ) 1) after 100 min with SO2 at Cf/Ci ) 0.4612, 463 ppm. O2 significantly enhance NOx adsorptivity but does not particularly influence SO2 (25,26). In a Wilde and Marin study (27), using Na-γ-alumina adsorbent, it was reported that no adsorptivity impact on SO2, due to O2, existed during a NOx-SO2 removal experiment. The experiment result shows that SO2 is dominant in adsorption affinity compared to NOx. SO2 initiated its breakthrough at 68 min. At the end of the 5 h reaction, concentrations were 807 ppm SO2, 778 ppm NO, and 390 ppm NO2. SO2 breakthrough increased slowly, in line with reaction time, its value reaching less than Cf/Ci ) 1, but, as the reaction time increased, NOx increased dramatically with Cf larger than Ci. Such excess NOx concentration is due to interaction between the adsorbates and the difference in adsorption affinity between the binarycomponent adsorbates on K-IAC. SO2 with higher adsorption affinity is adsorbed at the onset of the adsorption process. NOx is then adsorbed in the next available selective adsorption site along the K-IAC bed, but, as reaction time passes, SO2 is eventually distributed throughout the entire K-IAC bed. The unadsorbed SO2 moves to the next area of the K-IAC bed, desorbing the already adsorbed NOx and adsorbing in its place. Figure 3 shows the concentration of O2 emitted from the column outlet under the experiment conditions of Figure 2. Several studies report that O2 is chemisorbed on the activated carbon surface, as in eq 1, on the atomic state.

O2 f 2Oads

(1)

The column inlet concentration of O2 was 18.1% and rapidly decreased at onset as it passed the adsorption column. It can be seen that a significant amount of O2 was adsorbed to K-IAC. The O2 concentration was reduced to 17.95% in 12 min. The O2 concentration increased gradually thereafter, increasing to 18% at 79 min. O2 adsorption onto the K-IAC decreases as crystal, formed during the initial adsorption of SO2 and NOx, occupy the activated carbon surface pores. Especially, after 120 min, O2 adsorption was at 18.01-18.05, showing neither significant increase nor decrease. At this time, NOx recorded 1115 ppm, corresponding to Cf/Ci ) 1.1072, and SO2 recorded 559 ppm, Cf/Ci ) 0.5568. This illustrates that this is the point at which many selective adsorption sites are already occupied. 4930

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 22, 2002

FIGURE 3. O2 concentration profile for competitive adsorption of NOx-SO2 binary-component. Gas composition: 891 ppm NO, 116 ppm NO2, 1004 ppm SO2, 18.1% O2. Temperature ) 403 K. Linear velocity ) 29.971 cm/sec. Bold line ) Average trend line.

FIGURE 4. CO concentration profile for competitive adsorption of NOx-SO2 binary-component. Gas composition: 891 ppm NO, 116 ppm NO2, 1004 ppm SO2, 18.1% O2. Temperature ) 403 K. Linear velocity ) 29.971 cm/sec. Bold line ) Average trend line. Figure 4 shows the concentration curve of CO emitted from the column outlet under experiment conditions of Figure 2. Illa´n-Gome´z et al. (28) established that CO appearance has no direct correlation to NO reduction activity in the TPD experiment through the NO removal study result using activated carbon loaded with K catalyst. In actuality, CO emission from experiment results could not be confirmed at low temperature and CO started to increase noticeably at temperature above 873 K. Garcı´a-Garcı´a et al. (29) reported that CO begins to evolve at temperature over 873 K and NO, at this temperature, reached the saturation point. In other words, it was concluded that CO appearance is consistent with the complete disappearance of NO. Therefore, these results showed that CO does not evolve at low temperature. Although CO emission, as shown in Figure 4, is low, less than 10 ppm, it is meaningful in revealing of one of the reaction

FIGURE 5. SEM photograph of K-IAC: (a) 5 h NOx-SO2 adsorbed (×1000) and (b) 5 h NOx-SO2 adsorbed (×10000) from condition of Table 2.

FIGURE 7. AES survey scan obtained from K-IAC adsorbed for 5 h with 1007 ppm NOx and 1004 ppm SO2 at 403 K: (a) as received, (b) after 0.5 min depth sputtering, and (b) after 60 min depth sputtering (sputter rate based on SiO2 ) 145 Å/min). column outlet. This reaction time, with the highest concentration of NO2 and CO, is also viewed as the point at which NO is unable to react with the surface oxygen and is produced due to unstable, deficient carbon in the activated carbon as shown in eq 2.

