- Email: [email protected]
Fs in a municipal waste incinerator during different operating periods
Chemosphere 67 (2007) S177–S184 www.elsevier.com/locate/chemosphere
Formation and removal of PCDD/Fs in a municipal waste incinerator during different operating periods Hou Chuan Wang a, Jyh Feng Hwang a, Kai Hsien Chi b, Moo Been Chang a
c,*
Environmental Health and Air Pollution Division, Center for Environmental, Safety and Health Technology, Industrial Technology Research Institute, Hsin-Chu 310, Taiwan, ROC b Research Center for Environmental Changes, Academia Sinica, Taipei 115, Taiwan, ROC c Graduate Institute of Environmental Engineering, National Central University, Chungli 320, Taiwan, ROC Accepted 26 May 2006 Available online 14 February 2007
Abstract The PCDD/F concentrations and removal efficiencies achieved with air pollution control devices (APCDs) during different operating periods (start-up, normal operation, and shut-down) at an existing municipal waste incinerator (MWI) in Taiwan are evaluated via stack sampling and analysis. The MWI investigated is equipped with electrostatic precipitators (EP), wet scrubbers (WS), and selective catalytic reduction system (SCR) as APCDs. The sampling results indicate that the PCDD/F concentrations at the EP inlet during start-up period were 15 times higher than that measured during normal operation period. The PCDD/F concentration observed at shut-down period was close to that measured at normal operation period. The CO concentration was between 400 and 1000 ppm during startup period, which was about 50 times higher compared with the normal operation. Hence, combustion condition significantly affected the PCDD/F formation concentration during the waste incineration process. In addition, the distributions of the PCDD/F congeners were similar at different operating periods. During the normal operation and shut-down periods, the EP decreases the PCDD/F concentration (based on TEQ) by 18.4–48.6%, while the removal efficiency of PCDD/Fs achieved with SCR system reaches 99.3–99.6%. Nevertheless, the PCDD/F removal efficiency achieved with SCR was only 42% during the 19-h start-up period due to the low SCR operating temperature (195 C). 2007 Elsevier Ltd. All rights reserved. Keywords: Start-up; Shut-down; SCR; Stack sampling; Dioxin emission
1. Introduction Previous study indicates that the PCDD/Fs (polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans) in ambient air originate mainly from waste incineration processes (Gotoh and Nakamura, 1999) and this has been a great concern in Taiwan. The flue gases of combustion processes contain varieties of incomplete combustion products, such as chlorinated aromatics and polycyclic aromatic hydrocarbons (Wienecke et al., 1995). In general, waste incineration process comprises three *
Corresponding author. Tel./fax: +886 3 4226774. E-mail address: [email protected] (M.B. Chang).
0045-6535/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2006.05.152
operating periods, including start-up, normal operation and shut-down stages. To prevent the rapid increase of combustion temperature to protect the insulation material of the incinerator, the start-up procedure usually takes 8 to 19 h with feeding of auxiliary fuel. On the other hand, the shut-down procedure consists of two steps, i.e. (1) stopping feeding wastes and (2) operating auxiliary burners. Relevant study indicates that elevated PCDD/F formation has been observed during the start-up and shut-down stages at the MWI (Grosso et al., 2004). Start-up process has been recognized as an important source of the PCDD/F emissions from the modern MWIs (Blumenstock et al., 2000). During and right after start-up period of waste incineration process, significant changes in the PCDD/F
S178
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
homologues profiles have been detected. Blumenstock et al. (2000) also indicate that the carbon deposited in the boiler is one of the major causes leading to high PCDD/F emissions from the MWIs. To effectively reduce the PCDD/F emissions, four methods have been commonly used as engineering practice to either inhibit PCDD/F formation or to remove PCDD/ Fs from gas streams. They include (1) addition of inhibitors, (2) high temperature decomposition (secondary combustion), (3) applying activated carbon injection (ACI) to adsorb PCDD/Fs, and (4) decomposition with catalysts. To meet the stringent PCDD/F emission standards, waste incinerators are commonly equipped with various types of air pollution control devices (APCDs). Relevant studies indicate that the TiO2-based V2O5/WO3 catalysts originally designed for removal of nitrogen oxides (NOx) via selective catalytic reduction (SCR) are also effective in the decomposition of PCDD/Fs (over 95%) (Busca et al., 1999; Liljelind et al., 2001). Chi et al. (2005) indicate that compared to the ACI technology, which only transfers gas-phase PCDD/Fs to the reacted ash and would make ash disposal even more complicated, SCR system actually destroys the PCDD/Fs and can serve as a better control technology for removing the PCDD/Fs from gas streams from the perspective of total environmental management. This study is motivated to establish the database of PCDD/F concentrations in the flue gas and to evaluate PCDD/F removal efficiencies achieved with various APCDs at different operating periods, including start-up, normal operation and shut-down stages. 2. Experimental 2.1. Sampling sites The large-scale MWI investigated in this study started to operate in 1995 and consists of four parallel lines, each with its own mechanical type grate, secondary combustion chamber and steam boiler for energy recovery. The APCDs applied in this MWI include electrostatic precipitators (EP) for particulate matter removal, two-stage wet scrubber (WS) for acid gas removal and selective catalyst reduction (SCR) reactor as a retrofit technology for reducing dioxin emission to meet the standards (0.1 ng-TEQ/Nm3). In the initial stage, this MWI was only equipped with the EP and the WS for controlling particulate and acid gas emissions. As high as 4.81 ng-I-TEQ/Nm3 the PCDD/F con-
centration in stack gas was measured at this MWI back to early 2000. Since 2002 SCR technology has been retrofitted in this MWI for reducing the PCDD/F emissions. The measurement program was conducted at four sampling positions (EP inlet, EP outlet, SCR inlet, and stack) during different operating stages, including start-up, normal operation and shut-down periods. Table 1 summarizes the operating conditions of the MWI investigated. Characteristics of the SCR system used in this MWI are shown in Table 2. The flue gas flow sheets and PCDD/F sampling points of the MWI investigated are schematically shown in Fig. 1. 2.2. Sample collection In this study, PCDD/F samples were collected at the EP inlet, the WS inlet, the SCR inlet and the stack, respectively, for better understanding of the PCDD/F formation and removal during waste incineration process. The flue gas sampling was conducted with Graseby Anderson Stack Sampling System, complying with the USEPA Method 23. The gas-phase sample was collected by XAD-2 resin, whereas the particle-bound portion was collected by the fiber glass filter and by rinsing of the sampling probe thereafter. To avoid the error and bias caused by the sampling of PCDD/Fs bound to the particulate matter, isokinetic sampling was conducted to collect representative samples. 2.3. Sample analysis Once the sampling was completed, the samples were brought back to the laboratory under refrigeration. They were then spiked with known amounts of USEPA Method 23 internal standard solution. Thereafter, the XAD-2 and filter sample were Soxhlet extracted with toluene for 24 h. The toluene extract was then concentrated to about 1 ml by rotary evaporation and was replaced by 5 ml hexane for pretreatment process. Having been treated with concentrated sulfuric acid, the sample was then subjected to a serTable 2 Characteristics of the SCR system applied in the MWI investigated Catalyst
