Emission of Polycyclic Aromatic Hydrocarbons, Toxicity, and

mutagenicity (Ames test). The gas phase of smoke contributed g95% of 17 PAH, g96% of toxicity, and g60% of mutagenicity. The highest emission factor o...
0 downloads 0 Views 76KB Size
Environ. Sci. Technol. 2002, 36, 833-839

Emission of Polycyclic Aromatic Hydrocarbons, Toxicity, and Mutagenicity from Domestic Cooking Using Sawdust Briquettes, Wood, and Kerosene NGUYEN THI KIM OANH,* LE HOANG NGHIEM, AND YIN LATT PHYU Environmental Engineering, SERD, Asian Institute of Technology, Klongluang, Pathumthani 12120, Thailand

Smoke samples, in both gas and particulate matter (PM) phases, of the three domestic stoves were collected using U.S. EPA modified method 5 and were analyzed for 17 PAH (HPLC-UV), acute toxicity (Microtox test), and mutagenicity (Ames test). The gas phase of smoke contributed g95% of 17 PAH, g96% of toxicity, and g60% of mutagenicity. The highest emission factor of 17 PAH was from sawdust briquettes (260 mg/kg), but the highest emission of 11 genotoxic PAH was from kerosene (28 mg/ kg). PM samples of kerosene smoke were not toxic. The total toxicity emission factor was the highest from sawdust, followed by kerosene and wood fuel. Smoke samples from the kerosene stove were not mutagenic. TA98 indicated the presence of both direct and indirect mutagenic activities in PM samples of sawdust and wood fuel but only direct mutagenic activities in the gas phase. TA100 detected only direct mutagenic activities in both PM and gasphase samples. The higher mutagenicity emission factor was from wood fuel, 12 × 106 revertants/kg (TA100-S9) and 3.5 × 106 (TA98-S9), and lower from sawdust, 2.9 × 106 (TA100-S9) and 2.8 × 106 (TA98-S9). The low burning rate and high efficiency of a kerosene stove have resulted in the lowest PAH, toxicity, and mutagenicity emissions from daily cooking activities. The bioassays produced toxicity and mutagenicity results in correspondence with the PAH content of samples. The tests could be used for a quick assessment of potential health risks.

Introduction Domestic combustion is one of the major sources of indoor air pollution in developing countries and has been identified as a serious health hazard affecting hundreds of millions of people, especially women, children, and the elderly. Cooking smoke has been shown to cause respiratory diseases, such as chronic bronchitis, emphysema, expectorative cough, and dyspnea (1). Exposure to unvented indoor cooking smoke may cause cancer, particularly lung cancer. Results from a study program (2) in rural Xuan Wei County, China, have shown that the high lung cancer mortality rates are associated with exposure to unvented emissions from the indoor burning of smoky coal for cooking and heating. * Corresponding author phone: 662 524 5641; fax: 662 524 5625; e-mail: [email protected]. 10.1021/es011060n CCC: $22.00 Published on Web 01/18/2002

 2002 American Chemical Society

The emission of pollutants from domestic cooking is mainly the result of incomplete combustion of fuels and is a function of many variables, including, among others, fuel and stove types, air supply conditions, and the frequency and duration of stove use. The common use of low quality noncommercial fuels (fuel wood, agroresidue, dung cakes, etc.) in rural areas of developing countries may result in high pollution emissions. Residential stoves may produce higher emissions of some important air pollutants as compared to industrial combustion in many circumstances. The main pollutants in smoke from domestic cookstoves are particulate matters (PM), carbon monoxide (CO), and organic compounds. The latter consists of a wide range of substances. One study alone (3) has identified more than 180 polar, 75 aliphatic, and 225 aromatic compounds in wood smoke. Yet, there are still many unidentified compounds. Among the organic compounds emitted, of special interest are polycyclic organic matter (POM) and formaldehyde (4). POM is a chemical group that contains two or more benzene rings. One particular set of POM that are known to be carcinogens or mutagens are the polycyclic aromatic hydrocarbons (PAH). Most PAH in the atmosphere are released from the incomplete combustion of organic materials. In the atmosphere, PAH undergo transformations, and the derivatives are usually more toxic than PAH themselves (5), therefore increasing the potential harmful effects on human health. Besides PAH, the aza and imino arenes in POM have also been found to be potentially carcinogenic (3). Organic extracts of smoke samples from domestic combustion sources have demonstrated genotoxic and carcinogenic activities in experiments (3, 6). The indoor pollution concentration and exposure level depend not only on emission but also on ventilation conditions and on whether cooking stoves are physically isolated from main living areas. Much of developing Asia’s population cooks and heats using unvented stoves in poorly designed kitchens. As a result, the air pollution levels in Asian homes often exceed World Health Organization standards for ambient outdoor air as well as that for typical indoor levels in developed countries (7). Human exposure to cooking smoke has received increasing attention in indoor air pollution research. High consumption rates, low quality of fuels, and low efficiency of cookstoves used for domestic cooking in developing countries plus poorly designed kitchens lead to high health risks from cooking smoke. Yet, only limited data on PAH emissions and the toxic effects of cooking smoke are available in these countries. In this study, PAH emissions and potential toxicity and genotoxicity were monitored for three common domestic fuel-stove systems in Asia, namely, kerosene, sawdust briquettes, and wood. The purpose was to estimate the potential health risks associated with cooking activities and to identify safer cooking stove-fuel systems.

