SAM and ToF-SIMS on

Surface Chemical Characterization Using AES/SAM and ToF-SIMS on KOH-Impregnated Activated Carbon by Selective Adsorption of. NOx. Young-Whan Lee* ...
0 downloads 0 Views 239KB Size
Download PDF
Ind. Eng. Chem. Res. 2001, 40, 3337-3345

3337

Surface Chemical Characterization Using AES/SAM and ToF-SIMS on KOH-Impregnated Activated Carbon by Selective Adsorption of NOx Young-Whan Lee* and Dae-Ki Choi Environment & Process Technology Division, Korea Institute of Science & Technology, P.O. Box 131, Cheongryang, Seoul 130-650, Korea

Jin-Won Park Department of Chemical Engineering, Yonsei University, Seoul 120-749, Korea

KOH-impregnated activated carbon (K-IAC) was studied as a NOx adsorbent. When adsorbing NO2/air to K-IAC, as the concentration increases, the removal efficiency relatively decreases because the function of the selective adsorption site for NOx is lost faster. By using AES/SAM and ToF-SIMS, we examined the chemical characteristics created on K-IAC through adsorption of NOx. Oxide crystals formed on the surface of K-IAC after adsorption of NOx was a result of the chemical bond among K, N, and O. KNO3 was predominantly distributed on the external surface, and the pore of K-IAC showed blocking from the crystal growth of KNO3. The results showed that surface basic OH-, well developed on the surface because of KOH impregnation, acts as a selective adsorption site. This study pointedly illustrates selective adsorption behavior from strong chemical interaction between NOx and K-IAC. 1. Introduction Anthropogenic sources of NOx include transport vehicles such as automobiles and airplanes and stationary sources such as fossil fuel power plants, boilers, and incinerators. Other sources include direct inflow from the stratosphere and oxidization of NH3. Our study targets combustion flue gases that are categorized as a stationary source. Selective catalytic reduction (SCR) is being adopted by many commercialized technologies of today to control NOx emitted from stationary sources. Though SCR is known to be excellent in removing NOx, it is one of the most expensive NOx-controlling technologies. High initial investment is required for catalysts, state-of-the-art monitoring equipment, and instrumentation. Large operating expenses are also required for catalyst refills, NH3 consumption, and electricity.1-3 Besides, though NOx-inhibiting technologies have continuously been advanced, thanks to the efforts of many researchers, no omnipotent process is yet in place, because it is naturally very difficult to control NOx. A NOx-controlling technology should be commercially feasible in terms of relative efficiency and applicability. Especially, to be able to select the most optimum technology, economic evaluation based on cost efficiency should be made. Adsorption requires little electricity and needs no additional cost for treatment, because it does not cause secondary pollution. Adsorption processes are simple to design and take little space. Thus, it can be installed by adding a bypass line to an existing NOx treatment facility. The possibility of removing harmful gas using carbon materials during the adsorption process has been examined because the unique characteristics of carbon materials themselves.4-9 As a result, it was induced that * To whom correspondence should be addressed. E-mail: [email protected].

if alkali metal is impregnated in carbon material, the effect of controlling NOx is maximized. In particular, potassium hydroxide (KOH) used as the impregnant in this study has positioned itself as one of the promising chemical substances able to increase the removal efficiency of NOx by enhancing selective adsorptivity of NOx.10 Here, the reason for the increase in the removal efficiency of NOx is caused by changes in the surface functional group due to impregnation of KOH and the effect of catalytic action. Therefore, the information on transformation of the surface chemistry due to adsorption of NOx can be regarded as extremely significant in understanding the adsorption reaction. However, comprehensive examination of the surface chemistry resulting from NOx behavior on activated carbon has not yet occurred.11 To date, researchers have used mostly temperature-programmed desorption (TPD) or X-ray photoelectron spectroscopy (XPS) without a depth profile as the means to discover the characteristics of the surface chemistry regarding NOx adsorbed carbon materials.12-14 In studies where TPD functions as a means to reveal the adsorption and desorption mechanism by increasing surface temperature, it is difficult to obtain direct and detailed information on the phenomenon of chemical distribution and transformation at the surface level. In addition, though XPS acts as a powerful tool in producing information on the chemical state distributed on the surface, it is not easy to discriminate the peak of N because of the detection limit occurring at low-concentration NOx adsorption. Furthermore, because the roughness and irregularities of the surface of activated carbon bring changes in resolution, it can bring about confusion in the interpretation. Therefore, to acquire more credible surface information, a complementary system to use in conjunction with other analytical methods is needed. In addition, because information on diffusion inside the pore is very impor-

10.1021/ie001068q CCC: $20.00 © 2001 American Chemical Society Published on Web 06/30/2001

3338

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001 Table 1. Experimental Conditions

exp.a no.

adsorbentb

adsorbate

NO2 concn (ppm)

run 1 run 2 run 3

K-IAC K-IAC K-IAC

NO2/air NO2/air NO2/air

153 400 1014

contact time (s)

linear velocity (cm/s)

0.062 0.062 0.062

32.49 32.49 32.49

a Adsorption tests were performed on K-IAC at a temperature of 130 °C for 5 h. b 1.067 g of K-IAC consisting of 9.96 wt % potassium was introduced into the column.