V2O5/TiO2/WO3
Form SV (Nm3/h m3) AV (Nm3/h m2) Pressure drop (mm H2O/layer) Surface area (m2/m3)
Honeycomb 2558 2.9 50 870
Table 1 The operating temperature of the MWI investigated Parameter
Start-up (19-h)
Start-up (12-h)
Normal operation
Shut-down
Secondary chamber temperature (C) Boiler temperature (C) ESP temperature (C) WS temperature (C) SCR temperature (C)
320–800 220–570 180–205 55–64 199–200
330–800 220–650 180–200 55–63 203–215
900–1040 260–670 215–240 68 208–212
960–560 250–650 205–230 66 208–220
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
S179
SCR Boiler
EP
WS
Fig. 1. Layout of the MWI with the location of flue gas sampling points ( : EP inlet;
: WS inlet;
: SCR inlet;
or 19 h from ignition to reach stable combustion with operating temperature approaching 900 C. Relevant operating procedures of those two start-up models are listed in Table 3. Figs. 3 and 4 indicate the PCDD/F concentrations and TEQ concentrations measured at four sampling points during different operating stages, respectively. At EP inlet, PCDD/F concentrations in flue gas during the start-up periods were much higher than that of normal operation and shut-down periods. The PCDD/F concentrations at the EP inlet during start-up period were 15 times higher than that measured during normal operation period. CO concentration was between 400 and 1000 ppm during start-up period, which was about 50 times higher than that measured during normal combustion. Hence, combustion condition remarkably affected the PCDD/F formation during the waste incineration process (Weber et al., 2002). Under good and stable combustion conditions (e.g. during normal operation period), the PCDD/Fs were effectively destroyed inside the combustion chamber and the emissions were relatively low. However, bad and unstable combustion conditions during the start-up stage resulted in high concentrations of the products of incomplete combustion (PICs) like CO, soot, hydrocarbons and PCDD/Fs (Blumenstock et al., 2000). Table 3 also indicates that the wastes were fed into the incinerator as the combustion temperature reached 400 C. During the start-up stage, the
ies of clean-up columns including sulfuric acid silica gel column, acidic aluminum oxide column and Celite/Carbon column. Finally, the cleaned-up solution was spiked with known amounts of M-23 recovery standard solution, and then analyzed for seventeen 2,3,7,8-substituted PCDD/F congeners with high resolution gas chromatography/high resolution mass spectrometer (HRGC/HRMS) (JEOL JMS 700D), equipped with a DB5-MS capillary column (60 m · 0.25 mm · 0.25 lm film thickness). The mass spectrometer was operated with a resolution greater than 10 000 under the positive EI conditions, and data were obtained in the selected ion monitoring (SIM) mode. Based on the PCDD/F analysis method (NIEA.807.70C) used in Taiwan, the PCDD/F concentration in flue gas is expressed as ng-TEQ/Nm3, in which the normal (N) state refers to 1 atm, 0 C (273.15 K) and corrected for 11% oxygen content. 3. Results and discussion 3.1. PCDD/F concentrations in flue gas during different operating periods 3.1.1. Start-up and normal operation Two models of increasing temperature are available at the start-up stages. As shown in Fig. 2, it takes either 12 1,000
Temperature (0C)
800
600
400
200
19-hour start-up model 12-hour start-up model
0 0
2
4
6
8
: Stack).
10
12
14
16
18
Time (hour) Fig. 2. Temperature profiles for the two start-up models.
20
22
24
S180
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
Table 3 Operating procedure of 12-h and 19-h start-up model Operating procedure
12-h Model
19-h Model
Step 1
Step 4
The incinerator starts burning with diesel, after 8 h the combustion temperature increases to 400 C (50 C/h) Keep the combustion temperature at 400 C for 1 h Start feeding wastes into the incinerator, after 3 h the combustion temperature increases from 400 C to 900 C (170 C/h) Steady
Step 5
Steady
The incinerator starts burning with diesel, after 3 h the combustion temperature increases to 200 C Keep the combustion temperature at 200 C for 2 h Increase the combustion temperature from 200 C to 400 C in 4 h (50 C/h) Start feeding wastes into the incinerator and keep the combustion temperature at 400 C for 2 h Continually feeding wastes into the incinerator, after 8 h the combustion temperature increases to 900 C (60 C/h) and remains steady
Step 2 Step 3
5,000
PCDD/Fs (ng/Nm3)
EP inlet
WS inlet
SCR inlet
Stack
4,000
3,000
2,000
1,000
0 Start-up (19 hr)
Start-up (12 hr)
Normal operation
Shut-down
Fig. 3. The PCDD/Fs concentrations at four different operating models.
350 EP inlet
PCDD/Fs (ng-TEQ/Nm3)
300
WS inlet
SCR inlet
Stack
250 200 150 100 50 0 Start-up (19 hr)
Start-up (12 hr)
Normal operation
Shut-down
Fig. 4. The PCDD/Fs concentrations (on TEQ basis) at four different operating models.
combustion temperature was not high enough to reach complete combustion. Those PICs formed would eventually be carried to the steam boiler with the operating temperature between 220 C and 650 C, which was in the PCDD/F de novo synthesis temperature window (200– 450 C). Excess air (O2: 13–16%) and unburned carbon (soot) of the fly ash at temperature above 200 C would
cause significant formation of the PCDD/Fs at the boiler during the start-up period. 3.1.2. Shut-down period It was noted that the PCDD/F concentration measured at the EP inlet during the shut-down period was lower than that measured during the normal operation period. It
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
might be caused by the fact that the combustion temperature was still over 900 C and wastes were not fed into the incinerator during the shut-down period. The CO concentration was lower than 15 ppm. The HCl concentration was between 115 and 150 ppm before stopping of the waste feed and decreased to 60 ppm during the shut-down period. Although the concentration of HCl in flue gas became lower, the amounts of chlorides in fly ash (5000–6000 mg/kg) are sufficient to generate a small amount of PCDD/Fs. 3.2. PCDD/F congener distribution during different operating periods Dioxin congener concentrations and distributions at the EP inlet during four periods were shown in Figs. 5 and 6. It indicates that all the congener concentrations at the start-
normal operation
shut-down
S181
up mode were much higher than that at the normal operation and shut-down modes. During start-up period of the MSWI, a significant amount of lowly chlorinated congener with higher TEF (toxicity equivalency factor) and a small amount of OCDD and OCDF were generated. Interestingly, 19-h model results in a higher PCDD/F concentration but a lower TEQ value compared with 12-h model. It indicates that a significant amount of TCDDs and PCDFs with higher TEF (toxicity equivalency factor) was formed during the 12-h start-up model. At all operating periods, the PCDF/PCDD ratios at the EP inlet were all close to 1.0 (Table 4). It indicates that de novo synthesis might be the major mechanism responsible for the PCDD/F formation in the boiler at all operating periods. The characteristics also match with the results compiled in other countries (Hunsinger et al., 2002).
start-up (19 hours)
start-up (12 hours)
1000
3
Conc.(ng/Nm )
10000
100 010 001
D D PC s D Fs to ta lc on c I- . TE Q
PC
2 1 , ,3,7 2 1 , ,3, ,8-T 2 , 7, 3 8 C 1 , ,4,7 -Pe DD 2 , ,8 C 3 - D 1 , ,6,7 Hx D 2 , C 1 , ,3,7 8 -H DD 2 , ,8 x 3, ,9 CD 4, -H D 6, 7, xC 8- D H D pC D 2, O D 1 , 3 , 7 CD 2 , ,8 3 - D 2 , ,7,8 TC D 3 1 , ,4, -Pe F 2 , 7, C 3, 8 - DF 1 , 4,7 PeC 2 , ,8 3 -H D 1 , ,6,7 x C F 2 , ,8 3, -H DF 2 , 7,8 x 3 C 1, ,4,6 ,9 -H DF 2 , ,7 x 3, C 1, 4,6 ,8 -H DF 2, ,7 x 3, ,8 C 4, -H D 7, 8 , pC F 9- D H F pC D F O C DF
000
Fig. 5. Concentration of 2,3,7,8-substituted PCDD/F congeners at the EP inlet during different operating periods.