Materials and Methods Fuel Combustion and Source Sampling. Kerosene was purchased from a local market in Pathumthani, Thailand. The fuel was burned in a Vietnamese kerosene cookstove, which is of the wick type with eight wicks. The sawdust was purchased from a timber processing enterprise in Thailand, and the briquettes were prepared at the laboratory at the Asian Institute of Technology (AIT), by compressing sawdust alone at 250-300 °C using a screw type briquetting machine. The final briquettes are hollow, of a ring shape with outer diameter of 55-60 mm, an inner diameter of 15 mm, and a length of 25 mm. This fuel was VOL. 36, NO. 5, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

833

TABLE 1. Summary of Fuels, Stoves, and Source Sampling sample information fuel information types sawdust briquettes

ash gross heat moisture amount (%) (MJ/kg) (%) (kg) 2.5 21 (23)

wood fuel (Pterocarpus 2 indicus)

19a

kerosene (Thai market) 0.78 kg/L

41 (24)

a

4.5

10

250 283 276 287 293 290 360 511 540

5.679 6.96 7.904 7.971 7.925 7.485 9.852 11.162 13.324

90 92 97 98 98 90 103 91 98

73 62 63 71 74 73 67 70 67

92 84 85 143 152 133 2.5 5 1.8

Determined at AIT.

burned in a Vietnamese double-skinned cylindrical cookstove with a metal outer cover and a ceramic liner. The stove is of 25 cm height, 20 cm outer diameter, and 15 cm inner diameter. The fuel wood was in sticks (20-25 cm long, 3-4 cm thick) produced from logs of Pterocarpus indicus tree. The tree is naturally grown in tropical forests and can be used for timber and fuel. This fuel wood was burned in a Thai singlestage ceramic cookstove, which is cylindrical in shape with a ceramic ash-insulating layer. The stove is of 25 cm height, 28 cm outer diameter, and 20 cm inner diameter and is on a metal stand of 10 cm height. The stove has an opening 15 cm wide and 8 cm high for fuel stocking. Related information on fuel and cookstoves is provided in Table 1. The kerosene fuel was burning without stoking during the test. Sawdust briquettes and wood fuel were stoked through an opening on the periphery of the fuel beds with minimum disturbance to the existing fire. The cookstoves were ignited in the hood. Small wood sticks were used to start fire for sawdust briquettes and wood fuel. The two fuels were ignited fast, and fire was sustainable after less than 2 min. The sampling period covered the whole burning cycle from the moment of starting the fire to the end of the burning process, when visually the char burning stopped (sawdust briquettes and wood fuel). A hood, thermally insulated, was used to capture flue gases from the cookstoves. The hood and source sampling method used were the same as presented in the previous study (8). Samples were collected isokinetically using an Anderson-Graseby Auto5 semi-VOST in accordance to U.S. EPA modified method 5 (9). The small stack diameter (11 cm) could accommodate only one traverse point with the position determined by U.S. EPA method 1 (10). The sampling probe was thus fixed, and the sampling port was closed using metal-lining insulation material to minimize disturbances to gas flow. The sampling train consisted of an Amberlite resin XAD-2 (styrene divinylbenzene polymer) trap, containing 30 g of XAD-2 to adsorb gaseous organic compounds, which was followed by four impingers. The first impinger was the condensate knockout. The second impinger, of the standard Greenburg-Smith design, contained 100 mL of distilled and deionized water. Both the third and fourth impingers were of the modified Greenburg-Smith type. The former was empty while the latter contained 350 g of activated silica gel. During the sampling, the filter box and the sampling probe were heated at 125 °C. The sampling port was located at 4 m above the fire level. The average flue gas velocity at the sampling port was around 4.0-4.6 m/s. Taking into consideration of the configuration of the hood, the average residence time of flue gas in the hood before sampling was around 5-7 s. No spillage of the smoke or flame blown-away was observed. 834

stove

2.004 Vietnamese cylindrical 2.06 briquette stove, 1.957 η ) 25-30% (25) 3.614 Thai traditional wood 3.834 cookstove (single-stage), 3.59 η ) 25-30% (25) 0.624 Vietnamese kerosene 1.014 cookstove (wick), 0.858 η ) 50% (24)

vol, isokinetic collected time (m3, rate stack gas PM (min) (20 °C, 1 atm) (%) (°C) (mg)

