Environ. Sci. Technol. 2005, 39, 5844-5850
Fundamental Adsorption Characteristics of Carbonaceous Adsorbents for 1,2,3,4-Tetrachlorobenzene in a Model Gas of an Incineration Plant KENICHIRO INOUE AND KATSUYA KAWAMOTO* National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba, Ibaraki 305-8506, Japan
Carbonaceous adsorbents such as activated carbon have been used to reduce the emission of organic pollutants from incineration plants. However, with this method, the amount and type of adsorbent to be used are based only on empirical results, which may lead to overuse of the adsorbents. The fundamental adsorption characteristics of several kinds of activated carbon, activated coke, and carbide wood were examined using 1,2,3,4-tetrachlorobenzene as an adsorbate. The removal performance and various equilibrium adsorption characteristics of these adsorbents were analyzed using laboratory-scale adsorption equipment. The equilibrium adsorption amount increased by a factor of 1.9-3.2 at 150 °C compared with that at 190 °C. The effect of the moisture content on adsorption capacity was relatively small in comparison with that of the temperature. The micropore volume for pore diameters of 2 nm or less was the most important factor governing the adsorption capacity for all adsorbents. Activated carbon showed superior adsorption ability compared to activated coke and carbide wood, although all adsorbents were sufficient for practical use.
Introduction Exhaust gas from municipal solid waste (MSW) incinerators, of which a large amount is generated in many countries, can have an enormous environmental influence. The flue gas contains not only conventional inorganic pollutants such as NOX, but also organic micropollutants such as dioxins and chlorinated compounds. The concentration of organic pollutants could be reduced considerably by exhaust gas treatment processes or so-called end-of-pipe technologies such as bag filters for dust collection, removal of acidic gases through lime addition, or catalyst reactors (1-3). In a report by Kawamoto (4), when the concentration of chlorobenzenes at the inlet of a bag filter was 8.7 µg m-3N, the outlet concentration was reduced to 5.3 µg m-3N by keeping the temperature at 150 °C. However, more reliable technologies are required for the constant reduction of organic pollutants as bag filters cannot handle fluctuations in flue gas composition and/or concentration. Activated carbon is usually employed to control organic pollutants in the exhaust gas treatment process (5-8) because it is highly efficient and reliable. This is typically done by * Corresponding author phone: +81-29-850-2958; fax: +8129-850-2091; e-mail:
[email protected]. 5844
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
spraying powdered adsorbents into the ducts because the method can be easily applied to existing flue gas treatment facilities. Kim et al. (9) evaluated the dioxin removal efficiency in commercial-scale MSW incinerators with a capacity of more than 200 t d-1. The removal efficiency using a spray dryer absorber/bag filter system was 99% when a mixture of lime and activated carbon was used. The amount of mixture had to be minimized since this system produced relatively large amounts of fly ash. Chang et al. (10, 11) evaluated the dioxin removal efficiency at an MSW incinerator over a period of two years. The removal efficiency increased to 96.6% by using activated carbon at a rate of 115 kg d-1 during the first year, and reached 98.7% during the second year. The lower efficiency in the first year can be attributed to the memory effect: the dioxins, or precursors, desorbed slowly to bulk flue gas from the deposits and the dioxin concentration increased in the stack emissions. It is therefore important to consider this effect at the starting stage of activated carbon injection. These results have demonstrated the benefits of this adsorption method. However, the method may have disadvantages related to high temperature and moisture content in terms of the incineration flue gas. The influences of these conditions on the adsorption characteristics must be elucidated. Furthermore, the relationship between the physical properties and adsorption capacity of the adsorbents has not yet been clarified. The amount and type of adsorbents used are selected empirically at actual plants, which may lead to possible overuse of the adsorbents. If adsorbents are injected at an amount of 50 mg m-3N into flue gas with a flow rate of 4000 m3N h-1 (tons of solid waste)-1, the