normal operation
shut-down
start-up (19 hours)
start-up (12 hours)
60 50
(%)
40 30 20 10
Fs PC
D
s D D PC
2, 1, 3 ,7, 2, 8 1 , 3,7 -TC 2, , 8 3, -P DD 1 , 4,7 , eC 2, 8 3, -H DD 1 , 6,7 x C , 2, 8- D 1 , 3,7, H x D 2, 8, C D 9 3, 4, -H D 6, 7 , x CD 8H D pC D D 2, O 3, C 1, 7, DD 2, 8 3, -T 2, 7,8 CD -P 3 F 1, ,4,7 eC 2, ,8 D 3, -P F 1, 4,7 eC 2, ,8 3, -H DF 1, 6,7 , x C 2, D 8 3, -H F 2, 7,8 x C , 3, D 9 1 , 4,6, -Hx F 2, 3, 7 ,8 - CD 1 , 4,6 Hx F 2 , ,7 3, ,8- CD 4, 7, Hp F 8, C 9- D H F pC D F O C DF
0
Fig. 6. The PCDD/F congener distributions (on total basis) at the EP inlet during different operating periods.
S182
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
Table 4 PCDF/PCDD ratios at different operating modes Stages
PCDF/PCDD ratio for total concentration Start-up
In the boiler
19-h
12-h
1.0
1.4
Normal operation
Shut-down
1.2
1.4
Although the PCDD/F congener concentrations at the EP inlet were different, the congener distributions were similar except for the distribution of highly chlorinated congener (OCDD and OCDF) during the start-up periods. This may be attributed to the fact that the composition of the wastes burned might be quite similar during four different operating periods. 2,3,4,7,8-PeCDF is the major contributor to the TEQ at all operating periods (Fig. 7). 3.3. Removal efficiency of PCDD/Fs in flue gas during different operating periods Removal efficiencies of PCDD/Fs achieved with APCDs at different operating periods were evaluated in this study (Table 5). The PCDD/F concentration in the flue gas decreased after passing through the EP during start-up period, but increased during normal operation and shut-down periods. It may be attributed to the fact that EP operating temperature at normal operation was between 210 C and 240 C, which was in the de novo synthesis temperature
normal operation shut-down
windows. Although EP can effectively capture the particles and solid-phase PCDD/Fs, part of the synthesized PCDD/F on the fly ash may desorb to the gas-phase and increases the outlet concentration. Fig. 8 compares the removal efficiencies of seventeen 2,3,7,8-substituted PCDD/F congeners achieved with the EP at different operating periods. At the shut-down period, the removal efficiencies of PCDD/Fs achieved with the EP are mostly negative except for the TCDD/F and OCDD. Interestingly, the removal efficiencies of PCDD/Fs achieved with EP are all negative at normal operation. The operating temperature of the EP during start-up was lower (180–205 C) compared to the normal operation and shut-down periods and it reduces the possibility of the PCDD/F re-synthesis at the start-up stage. The EP could capture particulate matter with a high collection efficiency (over 99%). Hence, the PCDD/Fs adsorbed on particulate matter could be simultaneously removed, resulting in the decrease of the PCDD/F concentration at EP outlet. The removal efficiency of PCDD/F congeners during the start-up period increased with the increasing chlorination. It is partly attributed to the fact that the vapor pressures of highly chlorinated PCDD/Fs are lower than that of lowly chlorinated congeners and have higher tendency to condense on particles and be removed by the EP. The removal efficiency of the PCDD/Fs achieved with the WS was between 77% and 84% based on the total PCDD/Fs (between 69% and 79% in terms of TEQ). The
start-up (19 hours)
start-up (12 hours)
80 70 60 (%)
50 40 30 20 10
Ds PC D Fs
PC D
2, 3, 7,8 1, -T 2, C 3,7 1, ,8 D D 2, 3,4 -PeC DD 1 , ,7 ,8 2, 3,6 Hx CD , 7 1, 2, ,8-H D 3,7 xC 1, ,8 D 2, 3,4 ,9-H D ,6 xC ,7, D 8D H pC DD O 2, C 3, 7,8 DD 1, TC 2, 3,7 D 2 , ,8 -P F 3, 4,7 eC D 1, ,8 F 2, 3,4 -Pe C DF 1 , ,7 ,8 2, 3,6 Hx C ,7 ,8- D F 1, 2, 3,7 Hx CD 2 , ,8 ,9 F 3, 4,6 Hx 1, , 7 , CD 2, 8 3,4 -H F 1 , ,6 ,7 x CD ,8 2, F -H 3,4 pC ,7 ,8, 9- DF H pC DF O C DF
0
Fig. 7. The PCDD/F congener distributions (on TEQ basis) at the EP inlet during different operating periods.
Table 5 Removal efficiencies of the PCDD/Fs achieved with the APCDs adopted at different operating periods Removal efficiency (%) Start-up
EP WS SCR APCDs
Removal efficiency based on TEQ (%) Normal operation
19-h
12-h
76.2 83.7 42.9 97.8
72.6 77.3 93.4 99.6
34.1 37.9 98.6 98.9
Shut-down
16.2 28.1 99.1 99.3
Start-up
Normal operation
19-h
12-h
74.7 68.7 70.9 97.7
64.7 78.8 95.7 99.7
48.6 17.1 99.6 99.5
Shut-down
18.4 2.3 99.3 99.2
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
S183
100 80
Removal efficiency (%)
60 40
Start-up (19-hour model)
20
Start-up (12-hour model)
0
Normal operation
-20
Shut-down
-40 -60 -80
DF C O
2, 3,
O C DD 7, 8 1, -T 2, C 3, D 7, F 8Pe 2, 3, C 4, D 7, F 81, 2, Pe 3, C 4, D 7, F 8 1, -H 2, xC 3, 6, D 7, F 81, H 2, xC 3, 7, D 8, F 92, H 3, xC 4, 6, D 7, F 1, 82, H 3, xC 4, 6, D 7, 1, F 82, H 3, pC 4, 7, D 8, F 9H pC D F
1, 2,
2, 3,
7, 8TC 3, D 7 D ,8 1, -P 2, eC 3, 4, D 7, D 81, H 2, xC 3, 6, D 7, D 81, H 2, xC 3, 7, D 8, D 1, 9 2, -H 3, xC 4, 6, D 7, D 8H pC D D
-100
Fig. 8. The PCDD/F removal efficiencies achieved with the EP.
removal efficiency of PCDD/Fs increases with increasing chlorination during the start-up stage (Fig. 9). This may be attributed to relatively low operating temperature of the WS (55–64 C) and most of the PCDD/Fs are condensed on the suspended solids. The vapor pressures of highly chlorinated PCDD/Fs were lower than that of lowly chlorinated PCDD/Fs and their partitioning in solid phase was higher. Some particles in flue gas were eventually removed by the WS. Takaoka et al. (2003) indicate that the lowly chlorinated congeners were more volatile and easier to desorb and transfer to the flue gas. Hence, highly chlorinated congeners partition more in solid phase and are easier to remove by the WS. However, the removal efficiency was low (especially of lower chlorinated PCDD/F congener) at both normal operation and shut-down periods (17% and 2.3% on TEQ basis, respectively). The low removal efficiency of lowly chlorinated PCDD/F may be attributed to the memory effect taking place in the WS.