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 5, 2002

For each fuel-stove system, three burning batches were conducted. For each batch, two smoke samples were collected: a particulate matter (PM) and a gas-phase sample of organic compounds. The PM phase sample includes the PM on the filter paper (Whatman glass microfiber filter, 934AH, Cat. No. 1827 110) and PM rinsate, consisting of all of the PM recovered from the probe nozzle, sampling line, and other parts of the sampling train to the front half filter holder, inclusively. The gas-phase sample consists of XAD-2, the content of the condensate knockout, and the rinsate (dichloromethane) of all of the parts from the second half of the filter holder to the second impinger. More sampling data are given in Table 1. Sample Preparation and Analytical Procedure. The PM phase was extracted in combination with the PM rinsate. For the gas phase, the condensate was first extracted. The extract was then combined with XAD-2 and the gas-phase rinsate. Half of the extractable organic matters (EOM) were used for Microtox toxicity and Ames tests, and the other half were used for PAH analysis (Figure 1). PAH Analysis. The sample preparation procedure as well as all materials used for PAH analysis was the same as presented in the previous study (8). Reagents, filter, and XAD-2 blanks were tested for possible contamination. The analytical method was based on U.S. EPA method TO-13 (11). A high-performance liquid chromatograph, HP 1050, was used with a reversed-phase column (with guard column) from Hewlett-Packard Co. (Palo Alto, CA) and specified for PAH analysis (VYDAC 201 TP5 C-18 RP, 0.46 × 25 cm). A UV detector (HP 1050), working at λ ) 254 nm, was used to detect PAH compounds. A mixture of 17 PAH, including 16 U.S. EPA priority PAH, plus BeP, was used as the external standard for quantitative analyses. Four different dilutions of the standard mixture were used that covered the range of PAH present in the samples. Each concentration was injected 3 times. A good linear correlation between the compound concentrations and peak areas was found with R2 values in the range 0.980.99 for all 17 PAH compounds. The minimum detectable quantity of the method was determined and presented in Table 2 along with the analytical results. Microtox Toxicity. The Microtox test is a quick bioassay using marine bioluminescent bacterium (Photobacterium phosphoreum) as the test organism. The light emitted by the bacteria after they have been exposed to a sample of unknown toxicity is compared with the light emitted by the bacteria in the controls containing no sample. The decrease in the light emission due to the exposure, which is a function of metabolic inhibition in the bacteria, indicates the degree of toxicity of sample. The test is simple, fast, and follows standardized procedures. Nevertheless, the test bacteria have most often been shown to classify toxic substances in the same way as higher organisms (12). A study (13) reported a

a Names of compounds with minimum detectable quantity by UV detector in parentheses. NAPH, naphthalene (0.41 ng); ACY, acenaphthylene (0.35 ng); ACE, acenaphthene (0.22 ng); FLU, fluorene (0.11 ng); PHE, phenanthrene (0.03 ng); ANT, anthracene (0.04 ng); FTH, fluoranthene (0.10 ng); PYR, pyrene (0.06 ng); BaA, benzo[a]anthracene (0.03 ng); CHRY, chrysene (0.02 ng); BeP, benzo[e]pyrene (0.07 ng); BbF, benzo[b]fluoranthene (0.04 ng); BkF, benzo[k]fluoranthene (0.08 ng); BaP, benzo[a]pyrene (0.05 ng); DahA, dibenzo[a,h]anthracene (0.23 ng); BghiP, benzo[g,h,i]perylene (0.15 ng); IcdP, indeno[1,2,3-c,d]pyrene (0.15 ng). nd: not detected (lower than detection limits). b Total 17 PAH associated with PM. c Genotoxic PAH (11 PAH from FTH to IcdP) associated with PM. d PAH in gaseous phase in percentage of total PAH detected on both particulate matter and in the gaseous phase.

119 816 3.7 nd nd nd nd nd nd 0.92 0.9 97.8 7.17 7.12 99.3 0.81 0.78 96.3 0.43 0.43 100 1.22 1.1 90.2 1.28 1.22 95.3 5.05 4.97 98.4 11.33 11.25 99.3 0.55 0.48 87.3 1.19 1.15 96.6 2.57 2.51 97.7 12.5 12.0 95.9 4.4 2.93 66.6 total gaseous %d kerosene (0.104-0.12 kg/h)

17.6 17.2 97.6

31.6 35.8 43 0.26 nd 0 0.19 nd 0 0.24 nd 0 0.53 0.44 82.1 1.84 1.76 95.6 0.48 0.41 85.3 1.69 1.52 89.9 1.51 1.46 96.7 1.04 1.01 97.6 5.04 4.98 98.8 9.17 9.04 98.5 3.04 3.01 99 8.55 8.52 99.7 19.47 19.46 99.9 114.9 114.8 99.9 17.4 17.4 100 total gaseous %d sawdust briquettes (0.43-0.48 kg/h)

75.2 75.2 100

17.5 18 38.8 0.14 nd 0 0.09 nd 0 0.15 nd 0 0.41 0.36 87.7 0.85 0.82 96.5 0.45 0.42 93.3 0.98 0.89 90.8 0.95 0.93 97.9 0.62 0.61 98.4 4.47 4.44 99.3 12.6 12.5 99.7 2.54 2.54 99.9 6.64 6.63 99.8 3.48 3.48 99.9 16.5 16.5 100 11.6 11.6 100 3.96 3.96 100 total gaseous %d wood (0.74-0.79 kg/h)

gen/PM (mg/g)c PAH/PM (mg/g)b PM (mg/kg) IcdP BghiP DahA B aP BkF B bF B eP CHRY B aA PYR FTH

emission factor of individual PAH, (mg/kg fuel)

