The PCDD/F Removal Efficiency of a Medical Waste Incinerator Dual-Bag Filter System

This study investigated the polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs) removal efficiencies in two medical waste incinerators (MWIs) -MWI-A and MWI-B, with waste burning capacities of 9.12 and 12 tons per day, with dual-and single-bag filter systems, respectively. The PCDD/Fs in the stack flue gas and fly ashes of each air pollution control device (APCD) unit were collected concurrently. Based on mass balance, it was determined that the major PCDD/F removal occurred in the bag filter system with activated carbon injection (ACI). The total PCDD/F emission removal efficiencies for the whole dual-bag filter system, bag filter 1 (BF1), bag filter 2 (BF2) in MWI-A and the single bag filter system (BF) in MWI-B were 99.7%, 84.8%, 98.0%, 99.0% on the mass basis, respectively. The PCDD/F emission ratio (the escaped PCDD/F ratio after all APCDs) with the dual-bag filter system was over four times lower than that with a single bag filter. The results demonstrate a very efficient way to reduce the excess PCDD/F emissions incurred by intermittent operations and irregular waste feeding.


INTRODUCTION
Medical waste is discarded matter generated from medical activities, and has a high potential negative impact on both public health and the environment (Xie et al., 2009).The main methods available for disposal of medical waste are autoclaves, retorts, microwave disinfection systems, chemical disinfection and incineration (Lee et al., 2004).Among these, incineration is the most widely used treatment for the disposal of medical waste.Combustion has notable advantages in volume and weight reduction, destruction of pathogens and hazardous organics, and a relatively fast process, In addition, the waste heat generated from combustion can also be recycled.
However, many studies have highlighted the health risks associated with the emission of persistent organic pollutants (POPs) during incineration, such as polychlorinated dibenzop-dioxins and polychlorinated dibenzofurans (PCDD/Fs), polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), polybrominated dibenzo-p-dioxins and dibenzofurans (PBDD/Fs), and polybrominated diphenyl ethers (PBDEs) (Lee et al., 2002;Lee et al., 2003;Wang et al., 2010;Wang et al., 2011).A recent study indicated that the PCDD/F emissions from small medical waste incinerators (MWIs) are much higher than those from large ones (Choi et al., 2008), with the possible reasons including poor performance by air pollution control devices, intermittent operations, and irregular waste feeding.Coutinho et al. (2006) found that the shutdown of two MWIs in Porto, Portugal, caused a decrease in the atmospheric concentrations of PCDD/Fs of approximately 50%, underlining the fact that such emissions deserve greater attention.
The chlorine content of medical waste is usually higher than that of other municipal solid wastes, due to the heavy use of chlorinated plastics, such as polyvinyl chloride (PVC).Several researches reported that chlorine plays a significant role in the formation of PCDD/Fs via de novo synthesis, which is a key process in MWIs during precursor formation (Takeshi et al., 2005;Wang et al., 2008).Grochowalski (1998) reported that most MWIs with two-stage thermal processes in Poland can meet the emissions standard of 0.1 ng I-TEQ/Nm 3 that is adopted by most developed countries (Mininni et al., 2007;Streibel et al., 2007), although other studies reported that this is difficult to achieve in Italy, Colombia, China and Taiwan (Mininni et al., 2007;Hoyos et al., 2008;Lee et al., 2008;Gao et al., 2009).The 2004 European Dioxin Emission Inventory Project showed that are still an unknown number of medical waste incinerators with a large degree of variability in their emissions, and which can thus be considered important local sources of pollution (Quass et al., 2004).
The current emission limit of MWIs in Taiwan is 0.5 ng I-TEQ/Nm 3 , which is five times the European Union limit (Coutinho et al., 2006).The goal of this study is to reduce the PCDD/F emissions from two incinerators, MWI-A and MWI-B, both of which are equipped with systems that utilize two-stage combustion chambers and a combination of semi-dry scrubber, activated carbon injection and bag filters.These incinerators have designed capacities of 9.12 and 12.0 ton/d, respectively, and MWI-A has sometimes been found to have PCDD/F emission higher than 0.5 ng I-TEQ/Nm 3 .
In order to reduce PCDD/F emissions, MWI-A was modified from a single-to dual-bag filter system in 2009.The removal efficiencies between these two MWIs were compared in this study, as measured by the PCDD/F contents in bottom ash and fly ash from each air pollution control device (APCDs) unit.The results show that when the PCDD/F concentration in the exhaust gas is high, the removal efficiency of a single bag filter was unsatisfactory (85.8%), while the dual-bag filter system had a much better performance (99.2%).This study thus suggests that a dualbag filter system can greatly reduce the excess PCDD/F emissions resulting from intermittent operations and irregular waste feeding.

