Environ. Sci. Technol. 2003, 37, 3152-3157
Incineration of Different Types of Medical Wastes: Emission Factors for Particulate Matter and Heavy Metals MARIA C. M. ALVIM-FERRAZ* AND S EÄ R G I O A . V . A F O N S O LEPAE, Departamento de Engenharia Quı´mica, Faculdade de Engenharia, Universidade do Porto, R. Dr. Roberto Frias, 4200-465, Oporto, Portugal
Previously published results for emission factors of medical waste incineration do not include enough information about the incinerated waste composition. This paper reports the first emission factors estimated for particulate matter, As, Cd, Cr, Pb, Mn, Hg, and Ni, considering that medical waste is segregated in different types according to Portuguese legislation. The main purpose was to evaluate the influence of incinerated waste composition and segregation practice on emission factors. One “controlledair” incinerator without air pollution control devices was used for the incineration either of mixtures with a defined composition or of a specific waste type. Previously published emission factors are not associated with the composition of the incinerated mixture, and the results showed that the usefulness of those emission factors is very doubtful. The existence of different waste classifications also reduces the usefulness of previously published results. To protect human health, appropriate equipment to control atmospheric pollutants must be used, since the legal limits for pollutant concentrations were strongly surpassed (226 times higher than the limit for Hg), with risks for patients and workers of the hospital and exposed population. It was concluded that rigorous segregation practices and adequate management methodologies allow reducing 80% of the amount of wastes that must be incinerated, practically eliminating Hg and Pb emissions and reducing those of PM, As, Cd, Cr, Mn, and Ni, respectively, 98, 90, 92, 84, 77, and 92%.
Introduction The European Union has been making a special effort to standardize medical waste classification through the establishment of the Waste European Catalog. Meanwhile, many different classifications are yet to be considered in legislation and literature (1-4). For instance, while in the United States the segregation of medical waste is made usually considering three types (red bag, pathological, and general), the Environmental European Agency identifies “specific hospital waste” and “other hospital waste”, and the Portuguese Legislation settles for the following four types: (i) Group I - wastes similar to municipal ones; (ii) Group II - nonhazardous medical wastes that do not require specific treatment and can be considered similar to municipal wastes; * Corresponding author phone: 351-22-5081688; fax: 351-225081449; e-mail:
[email protected]. 3152
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
(iii) Group III - medical wastes with a biological risk that must be pretreated before elimination as municipal wastes; (iv) Group IV - specific medical wastes with compulsory incineration. These different classifications cause some confusion regarding the different meanings of similar terms and the different terms used for classification of similar waste types. Incineration is the most widely used treatment of medical waste. Three types of incinerators are being used: “multiplechamber”, “controlled-air”, and “rotary kiln”. “Controlledair” is the most extensively installed, the main disadvantage being the emission of pollutants to the atmosphere. Polychlorinated dibenzodioxins, dibenzofurans, and biphenyls and polycyclic aromatic hydrocarbons are the most toxic emissions. Particulate matter (PM) and heavy metals are also being considered of special environmental and health significance (5, 6). Previous research works showed that without control of atmospheric pollutants, the incineration of medical waste does not comply with the legal emission limits, even when correct practices of operation and maintenance are used (7). Thus, to protect public health, hospital incinerators should be provided with air pollution control devices (APCDs) to reduce atmospheric emissions. As most hospital incinerators do not possess such equipment, efficient methodologies should be developed in order to evaluate the safety of incineration procedures, through the comparison with legal limits. Emission factors have been used for different applications, allowing an easy estimation of emission rate and concentration of emitted pollutants (8-11). The emission factors for the incineration of medical waste allow an easy comparison of pollutant concentration in combustion gas with the respective legal limits (12, 13). It is well-known that emission factors strongly depend either on the process and conditions of incineration or on the waste type. Nevertheless, published values are generally averages from several hospitals, which means that they are not necessarily representative of a particular hospital. Moreover, to the best of our knowledge, just two publications presented emission factors for specific waste types, taking into account, nevertheless, different waste classifications. Lee et al. studied polycyclic aromatic hydrocarbons on emissions considering “special” and general wastes (14). Walker and Cooper analyzed a more complete range of pollutants according to a waste classification in three types: red bag, pathological, and general, this one including red bag and a not defined amount of municipal waste (1). Incineration practices greatly depend on the hospital concerned (most Portuguese hospitals incinerate groups III and IV and an unknown proportion of groups I and II), which generally leads to a poor knowledge of the incinerated mixture composition. Emission factors published by the EPA (15) and EEA (4) considered an incinerated mixture that also includes all waste types not specifying the proportion of each type in the incinerated mixture. Therefore, the knowledge related with emission factors for medical waste incinerators is still very scarce, mainly because they do not include enough information about the incinerated waste types and the respective proportion in incinerated mixture. The existence of different waste classifications also reduces the usefulness of previously published results.