2NO + 2C f 2CO + N2

FIGURE 6. X-ray diffraction patterns of K-IAC: (a) nonadsorbed, (b) NO2-air adsorbed (Ci ) 1007 ppm), (c) SO2-Air adsorbed (Ci ) 1004 ppm), and (d) NOx-SO2-O2 adsorbed (Ci ) 1007 ppm for NOx and 1004 ppm for SO2). paths during simultaneous adsorption of NOx and SO2 on K-IAC. Also, detailed data, even with low CO concentration analysis, has not been reported in published papers to date. The concentration of CO increased, until approximately 150 min, and began to decrease again as represented by the bold line (Figure 4), the average trend line, based on real value. Similarities were found in the NO2 breakthrough curve, by NOx adsorption, in Figure 2. It was confirmed, as a result, that NO2 and CO breakthrough was observed at highest concentration at approximately 150 min. As mentioned before, the point at which maximum NO2 is when NO is converted into NO2, due to surface oxygen, and then, unable to react with KOH, an impregnant, and emitted out to the

(2)

Suface Chemical Changes in Surface Characterization of NOx and SO2 Adsorbed K-IAC. Figure 5 is SEM images of NOx and SO2 adsorbed K-IAC under experiment conditions of Table 2. Figure 5(a) is the result of magnifying by 1000 times and Figure 5(b) that of magnifying by 10 000 times. These images of NOx and SO2 adsorbed K-IAC show differentiation between white crystals and the crystals of existing activated carbon surface. Such results can be noticed through comparison with SEM imagery of nonadsorbed K-IAC in our previous study (12). The surface of nonadsorbed K-IAC not found the certain white crystals. After NOx and SO2 adsorption, however, new crystals that form the white part as in Figure 5 were observed. Figure 6 is the results of comparing nonadsorbed K-IAC, pure-component adsorbed K-IAC of 1007 ppm NO2-Air and 1004 ppm SO2-Air, respectively, and binary-component adsorbed K-IAC of 1007 ppm NOx-1004 ppm SO2 in 18.1% O2 using XRD. This was attempted to observe the crystals newly created on the surface due to NOx and SO2 adsorption as shown in the SEM results of Figure 5. A high-intensity VOL. 36, NO. 22, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4931

FIGURE 8. Raw AES depth profile data obtained from K-IAC adsorbed for 5 h with 1007 ppm NOx and 1004 ppm SO2 at 403 K.

FIGURE 9. AES montage display obtained from K-IAC adsorbed for 5 h with 1007 ppm NOx and 1004 ppm SO2 at 403 K. peak of 2θ ) 26.580, from nonadsorbed K-IAC, Figure 6(a), was confirmed as resulting from GAC. This peak was not found with Figure 6(b) NO2-Air (Cf/Ci ) 0.8138), (c) SO2-Air (Cf/Ci ) 1), and (d) NOx-SO2-O2 (Cf/Ci ) 1.1599 for NOx, Cf/Ci ) 0.8038 for SO2) adsorbed on K-IAC. This may be derived as a phenomenon in which NO2 and SO2 produce new crystals from basic OH- group and potassium distributed on the external surface and which cover the surface of K-IAC as adsorption progresses. The main peak shown as [, b in Figure 6(b) and (c) is KNO3 and K2SO4, respectively. Using International Center for Diffraction Data PC software (JCPDS, PCPDFWIN v.200), anticipated substances were selected, and the experiment results tested were referenced under the same conditions for reagents of these substances. In Figure(d) NOx-SO2-O2, major peak which is checked to (b) NO2 adsorption, 2θ of 18.880, 23.420, 29.300, 32.220, 32.920, 33.720, 40.960, 42.900, 44.020 and (c) SO2 adsorption, 2θ of 21.240, 29.700, 30.740, 35.800, 37.040, 43.360 were all observed, but the main peak intensity of SO2 was relatively higher than that of NOx. This, matched with Figure 2 results, indicates that the surface chemical change, due to the impact of SO2 adsorption, is dominant on the surface. Figure 7 shows AES survey scan results, along with SAM images, so as to confirm surface composition of K-IAC adsorbed for 5 h with the column inlet concentration of 1007 ppm NOx and 1004 ppm SO2 as shown in Table 2. It is a image of the surface magnified 2500 times using SAM. On 4932