required dosage would be 40 t y-1 for a 200 000 t y-1 MSW incinerator (12). As these operating costs are extremely high, it is desirable to reduce the amount of adsorbent used and to choose superior, more cost-efficient adsorbents. Activated coke is known as an inexpensive adsorbent, and its use at facilities is increasing (13-17). The adsorbent is produced at a lower environmental load and cost than activated carbon because the degree of activation is considerably lower. The adsorption capacity of activated coke is generally smaller than that of activated carbon, although these details are still not clear. In addition, the carbides of raw wood materials are an attractive possibility as they use recycled ligneous waste. The objective of this study was to optimize adsorption treatment technology for organochlorine compounds in incineration exhaust gas using several kinds of activated carbon, activated coke, and carbide wood. The fundamental characteristics of the various adsorbents were evaluated in a laboratory-scale test system using 1,2,3,4-tetrachlorobenzene as an adsorbate. This compound had appropriate physical properties for generating the constant low-concentration gas needed for these experiments using the diffusion tube method described later. In a previous experiment, we attempted to generate the gas using 1,3,6,8tetrachlorodibenzo-p-dioxin. However, the vapor pressure was too low and it was difficult to achieve continuous gas generation. For these reasons, in this study we chose 1,2,3,4tetrachlorobenzene, which is a chlorinated aromatic compound with semi-volatility and hydrophobicity, as an adsorbate. This study examined the relationship between the physical properties and adsorption capacity of adsorbents, and the effects of the exhaust gas temperature, coexistence of moisture, and adsorbate concentration on the adsorption capacity. 10.1021/es0489745 CCC: $30.25
2005 American Chemical Society Published on Web 06/22/2005
TABLE 1. Physicochemical Properties of 1,2,3,4-Tetrachlorobenzene molecular weight melting point (°C) boiling point (°C) vapor pressure at 25 °C (Pa) 1-octanol/water partition coefficient
216 47.5 254 4.76 4.60
Materials and Methods Adsorbent. The adsorbents employed in this study consisted of five kinds of activated carbon made from palm shell and coal, two kinds of activated coke made from coal, and one kind of carbide wood. The activated carbons and cokes used here are currently available on the market for flue gas treatment. The carbide wood examined in this study is used as a soil enrichment product. All of the adsorbents were in the form of a fine powder and were used with an inert carrier media (granular calcium carbonate with a mean particle diameter of 2.4 mm) in the experiment. The inert carrier media was covered with the adsorbents at a weight ratio of 0.2% (w/w) prior to being packed into a glass column. The inner diameter of the glass column was 16 mm and the apparent packed volume was 9.8 cm3. The bulk density of the column was 1.3 g cm-3 and the absolute density was 2.7 g cm-3. To clarify the fundamental physical properties of the adsorbents, the mean particle diameter, specific surface area, and pore volume were measured. Adsorbate. In this study, 1,2,3,4-tetrachlorobenzene was used as an adsorbate. The physicochemical properties of this compound are shown in Table 1 (18, 19). The chlorobenzenes of concern in the exhausted gas of actual incineration plants are usually present in the order of 10 µg m-3 or less at the inlet of the bag filter (4, 20). However, at this concentration level it is difficult to supply gas at constant concentration for semi- and nonvolatile organic compounds. In this study, gaseous 1,2,3,4-tetrachlorobenzene was generated at a constant concentration using a Permeater PD-1B (Gastec Corp., Ayase, Japan), which is normally used to generate gaseous organic vapors. A diffusion tube that supplied the 1,2,3,4-tetrachlorobenzene in an open-type glass tube was used as the gas generation source in this experiment. The diffusion tube was installed in the constant temperature oven in the Permeater where the maximum temperature was 50 °C. The amount of 1,2,3,4-tetrachlorobenzene volatized was adjusted by controlling the temperature of the oven. The generated vapor was diluted with a carrier gas to produce the model gas. The concentration of 1,2,3,4-tetrachlorobenzene in the model gas was determined by actual measurement as the compound is a solid at room temperature and there is no data on the