Basic Information Concerning the Incinerators
The two medical waste incinerators (capacity: 9.12 and 12.0 ton/d) began operations in April and December 2004, respectively.Both incinerators utilize two-stage combustion chambers (800°C and 1200°C, respectively).Both APCDs include a semi-dry scrubber (250°C), activated carbon injection (3 kg/h), and bag filter system (150°C).Basic information about the operations of these MWIs is given in Table 1.Both MWIs operate for approximately 200 days per year, with the only difference being the bag filters, with MWI-A having a dual-bag filter system, while MWI-B has a single bag filter one.Since Taiwan's regulations require that medical waste can only be stored for one day under normal atmospheric temperature, or up to seven days at 5°C or below, intermittent operations and irregular waste feeding are unavoidable.Also, during start-up, both incinerators are preheated by diesel-fueled burners.

PCDD/F Sampling
The stack flue gas and fly ashes from the semi-dry scrubber (Scrubber), bag filter (BF), bag filter 1 (BF1), bag filter 2 (BF2) at each MWI were collected concurrently.All samplings and chemical analyses in this study were carried out by an accredited laboratory in Taiwan.The flue gas sampling duration was 2 h.Sampling was conducted three times at each stack, and samples were collected isokinetically from the flue gases according to USEPA Modified Method 23.The major gas components, including CO, CO 2 , O 2 , NO x , and SO x , as well as temperature, were monitored with a combustion gas analyzer every minute during the sampling.The fly ash samples were collected from each APCD and weighed every hour for one day.

Analyses of PCDD/Fs
The analyses of stack flue gas and fly ash samples followed the US EPA Modified Method 23 and Modified Method 1613, respectively.Prior to analysis, each collected sample was spiked with a known amount of the 13 C 12labeled internal standard, the extract was concentrated and then treated with concentrated sulfuric acid, and this was followed by a series of sample cleanup and fractionation procedures.The eluant was concentrated to approximately 1 mL and transferred to a vial.The solution was further concentrated to near dryness, using a stream of nitrogen.Prior to instrumental analysis, a standard solution for recovery checking was added to the sample.The recoveries of PCDD/F internal standards for the tetra-chlorinated homologues through hexa-chlorinated homologues were between 52% and 97%, which met the criteria within 40-130%, while those for the hepta-and octa-chlorinated homologues were between 60% and 87%, which met the criteria within 25-130%.A high-resolution gas chromatograph/high-resolution mass spectrometer (HRGC/HRMS) was used for PCDD/F analyses.The HRGC (Hewlett Packard 6970 Series, CA, USA) was equipped with a DB-5MS fused silica capillary column (L = 60 m, ID = 0.25 mm, film thickness = 0.25 µm) 5,000 7,000 FCC: First combustion chamber; SCC: Second combustion chamber; Scrubber: Semi-Dry Scrubber; BF: Bag Filter; BF1: Bag Filter 1; BF2: Bag Filter 2.
(J &W Scientific, CA, USA), with a splitless injection.The HRMS (Micromass Autospec Ultima, Manchester, UK) as equipped with a positive electron impact (EI+) source.The analyzer mode of the selected ion monitoring (SIM) was used with a resolving power of 10,000.

