Emission of polycyclic aromatic hydrocarbons and their carcinogenic potencies from cooking sources to the urban atmosphere.

Traffic has long been recognized as the major contributor to polycyclic aromatic hydrocarbon (PAH) concentrations. However, this does not consider the contribution of cooking sources of PAHs. This study set out, first, to assess the characteristics of PAHs and their corresponding benzo[a]pyrene equivalent (B[a]Peq) emissions from cooking sources to the urban atmosphere. To illustrate the importance of cooking sources, PAH emissions from traffic sources were then calculated and compared. The entire study was conducted on a city located in southern Taiwan. PAH samples were collected from the exhaust stacks of four types of restaurant: Chinese, Western, fast food, and Japanese. For total PAHs, results show that the fractions of gaseous PAHs (range, 75.9-89.9%) were consistently higher than the fractions of particulate PAHs (range, 10.1-24.1%) in emissions from the four types of restaurant. But for total B[a]Peq, we found that the contributions of gaseous PAHs (range, 15.7-21.9%) were consistently lower than the contributions of particulate PAHs (range, 78.1-84.3%). For emission rates of both total PAHs and total B[a]Peq, a consistent trend was found for the four types of restaurant: Chinese (2,038 and 154 kg/year, respectively) > Western (258 and 20.4 kg/year, respectively) > fast food (31.4 and 0.104 kg/year, respectively) > Japanese (5.11 and 0.014 kg/year, respectively). By directly adapting the emission data obtained from Chinese restaurants, we found that emission rates on total PAHs and total B[a]Peq for home kitchen sources were 6,639 and 501 kg/year, respectively. By combining both restaurant sources and home kitchen sources, this study yielded emission rates of total PAHs and total B[a]Peq from cooking sources of the studied city of 8,973 and 675 kg/year, respectively. Compared with PAH emissions from traffic sources in the same city, we found that although the emission rates of total PAHs for cooking sources were significantly less than those for traffic sources (13,500 kg/year), the emission rates of total B[a]Peq for cooking sources were much higher than those for traffic sources (61.4 kg/year). The above results clearly indicate that although cooking sources are less important than traffic sources in contributing to total PAH emissions, PAH emissions from cooking sources might cause much more serious problems than traffic sources, from the perspective of carcinogenic potency.

Polycyclic aromatic hydrocarbons (PAHs) are one of the first identified airborne carcinogenic pollutants containing two or more aromatic rings that are fused together in different arrangements (1). PAHs and derivatives are associated with the incomplete combustion of organic material arising partly from natural combustion such as volcano eruptions or forest fires, but most emissions arise from anthropogenic activities such as the burning of gasoline in motor vehicles, residential heating, home cooking, and industrial production activities (1). In the past 30 years, many studies have suggested increased risk for certain cancers in cooks and other food-service workers (2)(3)(4)(5)(6)(7). Because of this, many researchers have emphasized investigating PAH compositions in indoor air resulting from cooking processes. For example, Rogge et al. (8) found that the use of natural gas for cooking would increase the PAH concentrations in indoor air. Siegmann and Sattler (9) found that PAH concentrations contained in hot cooking oil fumes (range, 1.08-22.8 µg/m 3 ) were higher than those in an office room where 96 cigarettes were consumed within 6 hr (1.2 µg/m 3 ). In particular, van Houdt et al. (10) suggested that the cooking process was the most important contributor to the total mutagenic activity of indoor air. However, canopy hood ventilation has been widely used for cooking sources in many urban areas. Therefore, it can be expected that most PAHs emitted from cooking sources could be exhausted to the urban atmosphere. Many researchers have suggested that traffic is the major contributor to PAH concentrations in the atmosphere of urban and suburban areas (1). In one study, Harrison et al. (11) indicated that road traffic accounted for 88% of ambient benzo[a]pyrene at an urban location in Birmingham, United Kingdom. But to our knowledge, this estimate does not consider the contribution of cooking sources and hence warrants further investigation.