volatilization rate. The actual measured values for the concentration of 1,2,3,4-tetrachlorobenzene in the model gas were 70, 110, and 240 µg m-3 when the temperature of the diffusion tube was set at 35, 40, and 50 °C, respectively. Test Equipment. Figure 1 illustrates the adsorption test system including the equipment for generating organic vapor. The 1,2,3,4-tetrachlorobenzene vapor was diluted with purified air in the equipment, and then further mixed with carbon dioxide. Steam was also added to the gas mixture by injecting water into the front of the oven. There have been few reports of examples that included moisture in a laboratory-scale test system. As shown in Figure 1, the model gas flowed continuously into the adsorbent-packed column placed in the oven. It is difficult to continuously measure the adsorbate concentration in this column outlet gas because the concentration is at an ultratrace level. Therefore, a fixed amount of outlet gas was collected in n-hexane in a fixed period. Then, the concentra-
FIGURE 1. Schematic diagram of the adsorption test system. tion of this solution was determined quantitatively and hence the adsorbate concentration in the outlet gas was obtained. The outlet gas then passed into the water, and the moisture in the gas was trapped. Measurement of the concentration confirmed that the 1,2,3,4-tetrachlorobenzene did not dissolve in this trap water. Next, the adsorbate was collected by passing the outlet gas, with the moisture removed, to n-hexane. An intermittent automatic sampling device was used to collect the gas sample. Multiple small impingers filled with n-hexane were connected to this device as shown in Figure 1. Pump suction prompted the outlet gas to pass through one of these multiple impingers at a constant time and constant flow velocity. The suction was switched to a different impinger at a constant time interval by using an automatic 16-position pneumatic valve controlled by a timing program. In these experiments, the collection interval for the effluent gas was 120, 180, or 240 min, and the collection was carried out for up to 60 min. Gas collection was performed in the water tank, which was cooled to about 4 °C in order to prevent volatilization and to collect moisture and 1,2,3,4tetrachlorobenzene more effectively. Experimental Conditions. Various experimental conditions were used, taking into consideration the range of differences in the actual flue gas treatment. The space velocity, defined as the ratio of flow rate of model gas to packed volume of carrier medium, was fixed at 5000 h-1 for all experiments considering the breakthrough time. Usually, the gas discharged from the incinerator is rapidly cooled to 200 °C or less in order to restrain dioxin generation (21-23). Especially, it is lowered to about 150 °C to maintain a high efficiency of dioxin separation by the bag filter (1, 24, 25). For these reasons, the temperature of the column was adjusted to between 150 and 190 °C by changing the oven temperature. The concentration of 1,2,3,4-tetrachlorobenzene supplied was changed from 70 to 240 µg m-3 as mentioned above. The content of carbon dioxide, which exists at several % in exhaust gas (26-28), was fixed at 8% (v/v). The moisture is also included in the exhaust gas (26-28). In the incinerator using water spray quenching as a cooling method for exhaust gas, there are cases in which the moisture content is about 40% (29). Therefore, model gas conditions regarding moisture were 0, 20, or 40% (v/v). The total flow rate of model gas passing into the adsorbent-packed column was 815 mL min-1. Seven combinations of these conditions for each experiment are shown in Table 2. The conditions in each experiment described later are shown by the number of the condition in Table 2. Analytical Methods. The fundamental physical properties of the adsorbents were measured in order to examine their adsorption characteristics. The mean particle diameter was measured by laser diffraction method (30, 31). The specific surface area was obtained using the BET method (32) from VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5845
TABLE 2. Experimental Conditions number of condition temperature of column (°C) components of model gas (% (v/v))a air moisture concentration of 1,2,3,4-tetrachlorobenzene (µg m-3) a
1 170
2 150
3 190
4 170
5 170
6 170
7 170
52 40 240
52 40 240
52 40 240
92 0 240
72 20 240
52 40 70
52 40 110
The total flow rate of the model gas was 815 mL min-1, and the carbon dioxide content was fixed at 8% (v/v).