PCDD/F Concentrations in the Stack Flue Gases
The PCDD/F concentrations in the stack flue gas for these two MWIs are listed in Table 2.The mean concentrations were 0.231, 0.393 and 0.076 ng I-TEQ/Nm 3 in MWI-A, and 0.065, 0.043, and 0.078 ng I-TEQ/Nm 3 in MWI-B.MWI-A had higher PCDD/F concentrations before the original bag filter system was changed to a dual-bag one.There are two possible reasons for this, the memory effect and irregular feeding.Recent investigations revealed that PCDD/F emission sources which use bag filters in APCDs may all suffer from the PCDD/F memory effect due to aged bag filters (Li et al., 2011).In order to reduce this effect, MWI-A should replace its bag filters every two years.Even by doing so, the PCDD/F emissions may still be three to four times higher than those at MWI-B, and irregular waste feeding is assumed to be the major reason for this.In May 2009, MWI-A followed our suggestion and changed the single-bag filter system to a dual-bag design to reduce PCDD/F emissions.This improved bag filter system effectively reduced the PCDD/F concentration of the flue gas to a level close to the emission limit of 0.5 ng I-TEQ/Nm 3 required by Taiwan's EPA.Wang et al. (2008) studied five intermittent incinerators and found that the PCDD/F emissions of the first sample among three consecutive stack flue gas samples collected under a stable combustion condition after start-up was two to three times higher than the average.Unfortunately, according to Taiwanese law, it was necessary to measure PCDD/F emissions before formal operations when the APCD systems were changed.The sampling at MWI-A in August 2009 was performed only three days after start-up.At that time there was still a memory effect, and this contributed to the higher emissions of PCDD/Fs.By comparison, during normal continuous operations in May 2010, the PCDD/F concentrations were reduced to 0.076 ng I-TEQ/Nm 3 .It is noticeable that by using a dual-bag filter system the PCDD/F emission level can be lowered to below the limit of 0.1 ng I-TEQ/Nm 3 that has been adopted by many developed countries (Streibel et al., 2007).
Recent research (Gao et al., 2009) reported that the stack gas emissions of PCDD/Fs from fourteen MWIs in China, which had capacities ranging from 5 to 25 ton/d, exhibited a large variation in concentration (0.08-31.60 ng I-TEQ/Nm 3 ).Six of the MWIs, which did not use activated carbon, had PCDD/F emission concentrations ranging from 0.08 to 31.60 ng I-TEQ/Nm 3 , with a mean of 9.09 ng I-TEQ/Nm 3 .In contrast, the other eight MWIs, which used activated carbon in their APCDs, had PCDD/F emissions ranging from 0.10-17.67ng I-TEQ/Nm 3 , with a lower mean of 2.07 ng I-TEQ/Nm 3 .In addition, in both the MWI and fly ash treatment plant the PCDD/F contents (5.24 ng I-TEQ N/m 3 ) in the bag filter ashes of the single-bag filter system with an ACI of 40 kg/h (Li et al., 2007) was about five times higher than that without ACI (0.97 ng I-TEQ N/m 3 ) (Chi et al., 2006).Evidently, adding activated carbon can significantly reduce PCDD/F emissions, although large variations in such emissions still remain a challenging problem.Since in Taiwan it is only necessary that PCDD/F emissions from MWI be monitored once a year, it is highly likely that the annual PCDD/F emissions data is underestimated.