Figure 1. Schematic diagram of the experimental system.

tant in adsorption and catalytic reaction, an explanation of diffusion inside the pore only through survey scan is insufficient and, accordingly, research on the surface depth profile needs to be accompanied. In this study, Auger electron spectroscopy/scanning Auger spectroscopy (AES/SAM) and time-of-flight secondary ion mass spectroscopy (ToF-SIMS), differing from existing analysis and interpretation methods, have been used.15,16 These two instruments are, along with XPS, regarded as the three major types of instruments for surface analysis.17,18 Despite the fact that they are of much aid in understanding the surface phenomenon, they have not yet been adopted as analytical methods in the research on adsorption of NOx on impregnated activated carbon. The objective of this study was to compare adsorptivity according to the NO2 concentration on KOHimpregnated activated carbon (K-IAC) and to observe the distribution of surface elements and ion species from the adsorbed sample by using both AES/SEM and ToFSIMS, through which we attempted to clarify the surface chemical characteristics of K-IAC caused by NOx adsorption behavior. 2. Experimental Section 2.1. Adsorbent Preparation. The adsorbent was prepared by impregnating a KOH (Junsei Chemical Co.) solution into granular activated carbon (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 140 °C. Next the treated GAC was dried at 110 °C. KOH was impregnated in an aqueous solution state in the GAC via incipient wet impregnation and was used following dehydration at 130 °C. 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 to be 9.96 wt %. 2.2. Apparatus and Method. The fixed-bed adsorption system used in this experiment is shown in Figure 1. The fixed-bed adsorption column was a 316 stainless steel tube, with an inside diameter of 10.9 mm and a length of 400 mm. 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-integraldifferential (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 5000 ppm NO2/N2 was diluted to the desired concentration range of 153, 400, and 1014 ppm via a mass flow controller (Brooks Co., model 5280E). In the fore-end of the adsorption column, an in-line static mixer was installed to facilitate mixing. Concentrations of NO and NO2 that exhausted from the bypass line and adsorption column were analyzed by using a chemiluminescent NOx analyzer (Thermo Environmental Instruments Inc., model 42C). Daily NOx analyzer calibrations were performed with N2 (zero value) and NO and NO2 gases certified by manufacturer analysis having a concentration of near 80% of the analyzer full-scale range (span value). During the adsorption experiment, analysis on column outlet concentrations of NO, NO2, and NOx is controlled with software (Thermo Environmental Instruments Inc.). Analytical data are captured every minute and read into a computer, generating 300 analytical values from every of 5-hour-long experiment. For all experiments, 1.067 g of K-IAC (or equivalent to 2 cm of bed depth) is packed into the column, and conditions for adsorption are set as follows: concentrations of 153-1014 ppm, linear velocity of 32.49 cm/s, and temperature of 130 °C. To purge impurities from the system line before an adsorption is made, it is flowed with He for 30 min at the flow rate of 500 mL/min at 200 °C. Then, the temperature is reset to 130 °C for the experiment, and the column is packed with adsorbent and repurged with He for 10 min. Then, the zero value of the NOx analyzer is checked before proceeding with the experiment. Detailed conditions of each experiment are shown in Table 1. 2.3. Sample Characterization. The specific surface areas and pore volumes of GAC and K-IAC were obtained via an automatic volumetric sorption analyzer (Quantachrome, Autosorb 1) using N2 adsorption at 77 K. Prior to the measurements, the sample was outgassed at 200 °C under a N2 flow for at least 3 h. Although the true surface area of microporous materials is difficult to calculate by existing models, the BrunauerEmmet-Teller (BET) method is widely used to obtain an apparent surface area.19 The BET surface area of the samples was calculated from the N2 adsorption isotherms by assuming the area of a N2 molecule to be 0.162 nm2. The t-plot method was applied to calculate the micropore volume. Micropore volumes were estimated to be the liquid volumes of the adsorbate (N2) at a relative pressure of 0.993.