In this study we first focused on investigating the contents of PAHs that were emitted from stacks of four types of restaurants: Chinese, Western, fast food, and Japanese. Then, PAH emissions from home kitchen sources were estimated according to emission data obtained from Chinese restaurants. In addition, several PAH compounds have been classified by the International Agency for Research on Cancer (12) as "probable" human carcinogens (2A) or "possible" human carcinogens (2B). Therefore, the carcinogenic potency associated with PAH emissions from various cooking sources were also estimated. In this study, we assumed that PAH emissions from both restaurants and home kitchens represented those emitted from all cooking sources. To assess the effect of cooking sources on PAHs emitted into the urban atmosphere, we compared the above-estimated PAH emissions with those emitted from traffic sources in the same city by directly using the emission data presented in our previous studies (13)(14)(15).
sampling flow rate of approximately 10 L/min for 45 min per sample. Using a critical orifice flow calibrator (model GMW-25; General Metal Work, Taichung, Taiwan), we determined the accurate sampling flow rate by averaging the flow rates measured at the beginning and at the end of the sampling period. PAHs collected by a tube-type glass fiber filter (25 × 90 mm, Whatman glass fiber thimble) in the sampling probe (i.e., particulate PAHs) were stored in a prebaked glass bottle (wrapped with aluminum foil) for shipment before the chemical analysis. Gaseous PAHs collected by the glass cartridge, packed with a 5-cm polyurethane foam (PUF) plug, followed by a 2.5-cm XAD-16 resin supported by a 2.5-cm PUF plug, were stored in a clean screw-capped jar (with a Teflon cap liner) for transportation. Three breakthrough tests were investigated by using a two-layer XAD-16 cartridge with the sequence in the cartridge as PUF-1, XAD-16-1, PUF-2, XAD-16-2, and PUF-3 (Li-Tex Co., Kaoushing, Taiwan). No significant amounts of PAHs were found in the sections of PUF-2, XAD-16-2, or PUF-3.
PAH analysis. For PAH analysis, each collected sample (including particulate and gaseous PAH samples) was extracted in a Soxhlet extractor with a mixed solvent (n-hexane and dichloromethane; vol/vol, 1:1; 500 mL each) for 24 hr. The extract was then concentrated, cleaned up, and reconcentrated to exactly 1.0 or 0.5 mL. PAH contents were determined with a Hewlett-Packard (HP) gas chromatograph (GC) (HP 5890A; Hewlett-Packard, Wilmington, DE, USA) with a mass selective detector (MSD) (HP 59H72) and a computer workstation (Aspire C500; Acer, Taipei, Taiwan). This GC/MSD was equipped with a capillary column (HP Ultra 2, 50 m × 0.32 mm × 0.17 µm) and an automatic sampler (HP-7673A) and operated under the following conditions: injection volume of 1 µL, splitless injection at 310°C, ion source temperature at 310°C, oven from 50 to 100°C at 20°C/min; 100 to 290°C at 3°C/min; hold at 290°C for 40 min. The masses of primary and secondary ions of PAHs were determined using the scan mode for pure PAH standards. PAHs were qualified using the selected ion monitoring (SIM) mode (21).
The  Germany). Analysis of serial dilutions of PAH standards showed the limit of detection of GC/MSD to be between 0.021 and 0.384 ng for the 21 PAH compounds. The limit of quantification (LOQ) was defined as the limit of detection divided by the sampling volume. The LOQs of the 21 PAH compounds for PSS samples were between 0.047 and 0.853 ng/m 3 . Ten consecutive injections of a PAH 610-M standard yielded an average relative standard deviation of GC/MSD integration area of 3.0%, with a range of 0.8-5.1%. Following the same experimental procedures used for the treatment of samples, we determined recovery efficiencies by processing a solution containing known PAH concentrations. This study showed the recovery efficiencies for the 21 PAH compounds to range from 0.765 to 1.060, with an average value of 0.863. Analyses of field blanks, including the aluminum foil, polyethylene (PE) bottle, glass fiber filter, and PUF/XAD-16 cartridge, revealed no significant contamination (GC/MSD integrated area < detection limit).