PCDD/F Removal Efficiencies of the Dual-Bag Filter System
The following equations were used to calculate the PCDD/F removal efficiencies (RE) at each stage: (1) where A is the PCDD/F content of fly ashes from the scrubber times the PCDD/F generation rate of fly ashes from the scrubber, B is the PCDD/F content of fly ashes from BF1 times the PCDD/F generation rate of fly ashes from BF1, C is the PCDD/F content of fly ashes from BF2 times the PCDD/F generation rate of fly ashes from BF2, and D is the PCDD/F concentration of flue gas from stack times mean stack flue gas flow.The emission removal efficiencies for PCDD/Fs in the entire dual-bag filter system, scrubber, BF1, BF2, BF1 + BF2 in MWI-A, and the single-bag filter system, scrubber, and BF in MWI-B, are listed in Table 3.For MWI-A (with dual-bag filters) the removal efficiencies of total PCDD/Fs for the entire system, scrubber, BF1, BF2, and BF1 + BF2 were 99.7, 13.2, 84.8, 98.0, and 99.7% respectively, while for MWI-B (with a single bag filter) the removal efficiencies Scrubber: Semi-dry scrubber, BF1: bag filter 1, BF2: bag filter 2, BF: a single-bag filter system or BF1+BF2 for a dualbag system.b TeCDD: Tetrachlorodibenzo-p-dioxin, TeCDF: Tetrachlorodibenzofuran, PeCDD: Pentachlorodibenzo-p-dioxin, PeCDF: Pentachlorodibenzofuran, HxCDD: Hexachlorodibenzo-p-dioxin, HxCDF: Hexachlorodibenzofuran, HpCDD: Heptachlorodibenzo-p-dioxin, HpCDF: Heptachlorodibenzofuran, OCDD: Octachlorodibenzo-p-dioxin, OCDF: Octachlorodibenzofuran. for the entire system, scrubber, and BF were 99.1, 4.1, and 99.0%, respectively.
These results show that the dual-bag filter system has a higher PCDD/F removal efficiency than the single-bag one.With regard to health risks, the total PCDD/F I-TEQ removal efficiencies of the whole dual-bag filter system and whole single-bag filter system were 99.9% and 99.6%, respectively.Although the difference between them was only 0.3%, the emission toxicity, due to the unremoved, residual PCDD/Fs, was over four times higher in MWI-B than MWI-A (0.4% vs. 0.1%).
Recent studies ( Wang et al., 2003;Li et al., 2007;Li et al., 2007;Li et al., 2008) report that the greatest PCDD/F removal efficiencies occur in the bag filter systems of secondary aluminum smelters, fly ash treatment plants and crematories.For example, Wang et al. (2003) reported that the removal efficiency of a bag filter without ACI from a crematory was 55.1%.As seen in these results, while bag filters have the basic ability to remove fly ash and particulate phase PCDD/Fs, but their first order removal efficiency is not enough to protect public health.The present study found that the PCDD/F removal efficiency is affected by the AC content.In the single-bag filter system of a pilot plant scale test, the PCDD/F removal efficiency increased along with the amount of AC at first, but no considerable improvement was detected when AC usage was higher than 150 mg/Nm 3 (Kim et al., 2007).Nevertheless when a bag filter is combined activated carbon injection, the PCDD/F removal efficiency can be raised to 97% (Li et al., 2008).
Several novel studies on the use of dual bag filters for air pollutant control only used ACI in the second bag filter (Kim et al., 2007, andLin et al., 2008).In this study, in order to reduce the loading of AC, the bag filter system in MWI-A was modified by splitting the activated carbon injection equally into two bag filters in series.The first bag filter removed only 84.8% of PCDD/Fs, while the additional bag filter accomplished a higher removal efficiency of 98%, which contributed to the excellent overall removal efficiency of 99.9%.We speculate that the particulate phase PCDD/Fs and some gas phase PCDD/Fs were first removed by the BF1, and rest of the gas phase PCDD/Fs were adsorbed more completely by AC due to the lower fly ash load, and these were then removed by BF2.As a result, the removal efficiency decreases with the increase in chlorination in BF2.With a lower vapor pressure, the higher-chlorinated congeners condense more easily on particulates (Kaupp and Mclachlan, 1999).Since the highly chlorinated PCDDF congeners are primarily associated with particulates, the removal efficiencies of gas phase PCDD/Fs were higher than those for particle-bound PCDD/Fs for ACI (Karademir et al., 2004;Li et al., 2008).This explains why the OCDF had the lowest removal efficiencies in BF2.