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001 3339 Table 2. Physical Properties of K-IAC property m-3

bulk density, kg BET surface area, m2 g-1 mesopore area, m2 g-1 micropore volume, cm3 kg-1 average pore diameter, Å

K-IAC [GAC, loss (%)] 570 719 [1160, 38] 15 [42, 64] 232 [347, 33] 11.5

Before and after adsorption, surface analysis was performed for samples by utilizing AES/SAM (PerkinElmer, PHI model 670) and static-type ToF-SIMS (Perkin-Elmer, PHI model 700 ToF-SIMS/SALI). The AES survey scan and sputter depth profile were recorded using the following experimental 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 µA m2, was produced by a differentially pumped ion gun. In this analysis, AES was used quantitatively to help determine the chemical state of potassium (K), oxygen (O), carbon (C), and nitrogen (N) on the K-IAC. SAM was applied for the morphology analysis of the K-IAC. ToF-SIMS analysis was carried out using a system equipped with a two-stage reflectiontype 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. In this test, the background was corrected by measuring and comparing ICP-AES (Labtest Plasmascan, model 710) and ToF-SIMS. During NOx-adsorbed K-IAC analyses, tendencies were interpreted by comparison with nonadsorbed K-IAC, increasing data credibility.

Figure 2. K elemental dot mapping result of K-IAC.

3. Results and Discussion 3.1. Physical and Chemical Properties of the K-IAC. We observed the character changes of K-IAC manufactured with impregnation of KOH on GAC. According to the study by Illa´n-Gome´z and co-workers,20 after KOH impregnation, the N2 surface area decreased, which results from partial blocking of pores due to potassium. However, because potassium possesses high NO removal capability when present on the carbon surface at low temperature (below 200 °C), it is said that an increase in the adsorptivity may be expected. Lee and co-workers21 reported that the BET surface area decreased because of alkali processing and the degree of decrease is proportional to the alkali radius. The physical properties of the K-IAC and GAC used are tabulated in Table 2. Because of impregnation, the surface area of K-IAC decreased. The BET surface area of K-IAC indicated 719 m2/g, 38% decreased compared to GAC. The micropore volume decreased by 33%. The reason for such a decrease is because the pore of K-IAC was blocked because of KOH impregnation. Figure 2 illustrates the mapping image of potassium using electron probe microanalysis (EPMA; JEOL, model JXA8600) in order to check if the potassium on the surface is distributed uniformly on the surface of the activated carbon after KOH impregnation. EPMA was carried out

Figure 3. EDX spectrum of K-IAC.

with the accelerating voltage at 20 kV with 1000 times magnification. The white parts of the image represent potassium, which confirms that the impregnation is progressed properly because potassium is, in general, uniformly distributed on the K-IAC. However, on an extremely limited part of the surface, there were distributed clusters of potassium. Figure 3 shows the quantitative analysis of the component of the sample from the same portion as in Figure 2 using an EDX spectrum. We were able to confirm the distribution of a lot of potassium due to KOH impregnation. K-IAC showed large amounts of C and O with a K peak as the major, and impurities of Si, Mg, Na, and an extremely small amount of Fe were observed. Au was a peak that appeared because we conducted gold coating as a preanalysis process. 3.2. Correlation between the Effect Based on Concentration and the Amount of NO Produced in Adsorptivity. Figure 4 shows the tendency of the representative concentration curve of NO2 adsorption on K-IAC. It can be explained as the process of NO2 undergoing catalytic conversion into NO through selective adsorption site of K-IAC, through which K-IAC gradually loses its selective adsorption site function,

3340

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001

Figure 4. Representative concentration curve of NOx adsorption on K-IAC.

resulting in the increased amount of NO2 that exhausts without being adsorbed. In Figure 4, NO was produced more than NO2 in A because the selective adsorption site is predominant on the surface, and B becomes the opposite of A. B is the point at which the concentration of NO2 becomes higher than that of NO. Considering the basis of determining NOx adsorptivity, the optimum condition would be to delay the time of C, which is the crossing point of A and B. Figure 5 illustrates the experiment with varied NO2 concentrations of 153, 400, and 1014 ppm in the presence of 21% oxygen under the conditions of Table 1. According to research reports, higher adsorptivity is indicated with oxygen than without oxygen.22-24 Illa´n-Gome´z and co-workers25 reported that the enhancement of adsorptivity from the addition of oxygen, considering the fact that excessive amounts of oxygen exist in the typical emission gas, can be regarded as an advantage to differentiate from SCR. In Figure 5, the crossing point, C, of A and B at the respective concentration curve is marked. At a NO2 concentration of 153 ppm, the time of C indicated 65 min, but it has been reduced to 12 min at 400 ppm, and 8 min at 1014 ppm, showing the phenomenon of extreme reduction in selective adsorptivity as the concentration increased. Figure 6 illustrates the removal efficiency of NOx according to time under the same conditions as those in Figure 5. At an NO2 adsorption of 153 ppm, more than 20% of the removal efficiency was maintained for 300 min even when the reaction was completed, but in the case of 1014 ppm, the tendency was to reach saturation in 140 min. That is to say, Figure 6 shows that as the relative concentration of NOx increases, the efficiency of removal decreases. Table 3 shows the relative production amounts of NO and NO2 according to the reaction concentration under the same conditions as those in Figures 5 and 6. The produced amount of NO against the total NOx in terms of the relative production ratio of NO and NO2 was calculated. NO produced 30% of the total NOx for 300 min at 153 ppm, and the remaining 70% was occupied by NO2. The NO produced at 400 ppm was 15%, and at 1014 ppm, it remained at 10.7%, which confirmed that, as the concentration increases, the amount of NO produced against the total NOx decreases. If the amount of NO