Data analysis. In this study, the total PAH concentration was regarded as the sum of the concentrations of 21 PAH compounds for each collected sample. To assess PAH homolog distribution for each collected sample, we further classified total PAHs into three categories: low molecular weights (LM-PAHs, containing two-to three-ringed PAHs), middle molecular weights (MM-PAHs, containing four-ringed PAHs), and high molecular weights (HM-PAHs, containing five-to seven-ringed PAHs). Moreover, considering that several PAH compounds are known human carcinogens, the carcinogenic potencies associated with PAH emissions from each emission source were also determined. In principle, the carcinogenic potency of a given PAH compound can be assessed on the basis of its benzo[a]pyrene equivalent concentration (B[a]P eq ). Calculation of the B[a]P eq concentration for a given PAH compound requires the use of its toxic equivalent factor (TEF), which represents the relative carcinogenic potency of the given PAH compound, using benzo[a]pyrene as a reference compound to adjust its original concentration. Only a few proposals for TEFs are available (22)(23)(24)(25)

Results and Discussion
Characteristics of PAHs emitted from restaurant sources.  types of restaurant (ranges, 93.9-97.5% and 72.3-87.7%, respectively) were higher than the fractions of particulate PAHs (range 2.46-6.11%). But the fractions of gaseous PAHs in HM-PAHs (range 13.7-21.0%) were much lower than those of particulate PAHs (range 79.0-86.3%). Considering that total PAHs contained in the stack flue gas of the four types of restaurant were composed mainly of LM-PAHs (Table 3), it is not so surprising that gaseous PAHs accounted for higher fractions in total PAHs. Conversely, it is known that PAHs with higher molecular weights are associated with the higher TEFs ( However, it should be noted that the mean EF totPAH for Chinese restaurants was 1.80-and 7.43-fold higher than those for fast-food and Japanese restaurants, respectively. However, the mean EF totB[a]Peq for Chinese restaurants was 40.9-and 200-fold in magnitude higher than those for fast-food restaurants and Japanese restaurants, respectively. These results suggest that PAH emissions from both fast-food restaurants and Japanese restaurants not only contained lower total PAH contents but also had much lower carcinogenic potencies. At this stage, it is known that both EF totPAH and EF totB[a]Peq can be affected strongly by the cooking method and the type of food oil. Because the mechanisms associated with the formation of PAHs for various cooking sources were not known, this area warrants further investigation. PAH emission rates for restaurant and home kitchen sources. We assume that both the mean emission factors and food oil consumption rates obtained from this study are representative of the four types of restaurant. In addition, we assumed that all restaurants ran for 365 days per year and served only lunch and dinner each day. Based on these assumptions, the total PAH and total B[a]P eq emission rates (denoted ER totPAH and ER totB[a]Peq , respectively, in kilogram per year) for a given type of restaurant could be determined, respectively, according to the following two equations: where n was the total number of the given types of restaurant in Tainan, Taiwan (743,  88, 20, and 11 for Chinese, Western, fast food, and Japanese, respectively). Results show ER totPAH and ER totB[a]Peq for the four types of restaurant were Chinese (2,038 and 154 kg/year, respectively) > Western (258 and 20.4 kg/year, respectively) > fast food (31.4 and 0.104 kg/year, respectively) > Japanese (5.11 and 0.014 kg/year, respectively) ( Table 5).
In addition to restaurants, it is believed that home kitchens might also play an important role in PAH emissions into the urban atmosphere. According to the internal statistics data provided by the Taiwan Food Oil Producer Association, the personal consumption rate of food oil (PCR food-oil ) in the Taiwan area was approximately 58. 7 (Table 5). Traffic sources have long been recognized as the major contributor of PAHs in urban areas (1,11). Therefore, to assess the importance of cooking sources, PAH emissions from traffic were also estimated. According to the statistics data provided by the Transportation Bureau in Tainan (Table 5). These results suggest that ER totPAH from cooking sources (8,973 kg/year) was approximately 0.66-fold that from traffic sources, but ER totB[a]Peq from cooking sources (675 kg/year) was approximately 11.0-fold that from traffic sources. These results clearly suggest that in Tainan PAH emissions from cooking sources are much less important than those from traffic sources. However, the carcinogenic potency of PAH emissions from cooking sources was much greater than that from traffic sources.

Conclusions
In all restaurant sources studied, the emissions of gaseous PAHs were greater than those of particulate PAHs. However, the carcinogenic potency of gaseous PAH emissions was less than that of particulate PAH emissions. PAH emission intensities for the four types of restaurants for both EF totPAH and EF totPAH shared the same trend: Chinese > Western > fast food > Japanese. For cooking sources, we found that both ER totPAH and ER totB[a]Peq from home kitchen sources were consistently higher than those for restaurant sources. Nevertheless, these results suggest both home kitchens and restaurants should be included for estimating PAH emissions from cooking. To determine the significance of cooking sources, PAH emissions from traffic sources were estimated. We found that although ER totPAH from cooking sources was approximately 0.66-fold that from traffic sources, ER totB[a]Peq from cooking sources was approximately 11.0-fold that from traffic sources. These results clearly suggest that, in addition to PAHs emitted from traffic, cooking sources make an important contribution to PAH emissions into the ambient environment of an urban area. However, it should be noted that other city areas might have intrinsic differences in cooking methods, food oil consumption rates, restaurant compositions, and traffic conditions compared with those found in this study. Therefore, the importance of cooking sources as contributors to PAH emissions in other city areas could be somewhat different than that found in this study. Therefore, further investigation is needed.