PCDD/F Contents in the Discharged Ashes of the Different APCDs
The mean PCDD/F contents and their standard deviations (S.D.) in the discharged ashes of the combustion chamber, scrubber, BF1, BF2, BF1 + BF2 in MWI-A, and the singlebag filter system, scrubber, and BF in MWI-B, are listed in Table 4.The PCDD/F content in the fly ash, which was  (Yan et al., 2007;Bie et al., 2007;Chen et al., 2008;Zheng et al., 2010;Wu et al., 2011).These results provide evidence that AC is an ideal adsorbent for the removal of PCDD/Fs in MWIs.In this study, even though the AC may account for only 11% or less of the mass proportion of the fly ash in BF1, the PCDD/F contents are significantly higher than in the scrubber, implying that AC has much greater efficiency than the scrubber in terms of PCDD/F removal.Both MWI-A and MWI-B were equipped with systems which were able to undertake two-stage thermal processes with self-monitoring.The most notable reduction in PCDD/F content was found in the bottom ashes.The mean PCDD/F contents and their S.D. in the bottom ashes of MWI-A and MWI-B ranged from 0.24-0.34ng I-TEQ/g, with a mean of 0.30 ng I-TEQ/g (RSD: 16.7%), and 0.36-1.13ng I-TEQ/g, and 0.74 ng I-TEQ/g (RSD: 52.7%), respectively.Grochowalski (1998) reported that the bottom ashes taken from 18 MWIs which did not use two-stage thermal processes emitted PCDD/Fs at a level between 8 to 45 ng I-TEQ/g.In Taiwan, government policy with regard to incineration residues advocates their reuse as road subbases or secondary building materials, provided that the total PCDD/F content is below the legal limit (1 ng I-TEQ/g).Therefore, bottom ashes with total PCDD/F content below this limit can be collected and legally reused.
The formation of PCDD/Fs has two main pathways, i.e., de novo synthesis and precursor condensation.Huang and Buekens (1995) extensively reviewed papers regarding the formation mechanisms of PCDD/Fs, and concluded that de novo synthesis can produce PCDD/Fs with a PCDFs/PCDDs ratio < 1. Everaert and Baeyens (2002) also indicated that precursor formation occurs mainly at temperatures ranging from 300 to 600°C, with the typical PCDFs/PCDDs ratio > 1.The PCDFs are more dominant in the scrubber, BF1 and BF2 ashes (PCDFs/PCDDs ratio = 4.01-4.63)than in the bottom ashes (2.73) (see Table 1).At both MWI-A and MWI-B the operations were controlled at temperatures higher than 750°C (average 800°C) in the first combustion chamber, and more than 1000°C (average 1200°C) in the secondary one.These high temperatures effectively reduce PCDD/Fs and avoid incomplete combustion.Consequently, the increased PCDFs/ PCDDs ratio implies that de novo synthesis occurs as the flue gas cools down.

PCDD/F Congener Profile in the Discharged Ashes of Different APCDs
The congener profiles of the seventeen 2,3,7,8-chlorinated substituted PCDD/Fs detected in the discharged ashes from different APCDs are shown in Fig. 1.The profiles were calculated based on the fraction (%) of each congener and the total PCDD/F mass concentration.The congener profile of the fly ash obtained from bag filter system was dominated by 1,2,3,4,6,7,9-HpCDF, while the bottom ash was dominated by OCDF and OCDD.The results indicate that the lower chlorinated-substituted PCDD/F congeners, having higher vapor pressures, are vaporized and complete destroyed by the two-stage thermal process.Also, the major PCDD/F congeners adsorbed by AC were formed via de novo synthesis during the cooling of the flue gases in MWI-A and MWI-B.The I-TEQ contribution profiles of fly ashes were dominated by 2.3.4.7.8-PeCDF (see Fig. 2).The higher I-TEQ factor (0.5) was a result of the high contribution of 2.3.4.7.8-PeCDF, and the high removal efficiency (over 99%) of 2.3.4.7.8-PeCDF by AC raised the toxicity removal efficiency to 99.8% in the dual-bag filter system.