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 22, 2002

FIGURE 10. AES depth profile after Matlab application for C, K, O, S, and N obtained from K-IAC adsorbed for 5 h with 1007 ppm NOx and 1004 ppm SO2 at 403 K considering sensitivity factor. the surface imagery are black and white particles. The white is the result of the NOx and SO2 adsorption, as noted in the pure-component adsorption study results (11-14), and indicates oxide crystals formed on the surface. To obtain data on the quantified surface composition of NOx-SO2 adsorbed K-IAC, both white and black imagery have been included considering the representativeness of the sample carefully and the most flat side was selected. Figure 7(a) is the result of analyzing external surface of K-IAC, and Figure 7(b) and (c) are the result of measuring K-IAC after depth sputtering at 0.5 min and 60 min, respectively. For the sample that went through the depth sputtering of 60 min, the peak intensity of S was reduced noticeably. N was detected, though in very small quantities, and the diminutive quantity of

FIGURE 11. Relative integrated peak intensity of main negative ion by ToF-SIMS analysis obtained from K-IAC adsorbed for 5 h with 1007 ppm NOx and 1004 ppm SO2at 403 K. adsorbed NOx was probably not within instrument detection limits. This result is consistent with the above-mentioned result of Figure 2, confirming the fact that a larger amount of SO2 than NOx is adsorbed, dominating most of the selective adsorption site due to its high adsorption affinity to K-IAC compared to NOx. The similar level of intensity for C is shown, confirming that it did not participate in the reaction producing oxide crystals on the surface. Figure 8 is the result of the raw depth profile which investigated the composition distribution by depth sputtering the surface using 3.5 kV Ar ion beam. Such interpretation of the depth profile is clearly illustrated by referring to the Figure 9 showing the montage display of the AES spectrum. The sputter rate illustrates the depth of 145 Å based on the SiO2 per minute. When examining the trend of each element by depth, K and S show a continuous decrease. K and S were

well diffused until 35 min corresponds to depth of 5075 Å. Concerning the reaction of K-S-O, the distribution of O on the external surface is relatively low up to the sputter time of 6.5 min (942.5 Å). This is because the analyzed position of sample was selected for a large amount of SO2 rather than chemisorbed sites of the oxygen. O increased little by little until 17 min (2465 Å) and decreased very gradually afterward. Despite the fact that the reaction is adsorbed due to K-S-O or K-N-O, the reason O has a different pattern than K, S, N is because NOx is not uniformly adsorbed in pores and surface of K-IAC in addition to the fact that the distribution of the chemisorbed amount of O is different. N can be explained using Figure 9. When looking at AES Montage display of Figure 9, we can clearly see that NOx was adsorbed on K-IAC. NOx was distributed with different kinetic energy due to chemical shift with very irregular pattern according to depth. This can be interpreted as oxide crystal containing N having different oxidation number corresponding depth. Figure 10 shows the quantitative results in atomic % units through raw AES depth profile (Figure 8) and montage display (Figure 9). The representative curve fitting used the PHI matlap program to remove noise, separate peaks, and apply sensitivity factors per each elements. At onset, C,K,S,O,N ratio of atomic % was 52.59:30.25:9.31:7.75:0.11, respectively, and, after 60 min, 57.29:24.13:17.32:1.18:0.09. C was distributed in relatively small amounts through oxide crystal of K-S-O and K-N-O in the beginning of the sputter time, i.e., in the depth close to the external surface, but C existed in relatively larger amounts in proportion to sparser oxide crystal distribution per increased depth. For N, the relatively high points at all depths were 1.55 at sputter time of 2 min (290 Å), 1.83 at 8 min (1160 Å), 1.27 at 28 min (4060 Å), and 1.39 at 36 min (5220 Å). Very small amounts of N were adsorbed and very irregularly distributed according to depth. Previous study results, on NO2-Air in AES/SAM (11), proved

FIGURE 12. SIMS depth profiles of (a) SOx- (x ) 1,2,3,4) and (b) NOx- (x ) 1,2,3) ions obtained from K-IAC adsorbed for 5 h with 1007 ppm NOx and 1004 ppm SO2 at 403 K. VOL. 36, NO. 22, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4933