Figure 5. Effect of the NO2 concentration on K-IAC.

increases, A becomes broader, requiring more time to reach point C. It delays the breakthrough lead time as much. This, in turn, means that the removal efficiency will increase. When the production amount of NO (Table 3) and the removal efficiency (Figure 6) are compared for 153, 400, and 1014 ppm, a higher concentration of NO2 makes the driving force larger, leading to increased adsorption rate. In summary, while the adsorption rate increases under higher concentration on the same time, the removal efficiency decreases because NOx selective

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001 3341

Figure 6. Removal efficiency obtained from the inlet concentration of (a) 153 ppm NO2/air (run 1), (b) 400 ppm NO2/air (run 2), and (c) 1014 ppm NO2/air (run 3) on K-IAC.

Figure 7. AES profile obtained from K-IAC adsorbed for 5 h with 1014 ppm of the inlet NO2 concentration. Table 3. Relative Production Amount of NO and NO2 with Variation of the NO2 Concentration on K-IAC NO2 concn (ppm) 153 400 1014

NOx

relative production ratio

produced amount (%) against total NOx

NO NO2 NO NO2 NO NO2

0.429 2.330 0.176 5.686 0.120 8.361

30 70 15 85 10.7 89.3

adsorption site, which consists of KOH, loses its function more quickly. 3.3. AES/SAM Survey Scan. Figure 7 shows the survey scan of K-IAC adsorbed for 5 h with 1014 ppm of inlet NO2 concentration (run 3) using AES/SAM. From the results of observation of a 2500 times enlarged surface using SAM, we have assigned three parts with major differences and analyzed them. From the SAM image, the white part represents where the SAM itself occurs by charge from oxide. That is, the white part indicates formation of a certain oxide crystal caused by NO2 adsorption. In Figure 7, position 1 is the part where charging is not seen with the naked eyes, position 2 is the part where charging takes place in the form of a white color thinly coated on the surface, and position 3 is the part where oxide crystals is clearly distinguished so that charging occurs more strongly. By analyzing the AES survey scan, we found that there is a close relationship between the intensity of K and O with the existence of N. In position 3, the intensity of K and O

Figure 8. AES montage display obtained from (a) nonadsorbed K-IAC, (b) NO2-adsorbed K-IAC (400 ppm NO2/air, run 2), and (c) NO2-adsorbed K-IAC (1014 ppm NO2/air, run 3).

both appeared greater than those in positions 1 and 2. On the other hand, in the case of C, the intensity was lowered at position 3, implying that it did not take part in the reaction during oxide crystallization. Because of the element detection limit of AES (approximately 0.11%), it was difficult to distinguish the peak intensity of the N peak. However, because N exists in positions 1-3, the substance formed on the surface can be estimated to be an ionic crystal such as KNOx (x ) 2 and 3). In other words, it can be interpreted that, because of NOx adsorption, the intensity of each element from the AES survey scan is based on the degree of formation of oxide crystals in the form of KNOx. 3.4. AES Montage Display and Depth Profile Following Matlab Application. Figure 8 shows a AES montage display. Nonadsorbed K-IAC and NO2adsorbed K-IAC (run 2) are the results that gained 20 cycles by conducting sputtering for 10 min with 0.5 min intervals. Each element represents O, C, K, and N respectively from left to right. As for N, it is difficult to discern the result in nonadsorbed K-IAC and NO2adsorbed K-IAC (run 2) because of the AES detection limit. In nonadsorbed K-IAC, O, C, and K showed no significant changes in the intensity. However, for NO2adsorbed K-IAC (run 2), while there was little change in the intensity of C, the intensity of O and K showed tendencies to decrease. NO2-adsorbed K-IAC (run 3) is a result from gaining 60 cycles through 30 min of sputtering with 0.5 min intervals. As the concentration of NO2 becomes higher, the intensity of the major component of activated carbon, C, remains still to