TABLE 3. Analytical Conditions of Gas Chromatography/Mass Spectrometry instrument column
Saturn 2200, Varian Technologies Japan Ltd. CP5860, Varian Technologies Japan Ltd. 0.25 mm diameter, 30 m lengths, 0.25 µm thickness 40 °C (3 min) - 20 °C min-1 to 155 °C (1.5 min) - 30 °C min-1 to 280 °C (5 min) 250 °C 200 °C 70 eV 50:1 1 µL n-hexane helium
column oven temperature injection temperature ion source temperature ionization voltage split ratio injection amount solvent carrier gas
TABLE 4. Fundamental Physical Properties of the Employed Adsorbents adsorbent
A activated carbon palm shell 9.24 1213
raw material mean particle diameter (µm) specific surface area (m2 g-1) pore volume (mL g-1) pore diameter < 2 nm (micropore) 0.46 2-60 nm (mesopore) 0.10 60 nm - 200 µm (macropore) 1.20
B activated carbon palm shell 18.7 1097
C activated carbon anthracite 29.6 833
D activated carbon peat 21.0 590
E activated carbon peat 15.6 199
F activated coke lignite 23.5 285
G activated coke brown coal 15.0 383
wood 32.2 422
0.40 0.09 0.76
0.31 0.19 0.83
0.19 0.25 1.07
0.05 0.21 1.06
0.09 0.14 0.79
0.13 0.07 0.74
0.18 0.04 0.73
nitrogen adsorption isotherms. The pore volumes for pore diameters of up to 2 nm and 2-60 nm were calculated by the MP (33) and BJH (34) methods, respectively, from nitrogen adsorption and desorption isotherms. The pore volume for pore diameters of 60 nm - 200 µm was measured by mercury intrusion porosimetry (35, 36). These measurements were carried out based on well-accepted methods such as the Japanese Industrial Standards (JIS) (37-39). The collection and analysis of 1,2,3,4-tetrachlorobenzene were performed as follows. The gas was passed to the impinger filled with about 10 mL of n-hexane for 60 min at a flow rate of 250 mL min-1; therefore, 15 L of gas in total was collected. It was confirmed that the collection efficiency by this n-hexane trap was 100%. About 4 mL of n-hexane volatilized during this collection, so a constant volume was attained at 10 mL by adding n-hexane. The concentration of 1,2,3,4-tetrachlorobenzene in n-hexane solution was determined quantitatively by gas chromatography/mass spectrometry, the conditions of which are shown in Table 3. Quality criteria were based on the application of quality control and quality assurance measures such as analysis of a blank sample covering the complete analytical procedure. As additional performance checks, the mass pattern was calibrated using the standard substance for mass proofreading and equipment sensitivity was checked. The concentration of 1,2,3,4-tetrachlorobenzene included in the gas was calculated from this solution concentration, and the adsorption breakthrough curves were plotted, showing the relationship of gas supply time against C/C0 ratio (outlet gas concentration C to inlet gas concentration C0). An equilibrium adsorption amount was calculated by integrating the breakthrough curve obtained. The equilibrium adsorption char5846
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
H carbide
acteristics of various adsorbents were examined by this method.
Results Physical Properties. The fundamental physical properties of five kinds of activated carbon (A to E), two kinds of activated coke (F and G), and one kind of carbide wood (H) employed in this experiment are shown in Table 4. The mean particle diameter of the adsorbents ranged from 9 to 32 µm. The specific surface area of activated carbons A to D ranged from 590 to 1213 m2 g-1. These values were about 2-4 times larger than those of activated cokes F and G, which were 285 and 383 m2 g-1, respectively. This is the result of a higher degree of activation of the carbons in comparison with the activated cokes. However, the value for activated carbon E was only 199 m2 g-1, which is smaller than the value for both the activated cokes and carbide wood H (422 m2 g-1). The pores were divided into micropores (pore diameter: 2 nm or less), mesopores (2-60 nm), and macropores (60 nm - 200 µm) according to the classification by IUPAC (40). It should be noted that the boundary between meso- and macropore was changed from 50 to 60 nm for experimental reasons. Figure 2 shows the relationship between the specific surface area and each micro-, meso-, and macropore volume for the eight adsorbents. These results clarified that the specific surface area of these adsorbents was proportional to the micropore volume and independent of the meso- and macropore volume. This relationship can be expressed by:
S ) 2.7 × 103 vmicropore
(1)
where S is the specific surface area (m2 g-1) and vmicropore is
FIGURE 2. Relationship between pore volumes for pore diameters of up to 2 nm (b), 2-60 nm (0), and 60 nm - 200 µm (2) and the specific surface area of activated carbons, activated cokes, and carbide wood.
FIGURE 3. Breakthrough curves of activated carbon A, activated coke F, and carbide wood H for 1,2,3,4-tetrachlorobenzene (Condition 1).
FIGURE 4. Breakthrough curves of activated carbon A and activated coke F at various gas temperatures (Conditions 1, 2, 3). the micropore volume (mL g-1). The correlation coefficient for this relationship was 0.98. Adsorption Breakthrough Curves. Figure 3 shows the adsorption breakthrough curves for 1,2,3,4-tetrachlorobenzene under Condition 1 as shown in Table 2. The ratio of C/C0 increased with the experimental time until the adsorption saturation point was reached. However, differences can be seen in the breakthrough curves among the eight kinds of adsorbents. The adsorption capacity appears fairly high for the adsorbents with high specific surface areas. Influence of Gas Temperature. Adsorption breakthrough curves at 150-190 °C are shown in Figure 4 (Conditions 1, 2, 3). These results indicate that the time to adsorption saturation was reduced by an increase in temperature. In activated coke F, the curves rise rapidly, whereas the curves of activated carbon A are more gradual, with a maximum of around 30 h at 150 °C. Equilibrium adsorption amounts were obtained by integrating these breakthrough curves. The dependence of the equilibrium adsorption amount on gas temperature is shown in Figure 5. The figure indicates that the equilibrium adsorption amount rose exponentially according to the decrease in gas temperature. These relationships can be expressed by:
log q ) -at + b
(2)
FIGURE 5. Relationship between gas temperature and equilibrium adsorption amount.