PCDD/F Contribution Fractions and Emission Factors of the Different Output Media
Regarding the PCDD/F I-TEQ contribution fractions in the bottom ash, the discharged ashes of scrubber, BF1, BF2 and in the stack flue gases and bottom ash of MWI-A, and the discharged ashes of scrubber, BF and in the stack flue gases of MWI-B (see Table 5) were 1.95%, 13.2%, 72.8%, 12.0%, 0.100%, 12.8%, 23.2%, 63.7% and 0.281%, respectively.The dual-bag filter system can stop FCDD/F from escaping as stack flue gases at a greater efficiency than seen with the single-bag filter systems.Although the emission factors of stack flue gases in both bag filter systems appear similar (0.002 mg I-TEQ/ton, respectively), we speculate that a key factor is the kind of medical waste that is being disposed of.Comparing the emission factors found in this study to those of other emission sources (Wang et al., 2003;Lin et al., 2007), those in MWI-A and MWI-B were not particularly high when considered alongside those of iron ore sintering (0.002 mg I-TEQ/ton) and inter plants (0.00066-0.0031mg I-TEQ/ton), and both were lower than those found at another MWI (0.0046 mg I-TEQ/ton) in Taiwan.In addition to using a dual-bag filter system, Wang et al. (2003) reported that sinter plants which carry out selective catalytic reduction (SCR) can obtain a PCDD/F emissions factor (0.970 µg TEQ/ton) that is over three times lower than that seen with other systems (3.13 µg TEQ/ton) that do not use SCR.After changing to a dual-bag filter system, MWI-A had a lower I-TEQ in percentage terms in its stack gases.Consequently, although a significant amount of polyvinyl chloride (PVC) may exist in medical waste, and therefore may also serve as PCDD/F precursors during unstable combustion, the dual-bag filter system can still capture PCDD/Fs effectively.

CONCLUSIONS
This study provides a set of guidelines to improve the PCDD/F removal efficiencies in MWIs by splitting the  without resorting to extra use of activated carbon, the dualbag filter system in MWI-A achieves a higher overall removal efficiency (99.7%) than the single-bag filter system used in MWI-B (99.1%).Most importantly, the unremoved residual total I-TEQ data show that the dual-bag filter system has much lower toxicity (100% -99.9% = 0.1%) than the single-bag system (100% -99.6% = 0.4%).

Table 1 .
Basic information concerning the single-and dual-bag filter system MWIs.

Table 2 .
PCDD/F Concentrations in the flue gases of the MWI-A and MWI-B, respectively.MWI-A has a dual-bag filter system with ACI (3 kg/h), MWI-B has a single bag filter system with ACI (3 kg/h). a

Table 3 .
PCDD/F removal efficiencies in each APCD of the single-and dual-bag filter systems, respectively.

Table 4 .
PCDD/F content (mean ± S.D.) in the discharged ashes of the APCDs.

Table 5 .
The emission factors and the PCDD/F I-TEQ contribution fractions of the output media. in a traditional single bag filter into two bag filters in series.While the total consumption of the activated carbon remains the same, the overall removal efficiency is greatly enhanced.Based on the mass balance results, the scrubber can block over 11% of the PCDD/Fs which are formed during combustion and the cooling of the flue gases in MWI-A.Comparing the removal efficiencies of PCDD/Fs in both bag filter systems, it is clear that,