ANT PHE FLU ACE ACY NAPH phase

fuel (burning rate)

linear correlation coefficient of 0.91 between EC50 of Microtox 15-min tests and LC50 (lethal dose) of rainbow trout 96-h tests and a correlation coefficient of 0.94 between EC50 of Microtox 15-min tests and the chronic toxicity threshold of the Ceriodaphnia 7-day tests for various pulp and paper mill effluents. In this study, a Microtox model 500 analyzer (Microbics Corp., Carlsbad, CA) was used to run the tests for EOM in dimethyl sulfoxide (DMSO) according to the “basic test” protocol using organic solvent sample solubilization (14). The test organism and other necessities were purchased from Microbics Corp. The tests were conducted at the laboratory of the Environmental Engineering Program, AIT. A mixture of 1% DMSO in Microtox diluent was prepared and used as the diluent for the test. An amount of 20 µL of EOM sample was added to 1980 µL of the Microtox diluent, making the initial concentration of sample of 1%. Further necessary dilution was made according to the results of preliminary tests so that the data points in the Microtox plot were located on both sides of the effective concentration, EC50, which is the concentration of tested samples causing a 50% decrease in the light output. Both 5 and 15 min of exposure time were used, and the test temperature was 15 ( 0.5 °C. The test results are also expressed in toxicity unit (TU), which is 100/EC50. The division of this TU by the sample volume (m3) at standard conditions yielded the toxicity emission/m3 of smoke (TUV). Further, the toxicity emission rate (TER), TU/h and toxicity emission factor (TEF), TU/kg of fuels, were also determined. Testing Mutagenicity. Short-term tests for mutagenicity in microbial systems, such as Ames Salmonella test, are commonly used to identify possible carcinogens (15). Two

TABLE 2. Emission Factors of Individual PAH from Domestic Combustion of Wood, Sawdust Briquettes, and Kerosenea

FIGURE 1. Sample preparation procedure.

VOL. 36, NO. 5, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

835

TABLE 3. Summary of Emission Data of Total 17 PAH and Genotoxic PAH emission factor fuel wood sawdust briquettes kerosene a

vapor/total (%) total gen 98.9 99.4 95.5

96.9 93.8 98.4

gena (mg/kg)

total (mg/MJ)

gena (mg/MJ)

66 260 67

22 22 28

0.97 6.3 1.8

0.32 0.52 0.74

total (mg/h)

gena (mg/h)

50 112 6.5

16 9.8 2.7

concentration total (µg/m3)

gena (µg/m3)

366 805 59

117 67 23

Gen: genotoxic PAH include 11 carcinogenic and cocarcinogenic PAH, from FTH to IcdP.

tester strains of Salmonella typhimurium, TA98 and TA100, with and without metabolic activation of S9, were used in this study. The S9 mix was made from the livers of Aroclor1254-pretreated rats (a supernatant of rat liver homogenate containing microsomes) that is known to be efficient for detecting different classes of carcinogens (15). Tests with the S9 mix (+S9) are conducted to detect the indirect mutagenicity, which requires metabolic activation for full expression. The strain TA98 is sensitive to different frameshift mutagens, while TA100 detects mutagens that cause base-pair substitutions, primarily at one of the G-C pairs of DNA. The tests were conducted in the laboratory of the National Cancer Institute of Thailand. Four doses of the EOM in DMSO (10, 25, 50, and 100 µL) of PM and gas-phase smoke samples were tested in duplicate. The DMSO blanks were used as negative controls for spontaneous revertants in the Ames assay. In each experiment, positive mutagenesis controls were routinely included using 2-aminofluorene (2-AF) for -S9 tests and benzo[a]pyrene (BaP) for +S9 tests. The tests were performed following the plate incorporation method (15, 16). The number of revertant colonies on the plates was counted after 48 h of incubation in the dark at 37 °C, and dose-response curves were obtained. On the basis of the curves, the presence/absence of mutagenic activity was determined. The confirmed mutagenicity was expressed as the specific mutagenic activity (net revertant colonies/µL of the sample) and the mutagenic index (MI; net revertant colonies/m3 of smoke), as well as the mutagenicity emission factor (MEF; revertants/kg of fuel (rev/kg)).

Results and Discussion PAH Emission. All of the 17 PAH were detected in the wood fuel smoke and sawdust smoke, including 11 genotoxic compounds: carcinogens (BaA, chrysene, BbF, BkF, BaP, DahA, and IcdP) and cocarcinogens (fluoranthene, pyrene, BeP, and BghiP). In kerosene smoke, only 14 compounds were detected (Table 2). The emission factors of the total 17 PAH on both fuel weight basis (mg/kg) and energy basis (mg/MJ), presented in Table 3, are highest for the sawdust briquettes (260 mg/kg and 6.3 mg/MJ), the second highest for kerosene (67 mg/kg and 1.8 mg/MJ), and the lowest for the wood fuel (66 mg/kg and 0.97 mg/MJ). However, most of the contribution was from the first low molecular weight compounds, especially for sawdust briquettes. If only the 11 genotoxic PAH are taken into account, the emission factor ranking order was from the kerosene (0.74 mg/MJ) to sawdust briquettes (0.52 mg/MJ) and lowest for the wood fuel (0.32 mg/MJ). The emission factor for BaP, for example, found in this study for wood fuel (0.41 mg/kg) and for sawdust briquettes (0.53 mg/kg), is comparable to the emission factor found by the previous study (8) for eucalyptus wood (with a higher burning rate) of 0.7 mg/kg. Most of the PAH were found in the gas phase, especially the lower molecular weight compounds (Table 2). Relatively high levels of the lower molecular weight PAH were found on PM in kerosene smoke (Tables 2 and 3). The 11 genotoxic PAH associated with PM contributed the highest percentage 836

emission rate

total (mg/kg)