3342

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001

Figure 9. AES depth profile after Matlab application for K, O, and N obtained from K-IAC adsorbed for 5 h with 1014 ppm of the inlet NO2 concentration considering the sensitivity factor.

sputter time, while the intensity of K, N, and O showed trends of decrease. Therefore, this phenomenon is considered in that, because of NO2 adsorption, K, N, and O interact with each other and make a chemical bond on the surface of K-IAC, and with this, we can suggest a strong interaction between K and NOx. Here, it can be predicted that the form of adsorption on K-IAC occupies selective adsorption sites by forming ionic oxide crystals such as KNO2 or KNO3 due to potassium, because the potassium has a high reactivity and cannot remain as a single element but exists in a form of chemical compounds bonded with other elements. Figure 9 shows the quantitative result of illustrating the K-IAC that reached saturation (C/C0 ) 1) in adsorption of 5 h under 1014 ppm NO2/air in atomic percent of K, N, and O using the PHI matlab program through the AES montage display and depth profile. K, N, and O greatly decreased until 11 min sputter time, which indicates that on the surface K, N, and O coexist, forming a KNOx crystal. The initial atomic ratio of K:O:N was 9.7:5.6:1.3, then at 11 min 5.3:0.8:1.3, at 20 min 4.5:1.3:0.9, and at 30 min 4.3:0.8:0.5, respectively. For K, it mostly showed a certain amount of decrease, but as for N and O, we could confirm that they exist in a very irregular pattern according to the depth of the surface. The point where the amount of N and O is distributed uniformly is the point where a sputter time of 5 min has elapsed, which clarifies that the form of oxide in KNOx can differ. That is, KNO3 was distributed predominantly until the initial 5 min elapsed. 3.5. ToF-SIMS Qualitative Analysis. Figure 10 shows the ToF-SIMS spectrum of the external surface of both nonadsorbed K-IAC and NO2-adsorbed K-IAC (run 3) in order to see the changes in the main ion species. The result of analysis on the polarity of negative ions detected through the analysis was represented within the range of m/z 1-100 where information on the main peak from the interaction between KOH and NOx can be obtained. In the m/z outside this range, most of the fragmented ions appeared by a variety of functional groups existing in an aromatic compound of activated carbon, and the counts are relatively small compared to m/z 1-100, showing little change on the external surface. That is to say, we determined that these are not of great significance in dealing with the NO2-adsorbed K-IAC and is not dealt with in this paper. The result of comparing the qualitative analysis of the ToF-SIMS spectrum showed the following differences.

Figure 10. Negative spectrum of ToF-SIMS for qualitative analysis: (a) nonadsorbed K-IAC; (b) NO2-adsorbed K-IAC (1014 ppm NO2/air, run 3).

The main ions with increased counts compared to nonadsorbed K-IAC were CN- (m/z 26), CNO- (m/z 42), NO2- (m/z 46), and NO3- (m/z 62), while the main ions with decreased counts can be assigned to H- (m/z 1), C- (m/z 12), CH- (m/z 13), O- (m/z 16), OH- (m/z 17), C2- (m/z 24), C2H- (m/z 25), O2- (m/z 32), KO- (m/z 55), and CO3- (m/z 60). C-, CH-, C2-, and C2H- are assigned to aromatic carbon with and without bonds to hydrogen, respectively.26 From the comparison between the before and after of the adsorption, it was found that NO3- that was almost nonexistent before adsorption was formed in high intensity after adsorption. This made us think that the adsorption takes place in the form of NO3- or in the oxidation from NO2- to NO3- at the selective adsorption site. The amount of NO2- increased only a little. After the NOx adsorption on the external surface, the reaction reached its termination due to complete oxidation where most of them could not exist as NO2but as NO3-. Because the decrease of KO- and OH- as the main ion species of KOH was relatively clear, we could prove that KOH provides selective adsorption sites for NOx. To prove this even more clearly, we tried to obtain specific information on OH-, NO2-, and NO3-. Figure 11 is the result from integration of the respective peak intensity by counts of ions of nonadsorbed K-IAC and NO2-adsorbed K-IAC (run 3) to the external surface. For the sample that reached saturation (C/C0 ) 1) after

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001 3343

Figure 11. Integrated value of the ToF-SIMS peak intensity (IPI) of the main ion species for (a) OH- and (b) NO3- obtained from nonadsorbed K-IAC and NO2-adsorbed K-IAC to the external surface.