FIGURE 6. Breakthrough curves of activated carbon A and activated coke F at various moisture contents (Conditions 1, 4, 5).
FIGURE 7. Breakthrough curves of activated carbon A and activated coke F at various 1,2,3,4-tetrachlorobenzene supply concentrations (Conditions 1, 6, 7).
TABLE 5. a and b Values of Eq 2 adsorbent A D E F
a
b 10-2
1.3 × 1.2 × 10-2 7.0 × 10-3 9.7 × 10-3
2.3 × 103 1.0 × 103 6.5 × 101 1.3 × 102
where q is the equilibrium adsorption amount (mg g-1) and t is the gas temperature ( °C). The values of coefficients a and b in eq 2 are listed in Table 5. These relationships show that the equilibrium adsorption amounts increased 1.9-3.2 times between 150 °C and 190 °C. Influence of Moisture Content. Adsorption breakthrough curves at a moisture content of 0-40% (v/v) were measured for activated carbon A and activated coke F and are shown in Figure 6 (Conditions 1, 4, 5). Large differences were not observed at 170 °C, indicating that the influence of moisture content on the adsorption capacity of both adsorbents was relatively small in comparison with the effect of temperature. Influence of Supply Concentration. The dependence of the adsorption capacity on the 1,2,3,4-tetrachlorobenzene supply concentration was examined for activated carbon A and activated coke F, as shown in Figure 7 (Conditions 1, 6, 7). It was confirmed that the time to adsorption saturation VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5847
molecular weight compounds in the vapor phase. As for the activated carbons made from coal (C, D, and E), the mesopore volume is larger than for other adsorbents. This type of activated carbon has been successfully used for adsorption in the liquid phase, and effectively adsorbs solutes over a wide range of molecular sizes. In the case of the macropores, the pore volumes are larger than the other pores for all of the adsorbents. Furthermore, no relationship was observed between the meso- and macropore volume and the specific surface area (Figure 2). FIGURE 8. Adsorption isotherms of various adsorbents for 1,2,3,4tetrachlorobenzene (Conditions 1, 6, 7).
TABLE 6. k and n Values of Freundlich’s Formula adsorbent
k
n
A B C D E F G H
2.1 1.7 1.2 1.5 0.026 0.12 0.076 0.88
3.0 2.8 2.4 3.0 1.1 1.7 1.5 2.7
was reduced with an increase in the supply concentration. As for activated carbon A, the ratio of C/C0 reached saturation in around 80 h at a supply concentration of 70 µg m-3, whereas it took only around 40 h at 240 µg m-3. The difference was smaller in activated coke F. Similar experiments (Conditions 1, 6, 7) were conducted for the other six kinds of adsorbent, and adsorption isotherms were drawn. Adsorption Isotherms. Adsorption isotherms were obtained by plotting the relationship between the adsorbate equilibrium concentration and the equilibrium adsorption amount of each adsorbent as shown in Figure 8. The equilibrium adsorption amount increased with an increase in the equilibrium concentration. However, differences can be seen based on the kind of adsorbent. Adsorption isotherms are nearly straight lines on a logarithmic scale for all of the adsorbents and are accurately expressed as the adsorption isotherm of Freundlich’s formula:
q ) kC1/n
(3)
where q is the equilibrium adsorption amount (mg g-1) and C is the equilibrium concentration of 1,2,3,4-tetrachlorobenzene (µg m-3). Although this formula has been proposed for liquid-phase adsorption, it has also been recognized as applicable to gas-phase adsorption systems at low concentrations. The values of coefficients k and n in the formula are shown in Table 6. The k value ranges from 0.026 to 2.1 and the n value ranges from 1.1 to 3.0. The k value here is the equilibrium adsorption amount at an equilibrium concentration of 1 µg m-3.