Materials and Methods The study reported here was carried out from January till December 1999, in a 300 bed hospital, at the time repre10.1021/es026209p CCC: $25.00
2003 American Chemical Society Published on Web 05/21/2003
TABLE 1. Daily Production of Each Medical Waste Type: Average Weight (kg (bed.day)-1) and Percentagea group I+II
present study PERH (1999) Alvim Ferraz (2000) a
USUSeg RIGSeg
group III
group IV
DWPI+II
(%)
DWPIII
(%)
DWPIV
(%)
total DWP
1.9 (12%) 1.9 (9%) 2.1
49 49 61
1.6 (11%) 1.9 (10%) 1.3
41 49 37
0.4 (10%) 0.08 (8%) 0.08
10 2 2
3.9 3.9 3.5 3.8
Averages of 20 tests, relative standard deviation in parentheses.
sentative for Portugal, of the handling, storage, treatment, and ultimate disposal of medical waste. To assess the production of the different medical waste types, the daily waste production (DWP) of each one was estimated, based on the weight of the respective bags. Aiming to analyze the influence of segregation practices on the production of each type, two experimental sets were performed of 20 tests each (3 months), considering different practices of waste segregation. The wastes were treated by incineration, landfilling resultant ashes. The most usual incineration practice in Portugal was used, performing the tests in one “controlledair” incinerator without (APCDs), with a capacity of 2 × 180 kg h-1. Correct practices of maintenance and operation were employed; according to legislation, the incineration temperature, residence time, gas turbulence, and mixing of solids were controlled (12, 13). The main factor affecting pollutant emissions is temperature, maintained at 890 and 1100 °C, respectively, in the primary and secondary combustion chambers. The influence of incinerated waste composition, waste type, and segregation practice on particulate matter and metal emissions (As, Cd, Cr, Pb, Mn, Hg, and Ni) was analyzed. For that, four incineration scenarios were planed, each one considering the incineration either of mixtures with a defined composition or of a specific waste type. Nine tests were realized for each scenario, registering waste feed. To avoid memory effects in incineration process, two charges of similar wastes were incinerated before the first test. Temperature, pressure, flow rate, and water content were measured in the combustion gas. The concentration of particulate matter was measured according to standard procedures (16). Sampling was performed for 8 h, withdrawing isokinetically a partial volume from the flue gas, to get a representative sample for analysis. Particulate matter was collected using glass fiber filters (0.3 µm pore size). After that, the particulate matter was completely recovered, dried, and weighed. This method can be used to determine concentrations ranging from 0.005 to 10 mg m-3. Under ideal conditions, the inaccuracy of the method is about (10% for concentrations above 0.050 mg m-3. Metal concentrations in the combustion gas, either in particulate matter or in the gas phase, were estimated according to standard procedures (17). After particulate matter collection, gaseous emissions passed through an aqueous acidic solution of hydrogen peroxide (analyzed for all metals) and an aqueous acidic solution of potassium permanganate (analyzed for mercury). The recovered samples were digested to be analyzed. Cd, Cr, Pb, Mn, and Ni were quantified by Atomic Absorption Spectroscopy (AAS) using direct aspiration. As and Hg were also quantified by AAS, using, respectively, graphite furnace and cold vapor generation techniques. Standardization and calibration (of sampling train and spectrometer) were made according to standard procedures as well as the preparation of field blanks (17). The detection limits for Cd, Cr, Pb, Mn, Ni, As, and Hg were, respectively, 5, 50, 100, 10, 40, 0.1, and 1 ng cm-3. The blank