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 5, 2002

in sawdust smoke (6.2%), the second highest for wood (3.1%), and the lowest for kerosene (1.6%). The lower fraction of the total 17 PAH was associated with PM, and the rank is the opposite (i.e., highest for kerosene smoke (4.5%), followed by wood (1.1%), and lowest for sawdust (0.6%)). The apportioning is in the same range as that found by other studies. The previous study (8) found 5% of the total 18 PAH (plus coronene) associated with PM in the eucalyptus wood smoke. Another study (17) on 19 PAH (the 17 PAH plus methylnaphthalene and perylene) emitted from the open burning of forest wood in a wind tunnel found 7.6% out of 25.8 mg/kg of fir slash and 3.3% out of 31.7 mg/kg of pine slash associated with PM. The relatively high temperature of the stack gas (60-70 °C; Table 1) and the heating temperature of the sampling probe and filter box (125 °C) in this study partly explain the low fraction of PAH on PM. Also, the short residence time of the flue gas in the hood and along the sampling line may not be sufficient for the phase equilibrium to be established. In addition, the amount of PM emitted as well as the particle sorption characteristics may also result in the differences in the PAH fraction associated with PM for the three fuels. The PM emission factor was highest for sawdust and lowest for kerosene (Table 2). The 17 PAH and genotoxic PAH associated with a unit mass of PM was, however, found to be the highest for kerosene, the second highest for sawdust, and the lowest for wood. To assess exposure levels, the emission rate (mg/h) and concentration of pollutants in smoke (µg/m3) are also presented in Table 3. The lowest rate of burning, which was for kerosene (0.1-0.12 kg/h), resulted in the lowest emission rate of 17 PAH, 6.5 mg/h, and concentration in the smoke, 59 µg/m3, which are much lower than wood burning, 50 mg/h and 366 µg/m3, respectively, though the emission factors on a weight basis of the two fuels were almost the same. The highest emission rate and concentration of 17 PAH was from sawdust briquettes. However, the highest emission rate and concentration of genotoxic PAH in smoke was from wood burning. Toxicity. The Microtox test results for the EOM of PM and gas-phase samples for the three tested fuel-cookstove systems are presented in Table 4. The particulate phase of smoke from the kerosene cookstove from all three of the burning batches was nontoxic to Microtox bacteria. Both gas and PM smoke samples from sawdust briquettes and wood fuel burning were found toxic. The predominant toxicity was associated with the gas phase. For sawdust smoke, the gas-phase toxicity accounted for 98% of total toxicity in 5 min exposure tests and for 97% in 15 min test. The corresponding values for wood fuel are 96% and 98%, respectively. The gas-phase toxicity of sawdust briquettes and kerosene smoke decreased when the exposure time increased from 5 to 15 min, but for wood fuel samples, some increase was observed. The reverse trend was noted for the PM samples (i.e., increased with exposure time for sawdust briquettes but reduced for wood fuel). The reduction in toxicity with time indicated that the effects are short term for the Microtox

TABLE 4. Toxicity Emission from Selected Fuel-Cookstove Systemsa 5 min exposure test fuel

phase

keroseneb

PM gas

sawdust briquettes

PM

TU

gas total gas/total (%) PM

wood fuel

gas total gas/total (%)

0 6550 6450-6640 320 310-332 15 940 13 380-18 200 16 260 98 670 615-742 17 550 16 250-18 570 18 220 96

15 min exposure test

TUV (TU/m3)

TER (TU/h)

TEF (TU/kg)

NT 540 484-595 48 40-58 2330 2300-2360 2380

NT 62 800

NT 587 700

6100

13 500

299 700

673 100

305 800

686 600

11 400

15 000

298 800

392 600

310 200

407 600

86 82-94 2250 2170-2340 2340

TUV (TU/m3)

TER (TU/h)

TEF (TU/kg)

NT 428 361-494 56 47-68 1640 1620-1650 1690

NT 49 700

NT 462 500

7200

16 100

210 400

472 400

217 600

488 500

7200

9500

331 000

434 800

338 300

444 300

TU 0 5170 4820-5520 380 371-384 11 190 9 380-12 820 11 570 97 426 359-503 19 460 17 840-21 160 19 880 98

55 48-64 2490 2380-2670 2550

a For kerosene smoke, the PM phase is nontoxic (NT); hence, the gas-phase toxicity is the total toxicity. (TU ) 100/EC50, toxicity unit; TUV, toxicity unit in 1 m3 of smoke; TER, toxicity emission rate; TEF, toxicity emission factor.)