Discussion In general, the pore diameter on the adsorbent surface is determined by the degree of activation and the surface structure of the raw material. The micropore volume increases as the activation of carbonized raw material progresses, because areas of amorphous carbon on the surface are predominantly oxidized and micropores are formed at the areas (41). The volumes of micro- and mesopores for each adsorbent are compared in Table 4. The micropores of activated carbons A and B that are made from palm shell are larger than for other adsorbents. These materials are suitable and have been mainly used for the adsorption of low 5848
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
In the comparative breakthrough curves of Figure 3, apparent differences in the equilibrium adsorption amounts can be seen according to the kind of adsorbent, which may be related to the composition of the pore volumes. As shown in Table 4, these adsorbents have different pore distributions. Mesopores develop in activated carbons D and E. In activated carbons A and B with comparatively large adsorption capacity, although micropores develop, the mesopore volume is one-quarter or less of the volume. It appears that the adsorption capacity is dependent to some degree on the micropore volume and independent of the mesopore volume. As shown in Figure 4, the time to adsorption saturation was shortened by increasing the gas temperature. Given that adsorption phenomena are exothermic, this result can be explained by a temperature-dependent thermodynamic relationship. The fact that the vapor pressure of 1,2,3,4tetrachlorobenzene at 150 °C decreased to 5.6 kPa from 21 kPa at 190 °C (42) might have an additional influence on this temperature effect. The relationship between the gas temperature and the logarithm of the equilibrium adsorption amount is linear (Figure 5). As the value of a is larger in eq 2, the degree of decrease in the equilibrium adsorption amount with the rise in temperature is larger. At 150 °C, just 40 °C lower than 190 °C, the equilibrium adsorption amount increased 3.2 times in activated carbon A and 2.4 times in activated coke F. These findings suggest that it is desirable to operate at temperatures as low as possible to maximize the effect of the adsorbent. At 170 °C, the influence of the moisture content was modest (Figure 6). Kojima et al. (43) obtained the adsorption isotherms of water vapor on activated carbon through calculations and showed that moisture did not adsorb onto the surface at 170 °C based on the isotherm. Wey et al. (44) reported that moisture in the flue gas condensed on the activated carbon surface and occupied adsorption sites at 120 °C. They observed that the adsorption capacity of organic compounds such as phenylethyne and benzaldehyde decreased at 120 °C whereas there was a higher adsorption capacity at 150-180 °C. Therefore, moisture is not believed to be present on the adsorbent surface at 150-190 °C, the temperature range of actual exhaust gas. From this information and the results obtained in this study, the moisture content appears to have little effect and so it is not necessary to control the moisture content in flue gas. The parameter n is larger than 1 in all adsorbents, and a favorable adsorption isotherm is shown (Table 6). However, a large difference was observed in this value between adsorbents A to D and H, all having a specific surface area exceeding 422 m2 g-1, and adsorbents E to G, having an area of 383 m2 g-1 or less. In adsorbents E to G, n values were approximately 1. This means that the equilibrium adsorption amount is simply proportional to the equilibrium concentration, and that the relationship can be expressed by the Henry-type isotherm (q ) kC) in adsorbents that have lower equilibrium adsorption amounts. In adsorbents A to D and H, the value of n is larger than that of other adsorbents. Therefore, it is suggested that the degree of increase in the equilibrium adsorption amount is not large, even if the equilibrium concentration is raised further.
FIGURE 9. Relationships between pore volumes for pore diameters of up to 2 nm (a), 2-60 nm (b), and 60 nm - 200 µm (c) and equilibrium adsorption amount (O: activated carbon, b: activated coke, 0: carbide wood).