‘‘Memory effect’’ (Giugliano et al., 2002) was especially evident when the PCDD/F concentration in the flue gas was low as during the normal operation and the shut-down periods, but was insignificant when the concentration of the PCDD/F in flue gas was high (e.g. during the startup period). Removal efficiency of the PCDD/Fs achieved with SCR depends on operating temperature and inlet particle concentration. As high as 99% removal efficiencies of the PCDD/Fs were achieved with SCR system during normal operation and shut-down periods with operating temperature over 210 C and inlet particle concentration less than 20 mg/Nm3. The SCR system seems to decompose all congeners of tetra- to octa-PCDD/Fs (Fig. 10). However, during 19 h start-up period, the PCDD/F removal efficiency achieved with the SCR was only 42% and removal efficiency of the PCDD/Fs decreased with increasing chlorination. Cho and Ihm (2002) indicate that lowly
100
Removal efficiency (%)
80
60
Start-up (19-hour model)
40
Start-up (12-hour model) Normal operation
20
Shut-down 0
-20
DF C O
2, 3,
O C DD 7, 81, T 2, C 3, D 7, F 8Pe 2, 3, C 4, D 7, F 81, 2, Pe 3, C 4, D 7, F 81, H 2, xC 3, 6, D 7, F 81, H 2, xC 3, 7, D 8, F 92, H 3, xC 4, 6, D 7, F 1, 8 2, -H 3, xC 4, 6, D 7 1, F ,8 2, -H 3, pC 4, 7, D 8, F 9H pC D F
1, 2,
2, 3,
7, 8TC 3, D 7 D ,8 1, -P 2, eC 3, 4, D 7, D 8 1, -H 2, xC 3, 6, D 7, D 81, H 2, x 3, CD 7, 8, D 1, 92, H 3, xC 4, 6, D 7, D 8H pC D D
-40
Fig. 9. The PCDD/F removal efficiencies achieved with the WS.
S184
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184 120 100
Removal efficiency (%)
80
Start-up (19-hour model)
60
Start-up (12-hour model)
40
Normal operation
20
Shut-down
0 -20 -40
DF C O
O C DD 7, 8TC 3, D 7, F 8Pe 2, 3, C D 4, F 7, 81, Pe 2, 3, C 4, D 7, F 81, H 2, xC 3, 6, D 7, F 81, H 2, xC 3, 7, D 8, F 9 2, -H 3, xC 4, 6, D 7 F 1, ,8 2, -H 3, xC 4, 6, D 7, 1, F 82, H 3, pC 4, 7, D 8, F 9H pC D F 2, 3,
1, 2,
1, 2,
2, 3,
7, 8TC 3, D 7, D 1, 8 -P 2, eC 3, 4, D 7, D 81, H 2, xC 3, 6, D 7, D 8 1, -H 2, xC 3, 7, D 8 D 1, ,9 2, -H 3, xC 4, 6, D 7, D 8H pC D D
-60
Fig. 10. The PCDD/F removal efficiencies achieved with the SCR.
chlorinated congeners are effectively decomposed by the SCR, while highly chlorinated congeners are not effectively removed. Low removal efficiency achieved at this stage may be attributed to the fact that the SCR operating temperature (195 C) was not high enough to decompose highly chlorinated PCDD/Fs, which were easy to condense on particles. 4. Conclusions In general, this study has demonstrated the effectiveness of applying the SCR technology for removing the PCDD/ Fs from flue gas streams. The combustion condition of incinerator (start-up period) and operating temperature of the SCR (shut-down period) significantly affect the PCDD/F formation and destruction, respectively. Although the overall removal efficiencies of the PCDD/Fs achieved with whole APCDs (EP + WS + SCR) at start-up periods were higher than 97%, the PCDD/F concentrations measured at the stack were sometimes higher than the emission standard of 0.1 ng-TEQ/Nm3. This was attributed to the high concentration of the PCDD/Fs entering the SCR due to de novo synthesis taking place in EP. To further reduce the PCDD/F emission, wastes had to be better processed prior to combustion (drying, shredding, or separation) and the initial incineration temperatures during the start-up periods should be increased if possible. Acknowledgements The authors gratefully acknowledge the financial support provided by the Taipei City Government of the Republic of China and the cooperation program between National Central University and Industrial Technology Research Institute (NCU-ITRI 930302).
References Blumenstock, M., Zimmermann, R., Schramm, K.W., Kettrup, A., 2000. Influence of combustion conditions on the PCDD/F-, PCB-, PCBzand PAH-concentrations in the post-combustion chamber of a waste incineration pilot plant. Chemosphere 40, 987–993. Busca, G., Baldi, M., Pistarino, C., Gallardo, J.M., Sanchez, E.V., Finocchio, E., Romezzano, G., Bregani, F., Toledo, G.P., 1999. Evaluation of V2O5–WO3–TiO2 and alternative SCR catalysts in the abatement of VOCs. Catalysis Today 53, 525–533. Chi, K.H., Chang, M.B., Chang-Chien, G.P., Lin, C., 2005. Characteristics of PCDD/F congener distributions in gas/particulate phases and emissions from two municipal solid waste incinerators in Taiwan. Science of the Total Environment 347, 148–162. Cho, C.H., Ihm, S.K., 2002. Development of new vanadium-based oxide catalysts for decomposition of chlorinated aromatic pollutants. Environmental Science and Technology 36, 1600–1606. Giugliano, M., Cernuschi, S., Grosso, M., Miglio, R., Aloigi, E., 2002. PCDD/F mass balance in the flue gas cleaning units of a MSW incineration plant. Chemosphere 46, 1321–1328. Gotoh, Y., Nakamura, Y., 1999. Japanese source inventory, focusing on the emission reduction measures in the manufacturing industries sector. Organohalogen Compounds 41, 477–480. Grosso, M., Cernuschi, S., Palini, E., Lodi, M., Mariani, G., 2004. PCDD/Fs release during normal and transient operation of a full scale MSWI plant. Organohalogen Compounds 66, 1243–1249. Hunsinger, H., Jay, K., Vehlow, J., 2002. Formation and destruction of PCDD/F inside a grate furnace. Chemosphere 46, 1263–1272. Liljelind, P., Unsworth, J., Maaskant, O., Marklund, S., 2001. Removal of PCDD/Fs and related aromatic hydrocarbons from flue gas streams by adsorption and catalytic destruction. Chemosphere 42, 615–623. Takaoka, T., Liao, P., Takeda, N., Fujiwara, T., Oshita, K., 2003. The behavior of PCDD/Fs, PCBs, chlorobenzenes and chlorophenols in wet scrubbing system of municipal solid waste incinerator. Chemosphere 53, 153–161. Weber, R., Sakurai, T., Ueno, S., Nishino, L., 2002. Correlation of PCDD/PCDF and CO values in a MSW incinerator – indication of memory effects in the high temperature/cooling section. Chemosphere 49, 127–134. Wienecke, J., Kruse, H., Huckfeldt, U., Eickhoff, W., Wassermann, O., 1995. Organic compounds in the flue gas of a hazardous waste incinerator. Chemosphere 30, 907–913.