concentrations were always lower than the detection limits. To check results accuracy, recovery assays were carried out spiking subsamples with known quantities of the analyte (17). The recoveries were between 92 and 102%. Each average was based on at least three analyses, repeated till it reached a relative standard deviation lower than 5%. The calculations were based on the measuring of combustion gas flow rates and respective pollutant concentration. Emission rates (ER) and emission factors (EF) were calculated, considering the waste feed during sampling and the sampling time or the waste rate (see eqs 1 and 2, Supporting Information). The yearly emitted pollutant (YEP) can be estimated through the daily incinerated waste, using eq 3 of the Supporting Information. The pollutant concentration and the volume of combustion gas produced per mass of waste were compared with legal limits and published data. For that, they were converted to dry standard conditions (273 K and 101.3 kPa) referred to 11% O2 in combustion gas, being designated by Cstd@11%O2 and Vstd@11%O2 (12, 13) (see eqs 4 and 5, Supporting Information). The volume of combustion gas actually produced can be estimated throughout the ultimate composition of medical waste and the percentage of oxygen present in combustion gas (see eqs 6-8, Supporting Information) (1). To evaluate the accuracy of this procedure, the estimated values were compared with experimental ones, both converted to dry standard conditions referred to 11% O2 in combustion gas.
Results and Discussion Two experimental sets of 20 tests each were planed in order to assess the influence of waste segregation practice on the production of the different medical waste types. During the first one, carried out from March till May 1999, the wastes were collected according to the usual segregation practice previously used at the hospital (USUSeg). The second experimental set was performed from September till November 1999, after a sensitization program to stimulate hospital workers for a much more rigorous practice of segregation (RIGSeg). For the two experimental sets, the averages of the daily waste production of each waste type (DWPI+II, DWPIII, and DWPIV), normalized to the number of beds (relative standard deviation in parentheses), and the respective percentages, are compared in Table 1 with values previously published. Comparing the results of the two experimental sets with different segregation practices (USUSeg and RIGSeg), it can be observed that, for RIGSeg, group IV decreasing is reached at the cost of group III increasing. These results also allow for concluding that the amount of wastes that must be incinerated with the special requirements needed for medical waste (group IV) can be reduced 80% using a rigorous segregation practice. Despite the bigger percentage of groups I and II estimated in PERH (18), a good agreement is observed between the total amounts estimated for all the studies compared here. This probably means that the segregation practices justify the differences observed on the percentages. VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3153
The four incineration scenarios, planed to analyze the influence on emissions of incinerated waste composition, waste type, and segregation practice, were carried out considering the incineration either of mixtures of different waste types or of a specific waste type. The published results for emission factors, corresponding to the incineration of mixtures of different waste types, do not include enough information about the incinerated waste types and the respective proportion in the mixture (4, 15). For assessing how important the missing of this information is, the published results were compared with that determined for a representative mixture of the incineration practice in Portugal. This mixture had the following specified composition: group III (58.0%), group IV (14.5%), and groups I and II (27.5%), corresponding to the total amount of groups III and IV and to 40% of the total amount of groups I and II. To evaluate the relative contribution of groups III and IV, their incineration was considered separately, after the usual practice of segregation to be more representative. According to the legislation just group IV must be incinerated; therefore, to assess the influence of segregation practice, the incineration of group IV rigorously segregated was also studied. The following designations were considered: (i) INCMIX-spec, incineration of a mixture with a specified composition; (ii) INCIII-USUSeg, incineration of group III using the usual segregation practice; (iii) INCIV-USUSeg, incineration of group IV using the usual segregation practice; (iv) INCIV-RIGseg, incineration of group IV using the rigorous segregation practice. Moisture contents (as fired) of INCMIX-spec, INCIII-USUSeg, INCIV-USUSeg, and INCIV-RIGseg were, respectively, 16, 12, 27, and 85%; heat values (as fired) were, respectively, 22, 25, 20, and 2.3 MJ kg-1. The volumes of combustion gas produced per mass of waste (Vstd@11%O2), calculated through eq 5 of the Supporting Information, for INCMIX-spec, INCIII-USUSeg, INCIV-USUSeg, and INCIV-RIGseg are, respectively, 12.8, 12.8, 10.3, and 11.0 m3 kg-1, showing that they do not depend too much on the type of waste incinerated. This volume range includes the experimental value published by Walker and Cooper (11.2 m3 kg-1) and the value published by Alvim Ferraz et al. (11 m3 kg-1) estimated using eqs 6-8 of the Supporting Information and the generic ultimate composition of medical waste (1, 7). For the experimental conditions considered in the present study, the results show that a reasonable prevision of combustion gas volume can be possible through the simple utilization of the generic ultimate composition of medical waste. Equations 1-4 of the Supporting Information were used for the estimation of Cstd@11%O2, ER, EF, and YEP. The average of the nine tests made for each incineration scenario (relative standard deviation in parentheses) is expressed in Table 2 for the four scenarios. A correspondence between Portuguese and Walker and Cooper classifications can be approximately made to compare the present results with previously published ones. For that, red bag and pathological wastes were considered similar, respectively, to groups III and IV. However, a correspondence for general waste of Walker and Cooper classification is not so acceptable, according to the missing information about the proportion of municipal wastes (equivalent to groups I plus II) included in general wastes. Aiming to include the emission factors published by the EPA in the comparison, a new incineration scenario must be defined corresponding to the incineration of mixtures composed by different groups in a nonspecified proportion (INCMIX-notspec). Therefore, the incineration of general wastes of Walker and Cooper classification and of the mixture of wastes referred in the EPA publication belongs to this scenario. The results observed in Table 2 show that Cstd@11%O2, ER, EF, and YEP greatly depend on the type of waste incinerated. 3154
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
For some pollutants the concentrations estimated in this study were higher than those published by the EPA, enhancing the influence of specific waste constitution on the emitted concentrations (19). According to the legislation, just the group IV must be incinerated. To evaluate how emissions are influenced when other groups are incinerated, the concentration ratios relative to the concentration observed when just group IV is incinerated (CRrel-IV) were estimated. The CRrel-IV values of Table 3 show that while the highest concentrations of As, Cd, Cr, Mn, and Ni were observed for INCIV-RIGSeg (incineration of group IV collected using the rigorous segregation practice), this scenario is the one with the lowest concentration of PM, Pb, and Hg. According to this, the existence of group III mixed with group IV for INCIV-USUSeg (incineration of group IV collected using the usual segregation practice) justifies the decreasing of As, Cd, Cr, Mn, and Ni concentration and the increasing of PM, Pb, and Hg concentration. To evaluate if legislation was obeyed, the concentration ratios relative to the limit value (CRrel-lim) were estimated. According to Portuguese and European legislation for PM, the limit values of emission concentrations for daily and 30 min averages are, respectively, 10 and 30 mg m-3. For Hg the limit value for the 8 h average is 0.05 mg m-3. The limits for the other heavy metals were settled considering the sum of