TABLE 5. Mutagenicity of Smoke from Saw Dust Briquettes and Wood Cookstovesa TA98

TA100

-S9 sample PM gas total gas/total, % PM

MI

(rev/m3)

2500 1910-3390 7100 6830-7290 9600 74

8000 7330-8850 gas 12 100 11 490-12 750 total 20 100 gas/total (%) 60 a

+S9 MEF (rev/kg) 716 000 2 052 000 2 768 000

1 389 100 2 106 000 3 495 100

MI

(rev/m3)

3600 3330-3810 2700 2540-2860 6200

-S9

MEF (rev/kg)

MI

(rev/m3)

Sawdust Briquettes 1 017 500 2100 1990-2280 775 300 7900 7640-8140 1 792 800 10 000 79

9200 9070-9340 6800 6 470-7 290 16 000

+S9 MEF (rev/kg)

MI

(rev/m3)

606 300 1300 1160-1390 2 272 200 3200 3030-3390 2 878 500 4500

Wood Fuel 1 607 300 13 000 2 265 200 12 730-13 280 1 188 800 56 000 9 778 600 55 350-57 340 2 796 100 69 000 12 043 900 81

6700 6610-6820 10 800 10 270-11 420 17 500

MEF (rev/kg) 367 700 915 700 1 283 300

1 168 300 1 873 900 3 042 200

MI, mutagenicity index (rev/m3); MEF, mutagenicity emission factor (rev/kg of fuel); -S9, test without S9 mix; +S9, test with S9 mix.

bacteria and that they restored the light emission ability after a prolonged exposure. The total toxicity emission factor (both PM and gas phase) was the highest for sawdust briquettes, the second highest for kerosene, and the lowest for wood fuel burning. The slow burning rate of kerosene stove produced a much lower toxicity emission rate and toxicity content in one cubic meter (m3) of smoke than the other two fuels. Wood fuel, in fact, produced the highest toxicity emission rate (Table 4), which would lead to a higher exposure level. Mutagenicity. Results of mutagenicity tests indicated that both PM and gas phase of kerosene cookstove smoke were not mutagenic. For the other two fuels, there are positive mutagenic activities on both tester strains without the S9 mix (Table 5), indicating the presence of direct-acting mutagens which caused frameshift mutations (in strain TA98) and base pair substitutions (in strain TA100). More than 60% of the direct-acting mutagenicity was associated with the gas phase of the smoke. Thus, the predominant contribution to mutagenicity also comes from the gas phase, though it is lower than the contribution to the toxicity. With S9 added, the tester strain TA98 (TA98+S9) showed increased mutagenicity for PM phase samples of both sawdust briquettes and wood smokes. This indicates the presence of indirect mutagens in PM phase of smoke, causing frameshift mutations.

The TA98+S9 for the gas phase and TA100+S9 for both gas and PM phase samples of both fuels showed the reduction in mutagenicity as compared to the corresponding tests without the S9 mix. The microsomal enzymes thus reduced the mutagenic activity (due to adding the S9 mix), which indicates that virtually all the mutagenic activities of these samples were direct. Another study (3) reported results of TA98+S9 mutagenicity of PM phase samples from woodstoves of 15.8 × 106 rev/kg for pine and 2.9 × 106 rev/kg for oak. The values found in this study for the PM phase are 1.6 × 106 rev/kg for the wood fuel and around 106 rev/kg for sawdust briquettes (Table 5). For oil fired residential furnaces, the mutagenicity emission varies with the oil types. Heavier oil seems to have higher mutagenicity. The same study also cited a value of 0.04 × 106 rev/kg for oil no. 1 (0.84 kg/L) and higher value (0.36 × 106 rev/kg) for oil no. 2 (0.87 kg/L). The kerosene used in this study has a density of 0.78 kg/L (i.e., lighter than that of the cited oils), and the smoke from this kerosene stove showed no mutagenic activity. Toxicity, Mutagenicity, and PAH Content of Collected Samples. The PAH content of the samples (i.e., the actual amount of PAH in the samples in µg) is presented in Table 6 together with the toxicity and mutagenicity of the samples. Kerosene gas-phase samples contained the least PAH. The VOL. 36, NO. 5, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

837

TABLE 6. PAH Content and Toxicity, Mutagenicity of Collected Smoke Samples vapor phase

PM phase

smoke sample

17 PAH (µg)

13 PAH (µg)

gen (µg)

5 min (TU)

TA98a

wood sawdust kerosene

2800 6160 680

1280 760 310

900 490 295

17 550 15 940 6 550

9.42 4.87 NR

a

TA100a

17 PAH (µg)

13 PAH (µg)

gen (µg)

5 min (TU)

TA98a

TA100a

43.69 5.38 NRb

28.7 34.5 38.3

28.5 31.7 7

27.9 30.4 5.52

670 320 NTb

6.23 1.69 NR

10.12 1.43 NR

Reported values are specific direct mutagenic activity of samples, net rev/µL of sample extract (-S9 tests).

b

NR, no response; NT, nontoxic.