Formation and removal of PCDD/Fs in a municipal waste incinerator during different operating periods Hou Chuan Wang a, Jyh Feng Hwang a, Kai Hsien Chi b, Moo Been Chang a
c,*
Environmental Health and Air Pollution Division, Center for Environmental, Safety and Health Technology, Industrial Technology Research Institute, Hsin-Chu 310, Taiwan, ROC b Research Center for Environmental Changes, Academia Sinica, Taipei 115, Taiwan, ROC c Graduate Institute of Environmental Engineering, National Central University, Chungli 320, Taiwan, ROC Accepted 26 May 2006 Available online 14 February 2007
Abstract The PCDD/F concentrations and removal efficiencies achieved with air pollution control devices (APCDs) during different operating periods (start-up, normal operation, and shut-down) at an existing municipal waste incinerator (MWI) in Taiwan are evaluated via stack sampling and analysis. The MWI investigated is equipped with electrostatic precipitators (EP), wet scrubbers (WS), and selective catalytic reduction system (SCR) as APCDs. The sampling results indicate that the PCDD/F concentrations at the EP inlet during start-up period were 15 times higher than that measured during normal operation period. The PCDD/F concentration observed at shut-down period was close to that measured at normal operation period. The CO concentration was between 400 and 1000 ppm during startup period, which was about 50 times higher compared with the normal operation. Hence, combustion condition significantly affected the PCDD/F formation concentration during the waste incineration process. In addition, the distributions of the PCDD/F congeners were similar at different operating periods. During the normal operation and shut-down periods, the EP decreases the PCDD/F concentration (based on TEQ) by 18.4–48.6%, while the removal efficiency of PCDD/Fs achieved with SCR system reaches 99.3–99.6%. Nevertheless, the PCDD/F removal efficiency achieved with SCR was only 42% during the 19-h start-up period due to the low SCR operating temperature (195 C). 2007 Elsevier Ltd. All rights reserved. Keywords: Start-up; Shut-down; SCR; Stack sampling; Dioxin emission
1. Introduction Previous study indicates that the PCDD/Fs (polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans) in ambient air originate mainly from waste incineration processes (Gotoh and Nakamura, 1999) and this has been a great concern in Taiwan. The flue gases of combustion processes contain varieties of incomplete combustion products, such as chlorinated aromatics and polycyclic aromatic hydrocarbons (Wienecke et al., 1995). In general, waste incineration process comprises three *
Corresponding author. Tel./fax: +886 3 4226774. E-mail address: [email protected] (M.B. Chang).
0045-6535/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2006.05.152
operating periods, including start-up, normal operation and shut-down stages. To prevent the rapid increase of combustion temperature to protect the insulation material of the incinerator, the start-up procedure usually takes 8 to 19 h with feeding of auxiliary fuel. On the other hand, the shut-down procedure consists of two steps, i.e. (1) stopping feeding wastes and (2) operating auxiliary burners. Relevant study indicates that elevated PCDD/F formation has been observed during the start-up and shut-down stages at the MWI (Grosso et al., 2004). Start-up process has been recognized as an important source of the PCDD/F emissions from the modern MWIs (Blumenstock et al., 2000). During and right after start-up period of waste incineration process, significant changes in the PCDD/F
S178
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
homologues profiles have been detected. Blumenstock et al. (2000) also indicate that the carbon deposited in the boiler is one of the major causes leading to high PCDD/F emissions from the MWIs. To effectively reduce the PCDD/F emissions, four methods have been commonly used as engineering practice to either inhibit PCDD/F formation or to remove PCDD/ Fs from gas streams. They include (1) addition of inhibitors, (2) high temperature decomposition (secondary combustion), (3) applying activated carbon injection (ACI) to adsorb PCDD/Fs, and (4) decomposition with catalysts. To meet the stringent PCDD/F emission standards, waste incinerators are commonly equipped with various types of air pollution control devices (APCDs). Relevant studies indicate that the TiO2-based V2O5/WO3 catalysts originally designed for removal of nitrogen oxides (NOx) via selective catalytic reduction (SCR) are also effective in the decomposition of PCDD/Fs (over 95%) (Busca et al., 1999; Liljelind et al., 2001). Chi et al. (2005) indicate that compared to the ACI technology, which only transfers gas-phase PCDD/Fs to the reacted ash and would make ash disposal even more complicated, SCR system actually destroys the PCDD/Fs and can serve as a better control technology for removing the PCDD/Fs from gas streams from the perspective of total environmental management. This study is motivated to establish the database of PCDD/F concentrations in the flue gas and to evaluate PCDD/F removal efficiencies achieved with various APCDs at different operating periods, including start-up, normal operation and shut-down stages. 2. Experimental 2.1. Sampling sites The large-scale MWI investigated in this study started to operate in 1995 and consists of four parallel lines, each with its own mechanical type grate, secondary combustion chamber and steam boiler for energy recovery. The APCDs applied in this MWI include electrostatic precipitators (EP) for particulate matter removal, two-stage wet scrubber (WS) for acid gas removal and selective catalyst reduction (SCR) reactor as a retrofit technology for reducing dioxin emission to meet the standards (0.1 ng-TEQ/Nm3). In the initial stage, this MWI was only equipped with the EP and the WS for controlling particulate and acid gas emissions. As high as 4.81 ng-I-TEQ/Nm3 the PCDD/F con-
centration in stack gas was measured at this MWI back to early 2000. Since 2002 SCR technology has been retrofitted in this MWI for reducing the PCDD/F emissions. The measurement program was conducted at four sampling positions (EP inlet, EP outlet, SCR inlet, and stack) during different operating stages, including start-up, normal operation and shut-down periods. Table 1 summarizes the operating conditions of the MWI investigated. Characteristics of the SCR system used in this MWI are shown in Table 2. The flue gas flow sheets and PCDD/F sampling points of the MWI investigated are schematically shown in Fig. 1. 2.2. Sample collection In this study, PCDD/F samples were collected at the EP inlet, the WS inlet, the SCR inlet and the stack, respectively, for better understanding of the PCDD/F formation and removal during waste incineration process. The flue gas sampling was conducted with Graseby Anderson Stack Sampling System, complying with the USEPA Method 23. The gas-phase sample was collected by XAD-2 resin, whereas the particle-bound portion was collected by the fiber glass filter and by rinsing of the sampling probe thereafter. To avoid the error and bias caused by the sampling of PCDD/Fs bound to the particulate matter, isokinetic sampling was conducted to collect representative samples. 2.3. Sample analysis Once the sampling was completed, the samples were brought back to the laboratory under refrigeration. They were then spiked with known amounts of USEPA Method 23 internal standard solution. Thereafter, the XAD-2 and filter sample were Soxhlet extracted with toluene for 24 h. The toluene extract was then concentrated to about 1 ml by rotary evaporation and was replaced by 5 ml hexane for pretreatment process. Having been treated with concentrated sulfuric acid, the sample was then subjected to a serTable 2 Characteristics of the SCR system applied in the MWI investigated Catalyst