adsorption of 1014 ppm NO2 for 5 h, it showed that relatively 80% of OH- was lost compared to the nonadsorbed K-IAC and NO3- showed about a 98% production ratio. In Figure 11, if we look at the ratio of the area after the sample cleaning process, among ion species of NOx, NO3- was found to be distributed predominantly on the external surface because of termination of the complete oxidation. The reason NO3present on the surface was formed in higher counts than the loss of OH- can be explained from the phenomenon caused by the interaction on the external surface with a functional group present in existing GAC other than surface basic OH- ion. It supported the fact that the decrease in OH- is caused by NOx adsorption and provides a selective adsorption site for NOx. 3.6. SIMS Depth Profile. Figure 12 shows the sputter depth profile of NO2- and NO3- at NOx-adsorbed K-IAC of C/C0 ) 0.42 and C/C0 ) 1. In the depth profile of KIAC representing C/C0 ) 0.42, the counts of NO2according to sputter time decreased, showing a similar tendency with NO3-. For NO2-, the external surface increased only a little at C/C0 ) 1 compared to nonadsorbed K-IAC (Figure 10), but in the depth profile, the sputter time greatly increased until 180 s, but after that, it showed a gradual increase until 1000 s. However, NO3-, unlike NO2- at C/C0 ) 1, greatly increased until a sputter time of 180 s but showed a tendency to gradually decrease after that. That is to say, as for the sample of the reached saturation (C/C0 ) 1), only NO3showed a decrease. We interpreted the reason behind

Figure 12. SIMS depth profile of (a) NO2- and (b) NO3- ions obtained from K-IAC with C/C0 ) 0.42 (adsorbed for 40 min with 153 ppm of the inlet NO2 concentration) and C/C0 ) 1 (run 3 at Table 1), respectively.

such phenomenon to be as follows. KNO3 formed by a reaction-blocked pore, stopping any further progress of the surface reaction inside this pore, while nonreacted NO2 molecules and a significant amount of KNO2 existed on the inside of the pore blocked by KNO3, resulting in a continuous increase of NO2- and a decrease of NO3-. In other words, more rapid diffusion of NO2 on the external surface, where NO2 formed crystals of KNO2 and KNO3 with potassium on the external surface but with continuous adsorption with NO2, formed KNO2 is in the end oxidized and becomes KNO3. Inside pores blocked by KNO3, already diffused NO2- species could not further the progress of the reaction and remained there, failing to reach the termination reaction. Such a phenomenon clarifies the fact from the result of the previous AES depth profile (Figure 9) that the actual decrease of the peak intensity of K, N, and O results from the formation of KNO3 crystals. In addition, it was found that, in the sample that reached saturation by adsorption for 5 h under the concentration of 1014 ppm NO2 in the presence of 21% O2, KNO3 was present at the maximum amount on the surface under 180 s (Si based sputter rate ) 60 Å/min). Figure 13 shows the depth profile of OH- on nonadsorbed K-IAC (C/C0 ) 0) and NO2-adsorbed K-IAC (C/ C0 ) 1, run 3). OH- at NO2-adsorbed K-IAC (run 3) showed a great decrease until 180 s where NO2- and NO3- greatly increased and, after that, showed continuous and gradual decrease compared to the nonadsorbed K-IAC. Therefore, it demonstrates that surface basic OH-, well developed on the surface because of KOH impregnation, acts as a selective adsorption site for NOx

3344

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001

Figure 13. SIMS depth profile of the OH- ion obtained from K-IAC with C/C0 ) 0 (nonadsorbed) and C/C0 ) 1 (run 3 at Table 1), respectively.

by the following mechanism.

2KOH + 2NO2 f KNO2 + KNO3 + H2O

(1)

KNO2 + NO2 f KNO3 + NO

(2)