TABLE 7. Toxicity and Mutagenicity Emission from Daily Cooking for a Household of Four Persons mass of fuel required per day (kg/day)a Microtox, 5 min exposure test Microtox, 15 min exposure test TA98-S9 TA100-S9 17 PAH (mg) genotoxic PAH (mg)

kerosene

sawdust briquette

0.78

3.0

Toxicity Emission per Daily Cooking (TU/day) 473 000 2 060 000 369 000 1 466 000 Mutagenicity Emission per Daily Cooking (rev/day) NRb 8 304 000 NR 8 636 000 PAH Emission 53 22

780 66

wood 3.4 1 386 000 1 511 000 11 883 000 40 950 000 224 75

a Fuel amounts were calculated using a net useful heat requirement per person of 4 MJ/day and the efficiency of kerosene cookstove of 50% and the other two of 25% (see Table 1). b NR, no response.

PM phase of all of the samples contained low and approximately the same amount of 17 PAH. Most PAH are classified as toxic, but toxicity is different for different compounds. The relative toxicity of individual PAH generally varied with the molecular weight. A study (18) on PAH toxicity to marine and estuarine amphipods, using a model and experimental data from 33 field-sediment samples, found that among 13 studied PAH, the four lower molecular weight PAH (naphthalene, acenaphthene, acenaphthylene, and fluorene) each contributed less than 2% of total toxicity. The highest contribution was from BkF (17.4%), whereas BaP contributed 10.8%. In the present study, the first four compounds actually are predominant (more than 80%) in the PM phase of the kerosene smoke samples, which may be a reason for the low toxicity of the kerosene PM sample though the total 17 PAH are almost the same for 3 types of PM samples. The sum of the last 13 PAH (i.e., excluding the first four PAH) is also presented in Table 6. The most toxic vapor sample is wood smoke, followed by sawdust, and the least toxic is kerosene. The specific direct mutagenicity detected by both bacteria strains is also in the same ranking order. This corresponds with the ranking order of the last 13 PAH and genotoxic PAH (last 11 PAH) content of the samples but not with the content of all 17 PAH. There is not a clear relationship between PAH content and the toxicity/mutagenicity of PM samples. These PM samples contain low levels of PAH, but in general, individually, these higher molecular weight PAH should be more mutagenic and toxic. Wood smoke PM was most toxic and mutagenic, though PAH content of the samples was lower than that of the sawdust samples. Kerosene PM phase, which contained a low amount of the 13 PAH and genotoxic PAH, was not found to be toxic or mutagenic. The low burning rate of kerosene must result in a low emission rate and concentration of all other chemical compounds besides PAH. Emissions from combustion may contain hundreds of toxic and genotoxic compounds that are both aliphatic and POM (3, 19, 20). Other compounds present in smoke, such as formaldehyde, phenol, and so forth, could also contribute to the Microtox toxicity and genotoxicity (4). A group of authors (2) studied mutagenicity of the extract of PM collected 838

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 5, 2002

during cooking period (using smoky coal under unventilated conditions) in a Xuan Wei home inhabited by persons with lung cancer and found that the fraction containing PAH and alkylated PAH contributing up to 43% of mass and 61% of mutagenicity of the organic fraction of PM samples. The authors also suggested that the presence of three-to-fourring alkylated PAH is a significant factor that may link to the high incidence of lung cancer in the village. High mutagenic activities for a number of unsubstituted and substituted PAH were found in another study (21). The unsubstituted PAH and methylated PAH were found to be indirect mutagens, while the nitrated PAH were found to be direct-acting genotoxicants in the Escherichia coli PQ37 tests. Moreover, the genotoxicity of the nitrated PAH was increased with an increase in the extent of nitrification. The nitrated PAH, however, are the reaction products of PAH and reactive species present in polluted atmosphere (5), and they are not found to be significant constituents of major sources of primary POM, such as diesel soot in Los Angeles (22). A complete chemical characterization of the smoke is practically impossible. In addition, there are also possible synergistic and antagonistic effects of the chemical compounds. In this study, only 17 unsubstituted PAH were analyzed (i.e., just a small fraction of possible mutagenic and toxic compounds present in the smoke). It is therefore not expected to reveal the true relationship between the chemical composition of the smoke and its toxicity and mutagenicity. Nevertheless, the toxicity and mutagenicity detected by the quick bioassays were found following the same ranking order as the sum of 13 and 11 PAH of the gas-phase samples, which contained a high amount of PAH. Though there is no clear relationship between PAH content and the toxicity and mutagenicity of PM samples, it is shown that these samples contained a lower amount of PAH and a lower toxicity and mutagenicity than the gas-phase samples. These quick bioassays, hence, are useful screening tools and can provide indicators of a potential health risk from cooking activities. However, it worth noting that the tests results obtained for bacteria may not be directly extrapolated to human health effects.