V2O5/TiO2/WO3
Form SV (Nm3/h m3) AV (Nm3/h m2) Pressure drop (mm H2O/layer) Surface area (m2/m3)
Honeycomb 2558 2.9 50 870
Table 1 The operating temperature of the MWI investigated Parameter
Start-up (19-h)
Start-up (12-h)
Normal operation
Shut-down
Secondary chamber temperature (C) Boiler temperature (C) ESP temperature (C) WS temperature (C) SCR temperature (C)
320–800 220–570 180–205 55–64 199–200
330–800 220–650 180–200 55–63 203–215
900–1040 260–670 215–240 68 208–212
960–560 250–650 205–230 66 208–220
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
S179
SCR Boiler
EP
WS
Fig. 1. Layout of the MWI with the location of flue gas sampling points ( : EP inlet;
: WS inlet;
: SCR inlet;
or 19 h from ignition to reach stable combustion with operating temperature approaching 900 C. Relevant operating procedures of those two start-up models are listed in Table 3. Figs. 3 and 4 indicate the PCDD/F concentrations and TEQ concentrations measured at four sampling points during different operating stages, respectively. At EP inlet, PCDD/F concentrations in flue gas during the start-up periods were much higher than that of normal operation and shut-down periods. The PCDD/F concentrations at the EP inlet during start-up period were 15 times higher than that measured during normal operation period. CO concentration was between 400 and 1000 ppm during start-up period, which was about 50 times higher than that measured during normal combustion. Hence, combustion condition remarkably affected the PCDD/F formation during the waste incineration process (Weber et al., 2002). Under good and stable combustion conditions (e.g. during normal operation period), the PCDD/Fs were effectively destroyed inside the combustion chamber and the emissions were relatively low. However, bad and unstable combustion conditions during the start-up stage resulted in high concentrations of the products of incomplete combustion (PICs) like CO, soot, hydrocarbons and PCDD/Fs (Blumenstock et al., 2000). Table 3 also indicates that the wastes were fed into the incinerator as the combustion temperature reached 400 C. During the start-up stage, the
ies of clean-up columns including sulfuric acid silica gel column, acidic aluminum oxide column and Celite/Carbon column. Finally, the cleaned-up solution was spiked with known amounts of M-23 recovery standard solution, and then analyzed for seventeen 2,3,7,8-substituted PCDD/F congeners with high resolution gas chromatography/high resolution mass spectrometer (HRGC/HRMS) (JEOL JMS 700D), equipped with a DB5-MS capillary column (60 m · 0.25 mm · 0.25 lm film thickness). The mass spectrometer was operated with a resolution greater than 10 000 under the positive EI conditions, and data were obtained in the selected ion monitoring (SIM) mode. Based on the PCDD/F analysis method (NIEA.807.70C) used in Taiwan, the PCDD/F concentration in flue gas is expressed as ng-TEQ/Nm3, in which the normal (N) state refers to 1 atm, 0 C (273.15 K) and corrected for 11% oxygen content. 3. Results and discussion 3.1. PCDD/F concentrations in flue gas during different operating periods 3.1.1. Start-up and normal operation Two models of increasing temperature are available at the start-up stages. As shown in Fig. 2, it takes either 12 1,000
Temperature (0C)
800
600
400
200
19-hour start-up model 12-hour start-up model
0 0
2
4
6
8
: Stack).
10
12
14
16
18
Time (hour) Fig. 2. Temperature profiles for the two start-up models.
20
22
24
S180
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
Table 3 Operating procedure of 12-h and 19-h start-up model Operating procedure
12-h Model
19-h Model
Step 1
Step 4
The incinerator starts burning with diesel, after 8 h the combustion temperature increases to 400 C (50 C/h) Keep the combustion temperature at 400 C for 1 h Start feeding wastes into the incinerator, after 3 h the combustion temperature increases from 400 C to 900 C (170 C/h) Steady
Step 5
Steady
The incinerator starts burning with diesel, after 3 h the combustion temperature increases to 200 C Keep the combustion temperature at 200 C for 2 h Increase the combustion temperature from 200 C to 400 C in 4 h (50 C/h) Start feeding wastes into the incinerator and keep the combustion temperature at 400 C for 2 h Continually feeding wastes into the incinerator, after 8 h the combustion temperature increases to 900 C (60 C/h) and remains steady
Step 2 Step 3
5,000
PCDD/Fs (ng/Nm3)
EP inlet
WS inlet
SCR inlet
Stack
4,000
3,000
2,000
1,000
0 Start-up (19 hr)
Start-up (12 hr)
Normal operation
Shut-down
Fig. 3. The PCDD/Fs concentrations at four different operating models.
350 EP inlet
PCDD/Fs (ng-TEQ/Nm3)
300
WS inlet
SCR inlet
Stack
250 200 150 100 50 0 Start-up (19 hr)
Start-up (12 hr)
Normal operation
Shut-down
Fig. 4. The PCDD/Fs concentrations (on TEQ basis) at four different operating models.
combustion temperature was not high enough to reach complete combustion. Those PICs formed would eventually be carried to the steam boiler with the operating temperature between 220 C and 650 C, which was in the PCDD/F de novo synthesis temperature window (200– 450 C). Excess air (O2: 13–16%) and unburned carbon (soot) of the fly ash at temperature above 200 C would
cause significant formation of the PCDD/Fs at the boiler during the start-up period. 3.1.2. Shut-down period It was noted that the PCDD/F concentration measured at the EP inlet during the shut-down period was lower than that measured during the normal operation period. It
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
might be caused by the fact that the combustion temperature was still over 900 C and wastes were not fed into the incinerator during the shut-down period. The CO concentration was lower than 15 ppm. The HCl concentration was between 115 and 150 ppm before stopping of the waste feed and decreased to 60 ppm during the shut-down period. Although the concentration of HCl in flue gas became lower, the amounts of chlorides in fly ash (5000–6000 mg/kg) are sufficient to generate a small amount of PCDD/Fs. 3.2. PCDD/F congener distribution during different operating periods Dioxin congener concentrations and distributions at the EP inlet during four periods were shown in Figs. 5 and 6. It indicates that all the congener concentrations at the start-
normal operation
shut-down
S181
up mode were much higher than that at the normal operation and shut-down modes. During start-up period of the MSWI, a significant amount of lowly chlorinated congener with higher TEF (toxicity equivalency factor) and a small amount of OCDD and OCDF were generated. Interestingly, 19-h model results in a higher PCDD/F concentration but a lower TEQ value compared with 12-h model. It indicates that a significant amount of TCDDs and PCDFs with higher TEF (toxicity equivalency factor) was formed during the 12-h start-up model. At all operating periods, the PCDF/PCDD ratios at the EP inlet were all close to 1.0 (Table 4). It indicates that de novo synthesis might be the major mechanism responsible for the PCDD/F formation in the boiler at all operating periods. The characteristics also match with the results compiled in other countries (Hunsinger et al., 2002).
start-up (19 hours)
start-up (12 hours)
1000
3
Conc.(ng/Nm )
10000
100 010 001
D D PC s D Fs to ta lc on c I- . TE Q
PC
2 1 , ,3,7 2 1 , ,3, ,8-T 2 , 7, 3 8 C 1 , ,4,7 -Pe DD 2 , ,8 C 3 - D 1 , ,6,7 Hx D 2 , C 1 , ,3,7 8 -H DD 2 , ,8 x 3, ,9 CD 4, -H D 6, 7, xC 8- D H D pC D 2, O D 1 , 3 , 7 CD 2 , ,8 3 - D 2 , ,7,8 TC D 3 1 , ,4, -Pe F 2 , 7, C 3, 8 - DF 1 , 4,7 PeC 2 , ,8 3 -H D 1 , ,6,7 x C F 2 , ,8 3, -H DF 2 , 7,8 x 3 C 1, ,4,6 ,9 -H DF 2 , ,7 x 3, C 1, 4,6 ,8 -H DF 2, ,7 x 3, ,8 C 4, -H D 7, 8 , pC F 9- D H F pC D F O C DF
000
Fig. 5. Concentration of 2,3,7,8-substituted PCDD/F congeners at the EP inlet during different operating periods.