In conclusion, the result of analysis by ToF-SIMS, interpreted with AES/SAM data, showed that KOH directly provides selective sites for NOx. 4. Conclusions The basic characteristic of OH- provided on the surface, due to KOH, was a major factor of increasing adsorptivity of NOx. The adsorptivity of K-IAC can be explained according to the concentration of NOx, NO2 goes through selective adsorption site of K-IAC to catalytic conversion to NO and consequently loses its function as a selective adsorption site where the amount of NO2 that is not adsorbed goes out as it is increases. According to the relative ratio of NO and NO2 produced, the lower the concentration is during the 300 min of reaction time, the larger the parts of the production ratio NO occupies, with an increase in the removal efficiency. At high concentration, the rate of adsorption failed to offset the increasing amount of NOx to be eliminated. Therefore, because the concentration is higher, the rate of adsorption increases but the function of NOx selective adsorption sites consisting of surface basic OH- by KOH impregnation is diminished more rapidly, which results in rather decreased removal efficiency. From the result of AES/SAM analysis on the sample that reached saturation (C/C0 ) 1) by NOx adsorption, we found that there is a close relationship between the intensity of K and O and the existence of N, and the reason for this was due to the degree of formation of oxide crystals made in the form of KNOx, which showed differences in intensity. As the concentration of adsorbed NOx increased, the intensity of the major component of activated carbon, C, did not change but K, N, and O showed a gradual decrease. Thus, we could confirm that the substance created on the surface was the formation of KNOx by chemical reaction. This reflects a strong interaction between K and NOx and, therefore, it is understood that the form of surface adsorption occupies the adsorption site by forming crystals of either KNO2 or KNO3, which are ionic chemical compounds, due to potassium that acts as a catalyst. Such a result can be more soundly proved

because it matches with the result of analysis on the surface distribution of NO2- and NO3- by ToF-SIMS. For impregnated potassium through NOx adsorption according to the depth profile by AES, it mostly shows a fixed level of decrease, but for N and O, it was confirmed that they decrease in an irregular pattern according to the depth of the surface. Here, the reason for the decrease can be explained by the SIMS depth profile, which can be regarded as a phenomenon of NOx adsorption progressing the more rapid diffusion of NO2 on the external surface, which led to KNO3 crystal growth that blocks the pores. In addition, because of such a phenomenon, the NO2- species that already were diffused into the inside of blocked pores by KNO3 remain which exist in a status that cannot be progressed to the termination reaction, complete oxidation, because NO2species no longer show progress in the reaction. The conclusion of AES/SAM and ToF-SIMS, with mutually complementary interpretation, supported the surface chemistry distribution of adsorbed NO2- or NO3- species and OH- that act as the selective adsorption site due to NOx adsorption. In conclusion, the surface basic OH- ion acted as a delay of oxidation from KNO2 to KNO3, resulting in an increase in adsorptivity. In addition, the study on the surface characteristics based on NOx adsorption well explained the changes in the chemical properties of the surface due to selective adsorption behavior as NOx adsorption and strong chemical interaction between potassium and NOx. Last, the KNO3 newly formed on the surface from the reaction is environmentally harmless, and the fact that it can be recycled through proper regeneration for various industrial uses such as principal substances in gunpowder, fireworks, glass, medicine, or fertilizers gathers more attention for future studies. Nomenclature C ) effluent gas concentration [ppm] C0 ) influent gas concentration [ppm] NO2 adsorbed K-IAC (run 1) ) K-IAC that reached C/C0 ) 0.814 in adsorption of 5 h under 153 ppm NO2/air NO2 adsorbed K-IAC (run 2) ) K-IAC that reached C/C0 ) 0.935 in adsorption of 5 h under 400 ppm NO2/air NO2 adsorbed K-IAC (run 3) ) K-IAC that reached C/C0 ) 1 in adsorption of 5 h under 1014 ppm NO2/air

Literature Cited (1) Mochida, I.; Kawabuchi, Y.; Kawano, S.; Matsumura, Y.; Yoshikawa, M. High Catalytic Activity of Pitch-Based Activated Carbon Fibres of Moderate Surface Area for Oxidation of NO to NO2 at Room Temperature. Fuel 1997, 76 (6), 543-548. (2) Radojevic, M. Reduction of Nitrogen Oxides in Flue Gases. Environ. Pollut. 1998, 102 (S1), 685-689. (3) Pereira, C. J. Environmentally Friendly Processes. Chem. Eng. Sci. 1999, 54, 1959-1973. (4) Aarna, I.; Suuberg, E. The Role of Carbon Monoxide in the NO-Carbon Reaction. Energy Fuels 1999, 13, 1145-1153. (5) Tsuji, K.; Shiraishi, I. Combined Desulfurization, Denitrification and Reduction of Air Toxics using Activated Coke. Fuel 1997, 76 (6), 549-553. (6) Yun, J. H.; Choi, D. K.; Kim, S. H. Adsorption Equilibria of Chlorinated Organic Solvents onto Activated Carbon. Ind. Eng. Chem. Res. 1998, 37 (4), 1422-1427. (7) Yun, J. H.; Choi, D. K.; Kim, S. H. Equilibria and Dynamics for Mixded Vapors of BTX in an Activated Carbon Bed. AIChE J. 1999, 45 (4), 751-760.