FIGURE 13. Two predicted main mechanisms for adsorption by chemical reaction of (a) NOx and (b) SO2 on K-IAC in excess O2. that oxide crystals of KNO2 and KNO3 formed on the surface due to the chemical bond of K, N, and O. Accordingly, even as such irregular distribution occurs, there is confirmation that distribution is the result of the formation of oxide crystals of KNO2 and KNO3 or the C-N bonding caused by deficient carbon in the sample. SO2 was well diffused to 5075 Å, at approximately 35 min. The amounts of K, S, and O, diffused in the pores, decreased in proportion to increased depth, reducing in surface composition. This is believed to result from a pore blocking effect caused by the K-S-O chemical bond inducing oxide crystal production. K-S-O was relatively well diffused to 5075 Å depth, where reaction occurred, and adsorbed to the surface. Deeper than 5075 Å, the atomic % of K, S, O, and N, diffused into pores, reduced as the extent of the block increased due to the crystal production on the external surface. This is believed to obtain the smaller distribution of S and O amounts. AES/SAM analysis indicated that oxide crystal was produced partially because of NOx as SO2 with higher adsorption affinity became dominant in most selective adsorption sites over K-IAC where SO2 and NOx (NO:NO2 ) 7.7:1) coexisted. It was also noted that large amounts of oxygen were chemisorbed. Figure 11 shows the result of comparing the test sample with the integrated peak intensity analyzed using ToF-SIMS with the samples of depth sputtering for 5 min and 60 min with 1007 ppm NOx and 1004 ppm SO2 absorbed on K-IAC for 5 h. The SIMS depth profile was conducted, in addition to the quantitative analysis of the elements in AES, so as to clarify molecular ion data. The observed negative ions showed the result of SOx- (x ) 1,2,3,4), NOx- (x ) 1,2,3), OH-, KO-, Cx- (x ) 1,2), and CxH- (x ) 1,2). The integrated peak intensity is the result of integrating the peak area of molecular ion species from the analysis of ToF-SIMS. The amounts of SOxand NOx- were more distributed in the depth sputtering of 5 min. SOx- also decreased in larger amounts than NOx-. 4934

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 22, 2002

Meanwhile, amounts of OH- and KO-, in KOH, as selective adsorption sites, were reduced in the depth sputtering of 5 min. NOx and SO2 adsorption on K-IAC occurred by the chemical reaction between impregnant (KOH) and adsorbates (NOx and SO2), clearly indicating gradual function loss for KOH as selective adsorption site for NOx and SO2 adsorption. When NOx and SO2 coexist, competitive adsorption occurs, reducing OH- and KO- presence as ions of selective adsorption sites, resulting from reaction between KOH and SO2. Cx- and CHx- are groups mainly formed on the existing activated carbon surface and these groups provide positive evidence that K-S-O and K-N-O are concentrated near the external surface. Figure 12 is the result of observing SIMS depth profile after sputtering (a) SOx- (x ) 1,2,3,4) and (b) NOx- (x ) 1,2,3) in the direction of the surface depth. Four ions - SO-, SO2-, SO3-, SO4- - formed as oxide crystal, due to the reaction of K-S-O shown in AES/SAM, were mainly distributed near the external surface and were gradually decreased in counts in proportion to increasing depth. A small amount of N was confirmed in the AES result, but as it clearly exists in NOxform in ToF-SIMS could be proven. NOx- is also irregularly distributed across depth, the same as shown in the AES result, but it was distributed closer to the external surface. Analysis results confirmed the blocking effect caused by oxide crystal production from NOx and SO2 adsorption. SOx-, in particular, had higher counts than NOx- due to higher adsorption affinity to the K-IAC surface. NOx- also showed larger distribution in depths closer to the external surface, showing depth profile of irregular distribution. We concluded that the reason the counts of these ions were decreasing according to depth because the surface reaction inside the pore discontinued from pore blocking of K2SO3, K2SO4 or KNO2, KNO3. It may be concluded, from reviewing ToF-SIMS analysis interpreted in conjunction with

AES/SAM results, that KOH inherently provides selective adsorption sites for NOx and SO2. However, it can be proven that SO2 occupies most of the selective adsorption sites due to higher adsorption affinity, in comparison to NOx, to K-IAC. Predicted Main Mechanisms in Binary-Component Adsorption of NOx and SO2 on K-IAC. Figure 13 illustrates the two predicted main mechanisms for adsorption by chemical reaction of (a) NOx and (b) SO2 adsorbed on K-IAC. For NOx adsorption to K-IAC, as shown in 1, NO2 is well adsorbed on K-IAC, but in the case of NO, it has to go through step 2 and converted to NO2 and then adsorbed on K-IAC. Steps 3 and 3′ show that KNO2 may also exist as oxide crystal. Conversion from KNO2 to KNO3 (3 f 4, 3′ f 4′) is caused by the extent of surface becoming oxidized. NO produced in step 4 goes through step 2 again and is adsorbed at other available adsorption sites. NO2 produces NO during adsorption (3 f 4). NO production, therefore, indicates that adsorption is occurring. For SO2 adsorption, the main product on K-IAC is K2SO3 (5 f 6, 5 f 6′) and K2SO4 (6 f 7, 6′ f 7′). Because SO2, compared to NOx, has higher adsorption affinity to K-IAC, more amounts of SO2 were adsorbed in an atmosphere in which NOx and SO2 coexisted. As no evidence was found indicating that oxide crystal is produced as a result of the chemical bond of S and N, as shown in the surface characterization result, it may be concluded that oxide crystals are formed with SO2 and NOx adsorbed to K-IAC, respectively, through two reaction mechanisms.