normal operation
shut-down
start-up (19 hours)
start-up (12 hours)
60 50
(%)
40 30 20 10
Fs PC
D
s D D PC
2, 1, 3 ,7, 2, 8 1 , 3,7 -TC 2, , 8 3, -P DD 1 , 4,7 , eC 2, 8 3, -H DD 1 , 6,7 x C , 2, 8- D 1 , 3,7, H x D 2, 8, C D 9 3, 4, -H D 6, 7 , x CD 8H D pC D D 2, O 3, C 1, 7, DD 2, 8 3, -T 2, 7,8 CD -P 3 F 1, ,4,7 eC 2, ,8 D 3, -P F 1, 4,7 eC 2, ,8 3, -H DF 1, 6,7 , x C 2, D 8 3, -H F 2, 7,8 x C , 3, D 9 1 , 4,6, -Hx F 2, 3, 7 ,8 - CD 1 , 4,6 Hx F 2 , ,7 3, ,8- CD 4, 7, Hp F 8, C 9- D H F pC D F O C DF
0
Fig. 6. The PCDD/F congener distributions (on total basis) at the EP inlet during different operating periods.
S182
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
Table 4 PCDF/PCDD ratios at different operating modes Stages
PCDF/PCDD ratio for total concentration Start-up
In the boiler
19-h
12-h
1.0
1.4
Normal operation
Shut-down
1.2
1.4
Although the PCDD/F congener concentrations at the EP inlet were different, the congener distributions were similar except for the distribution of highly chlorinated congener (OCDD and OCDF) during the start-up periods. This may be attributed to the fact that the composition of the wastes burned might be quite similar during four different operating periods. 2,3,4,7,8-PeCDF is the major contributor to the TEQ at all operating periods (Fig. 7). 3.3. Removal efficiency of PCDD/Fs in flue gas during different operating periods Removal efficiencies of PCDD/Fs achieved with APCDs at different operating periods were evaluated in this study (Table 5). The PCDD/F concentration in the flue gas decreased after passing through the EP during start-up period, but increased during normal operation and shut-down periods. It may be attributed to the fact that EP operating temperature at normal operation was between 210 C and 240 C, which was in the de novo synthesis temperature
normal operation shut-down
windows. Although EP can effectively capture the particles and solid-phase PCDD/Fs, part of the synthesized PCDD/F on the fly ash may desorb to the gas-phase and increases the outlet concentration. Fig. 8 compares the removal efficiencies of seventeen 2,3,7,8-substituted PCDD/F congeners achieved with the EP at different operating periods. At the shut-down period, the removal efficiencies of PCDD/Fs achieved with the EP are mostly negative except for the TCDD/F and OCDD. Interestingly, the removal efficiencies of PCDD/Fs achieved with EP are all negative at normal operation. The operating temperature of the EP during start-up was lower (180–205 C) compared to the normal operation and shut-down periods and it reduces the possibility of the PCDD/F re-synthesis at the start-up stage. The EP could capture particulate matter with a high collection efficiency (over 99%). Hence, the PCDD/Fs adsorbed on particulate matter could be simultaneously removed, resulting in the decrease of the PCDD/F concentration at EP outlet. The removal efficiency of PCDD/F congeners during the start-up period increased with the increasing chlorination. It is partly attributed to the fact that the vapor pressures of highly chlorinated PCDD/Fs are lower than that of lowly chlorinated congeners and have higher tendency to condense on particles and be removed by the EP. The removal efficiency of the PCDD/Fs achieved with the WS was between 77% and 84% based on the total PCDD/Fs (between 69% and 79% in terms of TEQ). The
start-up (19 hours)
start-up (12 hours)
80 70 60 (%)
50 40 30 20 10
Ds PC D Fs
PC D
2, 3, 7,8 1, -T 2, C 3,7 1, ,8 D D 2, 3,4 -PeC DD 1 , ,7 ,8 2, 3,6 Hx CD , 7 1, 2, ,8-H D 3,7 xC 1, ,8 D 2, 3,4 ,9-H D ,6 xC ,7, D 8D H pC DD O 2, C 3, 7,8 DD 1, TC 2, 3,7 D 2 , ,8 -P F 3, 4,7 eC D 1, ,8 F 2, 3,4 -Pe C DF 1 , ,7 ,8 2, 3,6 Hx C ,7 ,8- D F 1, 2, 3,7 Hx CD 2 , ,8 ,9 F 3, 4,6 Hx 1, , 7 , CD 2, 8 3,4 -H F 1 , ,6 ,7 x CD ,8 2, F -H 3,4 pC ,7 ,8, 9- DF H pC DF O C DF
0
Fig. 7. The PCDD/F congener distributions (on TEQ basis) at the EP inlet during different operating periods.
Table 5 Removal efficiencies of the PCDD/Fs achieved with the APCDs adopted at different operating periods Removal efficiency (%) Start-up
EP WS SCR APCDs
Removal efficiency based on TEQ (%) Normal operation
19-h
12-h
76.2 83.7 42.9 97.8
72.6 77.3 93.4 99.6
34.1 37.9 98.6 98.9
Shut-down
16.2 28.1 99.1 99.3
Start-up
Normal operation
19-h
12-h
74.7 68.7 70.9 97.7
64.7 78.8 95.7 99.7
48.6 17.1 99.6 99.5
Shut-down
18.4 2.3 99.3 99.2
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184
S183
100 80
Removal efficiency (%)
60 40
Start-up (19-hour model)
20
Start-up (12-hour model)
0
Normal operation
-20
Shut-down
-40 -60 -80
DF C O
2, 3,
O C DD 7, 8 1, -T 2, C 3, D 7, F 8Pe 2, 3, C 4, D 7, F 81, 2, Pe 3, C 4, D 7, F 8 1, -H 2, xC 3, 6, D 7, F 81, H 2, xC 3, 7, D 8, F 92, H 3, xC 4, 6, D 7, F 1, 82, H 3, xC 4, 6, D 7, 1, F 82, H 3, pC 4, 7, D 8, F 9H pC D F
1, 2,
2, 3,
7, 8TC 3, D 7 D ,8 1, -P 2, eC 3, 4, D 7, D 81, H 2, xC 3, 6, D 7, D 81, H 2, xC 3, 7, D 8, D 1, 9 2, -H 3, xC 4, 6, D 7, D 8H pC D D
-100
Fig. 8. The PCDD/F removal efficiencies achieved with the EP.
removal efficiency of PCDD/Fs increases with increasing chlorination during the start-up stage (Fig. 9). This may be attributed to relatively low operating temperature of the WS (55–64 C) and most of the PCDD/Fs are condensed on the suspended solids. The vapor pressures of highly chlorinated PCDD/Fs were lower than that of lowly chlorinated PCDD/Fs and their partitioning in solid phase was higher. Some particles in flue gas were eventually removed by the WS. Takaoka et al. (2003) indicate that the lowly chlorinated congeners were more volatile and easier to desorb and transfer to the flue gas. Hence, highly chlorinated congeners partition more in solid phase and are easier to remove by the WS. However, the removal efficiency was low (especially of lower chlorinated PCDD/F congener) at both normal operation and shut-down periods (17% and 2.3% on TEQ basis, respectively). The low removal efficiency of lowly chlorinated PCDD/F may be attributed to the memory effect taking place in the WS.