FIGURE 10. Relationship between specific surface area and equilibrium adsorption amount (O: activated carbon, b: activated coke, 0: carbide wood). Activated carbons are usually sprayed into the exhaust gas flow at amounts of around 50-200 mg m-3N at incineration facilities (6-8, 11, 12), with the average amount of 100 mg m-3N. The total concentration of organic pollutants in exhaust gas has been estimated to be approximately 100 µg m-3N (26). The adsorption amount can thus be simply calculated to be 1 mg g-1. The results of the adsorption isotherm demonstrated that the equilibrium adsorption amounts of activated cokes F and G and carbide wood H are 1.8, 1.7, and 4.7 mg g-1, respectively, at an equilibrium concentration of 100 µg m-3. These equilibrium adsorption amounts are larger than the estimated adsorption amount. Therefore, these results suggest that activated coke and carbide wood may be effective adsorbents for the reduction of trace organic pollutants. In addition, the price of activated carbon A is about three times that of activated coke F, so activated coke may be a good adsorbent in practice. The relationship between the physical characteristics of the surface pore structure and adsorption potential was examined. The pores of each adsorbent were divided into micro-, meso-, and macropores (described earlier), and the pore volume for each kind of pore was obtained. The value of the equilibrium adsorption amount at an equilibrium concentration of 100 µg m-3 was used as an indicator of the adsorption capacity. This might be the overall value of the concentrations of organic pollutants in actual exhaust gas (26). Figures 9(a)-(c) show the relationships between each pore volume and the equilibrium adsorption amount. The results shown in Figure 9(a) suggest that micropore volume is the most important factor governing the adsorption capacity for all adsorbents. The adsorption capacity hardly depended on the meso- and macropore volume. The specific surface area of these adsorbents was proportional to the micropore volume as given by eq 1. These findings suggest that the specific surface area increases as the micropore volume becomes larger, increasing the equilibrium adsorption amount. This is supported by Figure 10, which shows the relationship between the specific surface area and the equilibrium adsorption amount. It is concluded that the micropore volume for pore diameters of 2 nm or less is the most important factor
governing the adsorption capacity for all adsorbents. Hence, micropore volume should be properly controlled through effective activation methods and raw material selection for efficient removal of pollutants. If an adsorbent abundant in such micropores is employed in packed towers, the life of the adsorbent could be prolonged. Furubayashi et al. (45) reported on the dioxin removal efficiency of a bag filter with activated carbon injected into the flue gas through quantitative estimation using a mass transfer model and linear isotherm system. They suggested that the dioxins are adsorbed on activated carbon, not when the activated carbon reaches the bag filter after injection into the flue, but while the exhaust gas is passing through the powder layer in the bag filter. The powder layer in the bag filter can thus be regarded as a fixed-bed adsorption layer. The application could be optimized by adjusting the amount and time of sediment on the filter, and the injected amount of adsorbent, based on the equilibrium adsorption and breakthrough characteristics obtained in this study.