concentrations of different metals (12, 13). The results for CRrel-lim are shown in Table 3. For PM they were estimated relative to daily and 30 min legal limits, since the sampling time (8 h) did not obey those times settled in legislation. If a limit was established for 8 h it will be surely between the daily and 30 min limits, which means that CRrel-lim will be between the values expressed in Table 3. Therefore, it can be concluded that the legal limit values for PM were greatly surpassed. For Hg the concentrations were 1.3 to 226 times higher than the limit, the highest corresponding to INCIII-USUSeg. Cd concentrations were also clearly higher than the limit settled for the sum of Cd and Tl concentrations (0.05 mg m-3). Some of the other heavy metals considered in legislation were not quantified. Nevertheless, just the sum of concentrations of quantified metals was 3 to 8 times higher than the legal limit (0.5 mg m-3). These results confirm that the incineration of medical waste without control of atmospheric pollutants does not obey the legal emission limits, even when correct practices of operation and maintenance are used, meaning that appropriate equipment to control atmospheric pollution must be used to protect human health (7). Emission factors shown in Table 2 have the smallest relative standard deviation for the incineration of group IV collected using the rigorous segregation practice, which is related to a bigger homogeneity of the respective wastes. Tables 2 and 3 show that emission factors are clearly dependent on the waste type. For the incineration of group IV collected using the rigorous segregation practice, they are the biggest for As, Cd, Cr, Mn, and Ni, while they are the smallest for PM, Pb, and Hg. INCIII-USUSeg has the smallest emission factor for As, Cd, Cr, Mn, and Ni and the biggest for PM, Pb, and Hg. According to this, the mixture of group III, in the incineration of group IV collected using the usual segregation practice, is responsible by the decreasing of emission factors of As, Cd, Cr, Mn, and Ni and by the increasing of emission factors of PM, Pb, and Hg. Table 2 shows that there is a very good agreement between the emission factors of this study and the published study by Walker and Cooper (1) for INCIII-USUSeg and INCIV-RIGSeg, which means that the correspondence established, between red bag waste and group III and pathological waste and group IV, is acceptable. A similarity between the results of this study for INCMIX-spec and those of the other studies obtained for INCMIX-notspec (Walker and Cooper and EPA) can be observed
TABLE 2. Pollutant Concentrations, Emission Rates, Emission Factors (Relative Standard Deviation in Parentheses), and Yearly Emitted Pollutant (Averages of Nine Tests) INCMIX-spec
Cstd@11%O2 (mg m-3) ER (kg h-1) EF (kg ton-1) YEP (kg year-1) Cstd@11%O2 (mg m-3) ER(g h-1) EF(g ton-1) YEP (kg year-1) Cstd@11%O2 (mg m-3) ER (g h-1) EF (g ton-1) YEP (kg
year-1)
Cstd@11%O2 (mg m-3) h-1)
ER (g EF (g ton-1) YEP (kg
year-1)
Cstd@11%O2 (mg m-3) ER (g h-1) EF (g ton-1) YEP (kg year-1) Cstd@11%O2 (mg m-3) ER (g h-1) EF (g ton-1) YEP (kg year-1) Cstd@11%O2 (mg m-3) ER (g h-1) EF (g ton-1) YEP (kg year-1) Cstd@11%O2 (mg m-3) ER (g h-1) EF (g ton-1) YEP (kg year-1)
INCMIX-notspec
INCIII-USUSeg
INCIV-USUSeg
INCIV-RIGSeg
PM present study EPA (1994) present study present study Walker (1992) EPA (1999) present study
422
503
443
372
0.909 6.23 (12%) 6.58
0.802 4.42 (11%)
0.671 4.00 (6.1%) 4.24
1090
194
35.0
0.00483
0.0166
0.0338
0.00874 0.0599 (11%) 0.0573
0.0301 0.166 (10%)
0.0610 0.364 (6.9%) 0.339
0.0105
0.00728
0.00319
0.147
0.335
0.585
0.266 1.82 (12%) 1.68
0.608 3.35 (11%)
1.06 6.29 (6.4%) 6.15
0.319
0.147
0.0551
0.0153
0.159
0.364
0.0276 0.189 (11%) 0.202
0.288 1.59 (11%)
0.658 3.92 (6.9%) 3.9
0.0331
0.0695
0.0343
4.03
2.43
0.405
7.29 49.9 (11%) 50.8
4.40 24.2 (10%)
0.730 4.35 (7.3%) 4.31
8.75
1.06
0.0381
0.00784
0.208
0.550
0.0142 0.0970 (10%) 0.0987
0.377 2.08 (11%)
0.992 5.91 (6.8%) 5.50
0.0170
0.0910
0.0518
11.3
4.06
0.0649
20.5 140 (11%) 110
7.35 40.5 (11%)
0.117 0.699 (6.8%) 0.637
24.6
1.78
0.00612
0.0110
0.0234
0.0415
0.0199 0.136 (11%)