Literature Cited (1) Harding, A. W.; Brown, S. D.; Thomas, K. M. Combustion Flame 1997, 107, 336. (2) Ko¨psel, F. W.; Halang, S. Fuel 1997, 76, 345. (3) Mochida, I.; Kawabuchi, Y.; Kawano, S.; Matsumure, Y.; Yoshikawa, M. Fuel 1997, 76, 543. (4) Tsuji, K.; Shiraishi, I. Fuel 1997, 76, 549. (5) Zhu, Z.; Liu, Z.; Liu, S.; Niu, H. Fuel 2000, 79, 651.

(6) Mun ˇ iz, J.; Herrero, J. E.; Fuertes, A. B. Appl. Catal. B 2000, 18, 171. (7) Burch R.; Millington, P. J.; Walker, A. P. Appl. Catal. B 1994, 4, 65. (8) Mun ˇ iz, J.; Marba´n, G. A.; Fuertes, B. Appl. Catal. B 2000, 27, 27. (9) Yun, J. H.; Choi, D. K.; Kim, S. H. Ind. Eng. Chem. Res. 1988, 37, 1422. (10) Yun, J. H.; Choi, D. K.; Kim, S. H. AIChE J. 1999, 45, 751. (11) Lee, Y. W.; Choi, D. K.; Park, J. W. Ind. Eng. Chem. Res. 2001, 40, 3337. (12) Lee, Y. W.; Choi, D. K.; Park, J. W. Sep. Sci. Technol. 2002, 37, 937. (13) Lee, Y. W.; Choi, D. K.; Park, J. W. Carbon 2002, 40, 1409. (14) Lee, Y. W.; Park, J. W.; Choung, J. H.; Choi, D. K. Environ. Sci. Technol. 2002, 36, 1086. (15) Illa´n-Gome´z, M. J.; Linares-Solano, A.; Radovic, L. R.; SalinasMartı´nez de Lecea, C. Energy Fuels 1995, 9, 97. (16) Garcı´a-Garcı´a A.; Chincho´n-Yepes, S.; Linares-Solano, A.; Salinas-Martı´nez de Lecea, C. Energy Fuels 1997, 11, 292. (17) Zhu, Z. H.; Lu, G. O. J. Catal. 1999, 187, 262. (18) Guo. J.; Lua, A. C. J. Chem. Technol. Biotechnol. 2000, 75, 971. (19) Tsuchiai, H.; Ishizuka, T.; Nakamura, H.; Ueno, T.; Hattori, H. Ind. Eng. Chem. Res. 1996, 35, 851. (20) Lizzio, A. A.; DeBarr, J. A. Energy Fuels 1997, 11, 284. (21) Cal, M. P.; Strickler, B. W.; Lizzio, A. A. Carbon 2000, 38, 1757. (22) Davini, P. Carbon 2000, 39, 419. (23) Lee, M. R.; Wolan, J. T.; Hoflund, G. B. Ind. Eng. Chem. Res. 1999, 38, 3911. (24) Kong Y.; Cha, C. Y. Carbon 1996, 34, 1027. (25) Aurora, M. R.; John, M. S. Fuel 1997, 76, 521. (26) Neathery, J. K.; Rubel, A. M.; Stencel. J. M. Carbon 1997, 35, 1321. (27) Wilde, J. D.; Marin, G. B. Catal. Today 2000, 65, 319. (28) Illa´n-Gome´z, M. J.; Linares-Solano, A.; Radovic, L. R.; SalinasMartı´nez de Lecea, C. Energy Fuels 1996, 10, 158. (29) Garcı´a-Garcı´a, A.; Illa´n-Gomez, M. J.; Linares-Solano, A.; Radovic, L.; Salinas-Martienez, C. Fuel 1997, 76, 499.

Received for review December 31, 2001. Revised manuscript received August 27, 2002. Accepted September 11, 2002. ES011510B

VOL. 36, NO. 22, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4935