Literature Cited (1) Watanabe, T.; Asai, M.; Kondo, T.; Shimizu, M.; Takeuchi, Y.; Aramaki, H.; Naito, M. Chemosphere 1996, 32, 177-187. (2) Horaguchi, M.; Ogawa, H.; Ose, K.; Tomisawa, S.; Matsuura, S. Chemosphere 1989, 18, 1785-1797. (3) Fujii, T.; Murakawa, T.; Maeda, N.; Kondo, M.; Nagai, K.; Hama, T.; Ota, K. Chemosphere 1994, 29, 2067-2070. (4) Kawamoto, K. J. Japan Soc. Air Pollut. 1993, 28, 266-278 (in Japanese). (5) Siret, B.; Gilman, K. Int. Conf. Inciner. Therm. Treat. Technol. 1996 1996, 583-586. (6) Sierhuis, W. M.; de Vries, C.; Born, J. G. P. Chemosphere 1996, 32, 159-168. (7) Abad, E.; Adrados, M. A.; Caixach, J.; Rivera, J. Environ. Sci. Technol. 2002, 36, 92-99. (8) Abad, E.; Caixach, J.; Rivera, J. Chemosphere 2003, 50, 11751182. (9) Kim, S.-C.; Jeon, S. H.; Jung, I.-R.; Kim, K.-H.; Kwon, M.-H.; Kim, J.-H.; Yi, J.-H.; Kim, S.-J.; You, J.-C.; Jung, D.-H. Chemosphere 2001, 43, 773-776. (10) Chang, M. B.; Lin, J. J. Chemosphere 2001, 45, 1151-1157. (11) Chang, M. B.; Lin, J.-J.; Chang, S.-H. Atmos. Environ. 2002, 36, 279-286. (12) Everaert, K.; Baeyens, J.; Degre`ve, J. Environ. Sci. Technol. 2003, 37, 1219-1224. (13) Kim, K.-T.; Koh, D. J.; Kim, J. H. Organohalogen Compd. 2002, 56, 397-400. (14) Kim, K.-T.; Koh, D.-J.; Kim, J.-H. Organohalogen Compd. 2003, 63, 228-231. (15) Kassebohm, B.; Streng, S. Proc. Annu. Meet. Air Waste Manag. Assoc. 1993, 93-MP-5.05, 1-14. (16) Streng, S.; Kassebohm, B. Fuel Process. Technol. 1994, 39, 431444. (17) Hartenstein, H.-U. Chemosphere 1994, 29, 2071-2081. (18) Handbook of Physical Properties of Organic Chemicals; Howard, P. H., Meylan, W. M., Eds.; CRC Press: Boca Raton, FL, 1997. (19) Mackay, D.; Shiu, W. Y.; Ma, K. C. Illustrated Handbook of Physical-Chemical Properties and Environmental Fate for Organic Chemicals, Volume I.; CRC Press: Boca Raton, FL, 1992. (20) Ide, Y.; Kashiwabara, K.; Okada, S.; Mori, T.; Hara, M. Chemosphere 1996, 32, 189-198. VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5849
(21) Prescott, W. Int. Conf. Inciner. Therm. Treat. Technol. 1996 1996, 617-619. (22) Tejima, H.; Nakagawa, I.; Shinoda, T.; Maeda, I. Chemosphere 1996, 32, 169-175. (23) Ogushi, Y.; Nose, H. Organohalogen Compd. 1994, 19, 437441. (24) Sato, Y.; Shizuma, M.; Sasaki, M.; Futamura, O. Organohalogen Compd. 1994, 19, 389-393. (25) Futamura, O.; Sasaki, M.; Satou, Y.; Tsukamoto, T. Organohalogen Compd. 1999, 40, 499-502. (26) Karademir, A.; Bakoglu, M.; Taspinar, F.; Ayberk, S. Environ. Sci. Technol. 2004, 38, 1201-1207. (27) Takeda, N.; Nakamura, K.; Fujii, T.; Tsunemi, T. Organohalogen Compd. 1994, 19, 465-470. (28) Bassetti, A.; Bodini, M.; Donega`, M.; Miglio, R.; Tirler, W.; Voto, G. Organohalogen Compd. 2001, 54, 98-101. (29) Kawamoto, K.; Kimura, T.; Sato, M. Trans. Mater. Res. Soc. Jpn. 1994, 15A, 127-130. (30) Tabibian, R. M.; Heller, W.; Epel, J. N. J. Colloid Sci. 1956, 11, 195-213. (31) Oster, G. Chem. Rev. 1948, 43, 319-365. (32) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309-319. (33) Mikhail, R. S. H.; Brunauer, S.; Bodor, E. E. J. Colloid Interface Sci. 1968, 26, 45-53. (34) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373-380. (35) Ritter, H. L.; Drake, L. C. Ind. Eng. Chem., Anal. Ed. 1945, 17, 782-786. (36) Drake, L. C. Ind. Eng. Chem. 1949, 41, 780-785.
5850
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
(37) JIS R 1629:1997; Japanese Industrial Standard, Determination of particle size distributions for fine ceramic raw powders by laser diffraction method; Japanese Standards Association: Tokyo, 1997. (38) Pieters, W. J. M.; Gates, W. E. Method and apparatus for determining the amount of gas adsorbed or desorbed from a solid; 1984; United States Patent 4489593. (39) JIS K 1150:1994; Japanese Industrial Standard, Test method for silica gel; Japanese Standards Association: Tokyo, 1994. (40) IUPAC Manual of Symbols and Terminology for Physicochemical Quantities and Units, Pure Appl. Chem. 1972, 31, 577-638. (41) Morioka, Y.; Kuroda, T.; Kosugi, H.; Ueno, A.; Kobayashi, J. NIPPON KAGAKU KAISHI (Journal of the Chemical Society of Japan, Chemistry and Industrial Chemistry) 1995, 9-12 (in Japanese). (42) Shiu, W.-Y.; Ma, K.-C. J. Phys. Chem. Ref. Data 2000, 29, 387462. (43) Kojima, T.; Okajima, S.; Ueoka, S. Proceedings of the 11th Annual Conference of The Japan Society of Waste Management Experts 2000, 657-659 (in Japanese). (44) Wey, M.-Y.; Yu, L.-J.; Jou, S.-I.; Chiang, B.-C.; Wei, M.-C. J. Environ. Eng. 1999, 125, 925-932. (45) Furubayashi, M.; Nagai, K. KAGAKU KOGAKU RONBUNSHU 2004, 30, 54-64 (in Japanese).
Received for review July 5, 2004. Revised manuscript received April 26, 2005. Accepted May 13, 2005. ES0489745