‘‘Memory effect’’ (Giugliano et al., 2002) was especially evident when the PCDD/F concentration in the flue gas was low as during the normal operation and the shut-down periods, but was insignificant when the concentration of the PCDD/F in flue gas was high (e.g. during the startup period). Removal efficiency of the PCDD/Fs achieved with SCR depends on operating temperature and inlet particle concentration. As high as 99% removal efficiencies of the PCDD/Fs were achieved with SCR system during normal operation and shut-down periods with operating temperature over 210 C and inlet particle concentration less than 20 mg/Nm3. The SCR system seems to decompose all congeners of tetra- to octa-PCDD/Fs (Fig. 10). However, during 19 h start-up period, the PCDD/F removal efficiency achieved with the SCR was only 42% and removal efficiency of the PCDD/Fs decreased with increasing chlorination. Cho and Ihm (2002) indicate that lowly
100
Removal efficiency (%)
80
60
Start-up (19-hour model)
40
Start-up (12-hour model) Normal operation
20
Shut-down 0
-20
DF C O
2, 3,
O C DD 7, 81, T 2, C 3, D 7, F 8Pe 2, 3, C 4, D 7, F 81, 2, Pe 3, C 4, D 7, F 81, H 2, xC 3, 6, D 7, F 81, H 2, xC 3, 7, D 8, F 92, H 3, xC 4, 6, D 7, F 1, 8 2, -H 3, xC 4, 6, D 7 1, F ,8 2, -H 3, pC 4, 7, D 8, F 9H pC D F
1, 2,
2, 3,
7, 8TC 3, D 7 D ,8 1, -P 2, eC 3, 4, D 7, D 8 1, -H 2, xC 3, 6, D 7, D 81, H 2, x 3, CD 7, 8, D 1, 92, H 3, xC 4, 6, D 7, D 8H pC D D
-40
Fig. 9. The PCDD/F removal efficiencies achieved with the WS.
S184
H.C. Wang et al. / Chemosphere 67 (2007) S177–S184 120 100
Removal efficiency (%)
80
Start-up (19-hour model)
60
Start-up (12-hour model)
40
Normal operation
20
Shut-down
0 -20 -40
DF C O
O C DD 7, 8TC 3, D 7, F 8Pe 2, 3, C D 4, F 7, 81, Pe 2, 3, C 4, D 7, F 81, H 2, xC 3, 6, D 7, F 81, H 2, xC 3, 7, D 8, F 9 2, -H 3, xC 4, 6, D 7 F 1, ,8 2, -H 3, xC 4, 6, D 7, 1, F 82, H 3, pC 4, 7, D 8, F 9H pC D F 2, 3,
1, 2,
1, 2,
2, 3,
7, 8TC 3, D 7, D 1, 8 -P 2, eC 3, 4, D 7, D 81, H 2, xC 3, 6, D 7, D 8 1, -H 2, xC 3, 7, D 8 D 1, ,9 2, -H 3, xC 4, 6, D 7, D 8H pC D D
-60
Fig. 10. The PCDD/F removal efficiencies achieved with the SCR.
chlorinated congeners are effectively decomposed by the SCR, while highly chlorinated congeners are not effectively removed. Low removal efficiency achieved at this stage may be attributed to the fact that the SCR operating temperature (195 C) was not high enough to decompose highly chlorinated PCDD/Fs, which were easy to condense on particles. 4. Conclusions In general, this study has demonstrated the effectiveness of applying the SCR technology for removing the PCDD/ Fs from flue gas streams. The combustion condition of incinerator (start-up period) and operating temperature of the SCR (shut-down period) significantly affect the PCDD/F formation and destruction, respectively. Although the overall removal efficiencies of the PCDD/Fs achieved with whole APCDs (EP + WS + SCR) at start-up periods were higher than 97%, the PCDD/F concentrations measured at the stack were sometimes higher than the emission standard of 0.1 ng-TEQ/Nm3. This was attributed to the high concentration of the PCDD/Fs entering the SCR due to de novo synthesis taking place in EP. To further reduce the PCDD/F emission, wastes had to be better processed prior to combustion (drying, shredding, or separation) and the initial incineration temperatures during the start-up periods should be increased if possible. Acknowledgements The authors gratefully acknowledge the financial support provided by the Taipei City Government of the Republic of China and the cooperation program between National Central University and Industrial Technology Research Institute (NCU-ITRI 930302).
References Blumenstock, M., Zimmermann, R., Schramm, K.W., Kettrup, A., 2000. Influence of combustion conditions on the PCDD/F-, PCB-, PCBzand PAH-concentrations in the post-combustion chamber of a waste incineration pilot plant. Chemosphere 40, 987–993. Busca, G., Baldi, M., Pistarino, C., Gallardo, J.M., Sanchez, E.V., Finocchio, E., Romezzano, G., Bregani, F., Toledo, G.P., 1999. Evaluation of V2O5–WO3–TiO2 and alternative SCR catalysts in the abatement of VOCs. Catalysis Today 53, 525–533. Chi, K.H., Chang, M.B., Chang-Chien, G.P., Lin, C., 2005. Characteristics of PCDD/F congener distributions in gas/particulate phases and emissions from two municipal solid waste incinerators in Taiwan. Science of the Total Environment 347, 148–162. Cho, C.H., Ihm, S.K., 2002. Development of new vanadium-based oxide catalysts for decomposition of chlorinated aromatic pollutants. Environmental Science and Technology 36, 1600–1606. Giugliano, M., Cernuschi, S., Grosso, M., Miglio, R., Aloigi, E., 2002. PCDD/F mass balance in the flue gas cleaning units of a MSW incineration plant. Chemosphere 46, 1321–1328. Gotoh, Y., Nakamura, Y., 1999. Japanese source inventory, focusing on the emission reduction measures in the manufacturing industries sector. Organohalogen Compounds 41, 477–480. Grosso, M., Cernuschi, S., Palini, E., Lodi, M., Mariani, G., 2004. PCDD/Fs release during normal and transient operation of a full scale MSWI plant. Organohalogen Compounds 66, 1243–1249. Hunsinger, H., Jay, K., Vehlow, J., 2002. Formation and destruction of PCDD/F inside a grate furnace. Chemosphere 46, 1263–1272. Liljelind, P., Unsworth, J., Maaskant, O., Marklund, S., 2001. Removal of PCDD/Fs and related aromatic hydrocarbons from flue gas streams by adsorption and catalytic destruction. Chemosphere 42, 615–623. Takaoka, T., Liao, P., Takeda, N., Fujiwara, T., Oshita, K., 2003. The behavior of PCDD/Fs, PCBs, chlorobenzenes and chlorophenols in wet scrubbing system of municipal solid waste incinerator. Chemosphere 53, 153–161. Weber, R., Sakurai, T., Ueno, S., Nishino, L., 2002. Correlation of PCDD/PCDF and CO values in a MSW incinerator – indication of memory effects in the high temperature/cooling section. Chemosphere 49, 127–134. Wienecke, J., Kruse, H., Huckfeldt, U., Eickhoff, W., Wassermann, O., 1995. Organic compounds in the flue gas of a hazardous waste incinerator. Chemosphere 30, 907–913.