Ind. Eng. Chem. Res., Vol. 40, No. 15, 2001 3345 (8) Munˇiz, J.; Herrero, J. E.; Fuertes, A. B. Treatments to Enhance the SO2 Capture by Activated Carbon Fibres. Appl. Catal. B 2000, 18, 171-179. (9) Tsuchiai, H.; Ishizuka, T.; Nakamura, H.; Ueno, T.; Hattori, H. Removal of Sulfur Dioxide from Flue Gas By the Absorbent Prepared form Coal Ash: Effects of Nitrogen Oxide and Water Vapor. Ind. Eng. Chem. Res. 1996, 35, 851-855. (10) Illa´n-Gome´z, M. J.; Linares-Solano, A.; Radovic, L. R.; Salinas-Martı´nez de Lecea, C. NO Reduction by Activated Carbons. 2. Catalytic Effet of Potassium. Energy Fuels 1995, 9, 97103. (11) Zhu, Z.; Liu, Z.; Liu, S.; Niu, H. Adsorption and Reduction of NO over Activated Coke at Low Temperature. Fuel 2000, 79, 651-658. (12) Garcı´a-Garcı´a, A.; Illa´n-Gome´z, M. J.; Linares-Solano, A.; Salinas-Martı´nez de Lecea, C. Potassium-containing Briquetted Coal for the Reduction of NO. Fuel 1997, 76, 499-505. (13) Guo, J.; Chong, A. C. Effect of Surface Chemistry on GasPhase Adsorption by Activated Carbon Prepared from Oil-Palm Stone with Pre-Impregnation. Sep. Purif. Technol. 2000, 18, 4755. (14) Chambrion, P.; Suzuki, T.; Zhang, Z. G.; Kyotani, T.; Tomita, A. XPS of Nitrogen-Containing Functional Groups Formed during the C-NO Reaction. Fuel 1997, 11, 681-685. (15) Spevack, P.; Deslandes, Y. ToF-SIMS Analysis of Adsorbate/ Membrane Interactions. Appl. Surf. Sci. 1996, 99, 41-50. (16) Lamontagne, B.; Semond, F.; Adnot, A.; Roy, D. SIMS Investigation of the Si(111) Oxidation Promoted by Potassium Overlayers. Appl. Surf. Sci. 1995, 90, 447-454. (17) Rivie´re, J. C. Surface Analytical Techniques; Clarendon: Oxford, U.K., 1990. (18) Vickerman, J. C. Surface AnalysissThe Principal Techniques; John Wiley & Sons: New York, 1999.

(19) Hu, Z.; Srinivasan, M. P. Preparation of High-Surface-Area Activated Carbons from Coconut Shell. Micropor. Mesopor. Mater. 1999, 27, 11-18. (20) Illa´n-Gome´z, M. J.; Raymundo-Pinˇero, E.; Garcı´a-Garcı´a, A.; Brnada´n, S.; Linares-Solano, A.; Salinas-Martı´nez de Lecea, C. Catalytic NOx Reduction by Carbon Supporting Metals. Appl. Catal. B 1999, 20, 267-275. (21) Lee, M. R.; Wolan, J. T.; Hoflund, G. B. Adsorption of NOx on Alumina Treated with Alkali Carbonates and Alkali Hydroxides. Ind. Eng. Chem. Res. 1999, 38, 3911-3916. (22) Neathery, J. K.; Rubel, A. M.; Stencel, J. M. Uptake of NOx by Activated Carbons; Bench-Scale and Pilot-Plant Testing. Carbon 1997, 35 (9), 1321-1327. (23) Kong, Y.; Cha, C. Y. NOx Adsorption on Char in the Presence of Oxygen and Moisture. Carbon 1996, 34, 1027-1033. (24) Yang, J.; Mestl, G.; Herein, D.; Schl gl, R.; Find, J. Reaction of NO with Carbonaceous Materials. 2. Effect of Oxygen on the Reaction of NO with Ashless Carbon Black. Carbon 2000, 38, 729740. (25) Illa´n-Gome´z, M. J.; Linares-Solano, A.; Radovic L. R.; Salinas-Martı´nez de Lecea, C. NO Reduction by Activated Carbons. 7. Some Mechanistic Aspects of Uncatalyzed and Catalyzed Reaction. Energy Fuels 1996, 10, 158-168. (26) Darmstadt, H.; Pantea, D.; Smmchen, L.; Roland, U.; Kaliaguine, S.; Roy, C. Surface and Bulk Chemistry of Charcoal Obtained by Vacuum Pyrolysis of Bark: Influence of Feedstock Moisture Content. J. Anal. Appl. Pyrolysis 2000, 53, 1-17.

Received for review December 11, 2000 Revised manuscript received April 5, 2001 Accepted May 8, 2001 IE001068Q