To assess the risks associated with domestic cooking, the emissions of PAH, toxicity, and mutagenicity from daily cooking was estimated and presented in Table 7. For the estimation, it was assumed that, on average, 4 MJ net (useful) heat content is required for daily cooking per person (3). The emission of 11 genotoxic PAH is highest from the wood fuel and lowest from kerosene. Among the three fuel-cookstove systems, the highest risk of exposure to mutagenicity and toxicity is associated with wood fuel, followed by the sawdust briquettes, and the lowest risk is from kerosene. The kerosene cookstove thus can provide safer household cooking. This is especially recommended for urban dwellings in developing countries, where home and kitchen are not physically separated and poorly ventilated, if safer fuels such as electricity or LPG are not yet affordable.

Acknowledgments This study was partly funded by the Swedish Agency for Research Cooperation with Developing Countries (SAREC) of the Swedish International Development Cooperation Agency (Sida) through the Asian Regional Research Program in Energy, Environment and Climate at the Asian Institute of Technology. The authors thank the staff from the National Institute for Cancer Research of Thailand, the Biochemistry and Chemical Carcinogenesis Section, especially Dr. Wannee R. Kusamran, Dr. Piengchai Kupradiun, and khun Anong Tepsuwan for their kind help and support and also for allowing the use of the laboratory facility for mutagenicity testing.

Literature Cited (1) Teschke, K.; Hertzman, C.; Netten, C. V. Environ. Res. 1989, 50, 296-308. (2) Chang, J. C.; Wise, S. A.; Cao, S. R.; Mumford, J. L. Environ. Sci. Technol. 1992, 26, 999-1004. (3) Smith, K. R. Biofuels, Air Pollution, and HealthsA Global Review; Plenum Press: New York, 1987. (4) Godish, T. Indoor Air Pollution Control, TD883.1.G63.P.697; Lewis Publishers: Pearl River, NY, 1989. (5) Nikolaou, K.; Masclet, P.; Mouvier, G. Sci. Total Environ. 1984, 32, 103-132. (6) McCrillis, R. C.; Watts, R. R.; Warren, S. H. J. Air Waste Manage. Assoc. 1992, 42, 691-694. (7) Urban air pollution, UNEP/GEMS Environment Library No. 4; United Nations Environmental Programme: 1991.

(8) Kim Oanh, N. T.; Reutergårdh, L.-B.; Dung, N. T. Environ. Sci. Technol. 1999, 33, 2703-2709. (9) U.S. EPA Modified Method 5 (Method 0010) Sampling Train, Revision 0, September 1986; U.S. EPA: Washington DC, 1986. (10) Graseby-Andersen, Auto5 Sampler Operator’s Manual, Revision July 31, 1995; Graseby-Andersen: Smyrna, GA, 1995. (11) Compedium of Methods for Determination of Toxic Organic Compounds in Ambient Air, June 1988; U.S. EPA/600/4-89/017; U.S. EPA: Washington, DC, 1990. (12) Biological-Chemical Characterization of Industrial Wastewater; INFORMS, Swedish Environmental Protection Agency: Stockholm, Sweden, 1990. (13) Firth, B. K.; Backman C. J. A comparison of Microtox testing with rainbow trout acute and Ceriodaphnia chronic bioassays using pulp and paper mill wastewaters. Proceedings of the Environmental Conference, Book 2, Atlanta, GA, April 1990; TAPPI Press: Atlanta, GA, 1990; pp 621-626. (14) Microtox Manual, A Toxicity Testing Handbook; Microbics Coporation: Carlsbad, CA, 1990; Vols. 1-5. (15) Ames, B. N.; Mccann, J.; Yamasaki, E. Mutat. Res. 1975, 31, 347364. (16) Maron, D. M.; Ames, B. N. Mutat. Res. 1983, 113, 173-215. (17) Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. Atmos. Environ. 1996, 30, 3825-3835. (18) Swartz, R. C.; Schults, D. W.; Ozretich, R. J.; Lamberson, J. O.; Cole, F. A.; DeWitt, T. H.; Redmond, M. S.; Ferraro, S. Environ. Toxicol. Chem. 1995, 14, 1977-1987. (19) Indoor Air Quality: Organic pollutants; EURO Reports and Studies 111, Regional Office for Europe, World Health Organization: 1989. (20) Zhuang, Y.-H. Emissions from Biomass Combustion in Asia. Environmental systems reviews, No. 40; ENSIC, Asian Institute of Technology: Pathumthani, Thailand, 1996. (21) Mersch-Sundermann, V.; Mochayedi, S.; Kevekordes, S.; Kern, S.; Wintermann, F. Anticancer Res. 1993, 13, 2037-2044. (22) Pitts, J. N., Jr. Atmos. Environ. 1987, 21, 2531-2547. (23) Proceedings of Asean Conference on Energy Conservation, Chiang Mai, Thailand, 1985. (24) Kandpal, J. B.; Maheshwari, R. C.; Kandpal, T. C. Energy Convers. Manage. 1995, 36, 1067-1072. (25) Foley, G.; Moss, P. Improved Cooking Stoves in Developing Countries, Technical Report No. 2, Energy Information Programme, Earthscan: London, U.K., 1983.

Received for review June 13, 2001. Revised manuscript received November 8, 2001. Accepted November 20, 2001. ES011060N

VOL. 36, NO. 5, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

839