Characterization of cooking effluent from seven commercial kitchen appliances and representative food products.
Characterization of emissions from cooking processes continues to be of interest, with an emphasis on grease accumulation on surfaces in kitchens and within exhaust systems and its impact on cleaning and fire potential. Indoor airborne emissions may create potential health hazards to cook staff and others within the building; widespread dissemination to outdoors will impact atmospheric chemistry and air quality.
Several regulations on emissions from commercial kitchens have been adopted in the U.S. The South Coast Air Quality Management District (SCAQMD 1997) implemented Rule 1138 that requires emission controls on certain chain-driven charbroilers within the district. More recently, the Bay Area Air Quality Management District (BAAQMD 2007) adopted Rule 2 of Regulation 6 that also requires emission controls on chain-driven charbroilers beginning January 1, 2009. In addition, control of underfired charbroiler emissions is scheduled to be phased in between January 1, 2010, and January 1, 2013.
An annotated bibliography was written (Gerstler et al. 1996) that includes 74 references on measured cooking effluent and technologies related to its sampling and analysis. Since the review's publication, several additional studies were reported that add to the archival literature.
Residential applications have been the focus of some research. A series of papers document a study conducted by the National Exposure Research Laboratory in Reston, Virginia. The first paper (Wallace 2000) documented several sources of particles and evaluated the levels of polycyclic aromatic hydrocarbons (PAH) and carbon monoxide in a townhouse with no external ventilation from the kitchen. Cooking was found to be the major source of particles in the building. The levels of PAH were deemed insignificant, so they were not monitored in future work at the site. A subsequent paper (Wallace et al. 2004) focused on particles from 10 nm to 2.5 [micro]m in size. Forty-four cooking episodes were selected for analysis, as it was determined that cooking (mostly frying in a skillet) was capable of producing more than ten times the ultrafine particle number emissions than other activities. Typical cooking episodes produced particles at a rate of about [10.sup.14] particles/h during a 5-15 min cooking cycle. Ultrafine (<100 nm) and accumulation mode (0.1-1 [micro]m) particle sources were analyzed for various activities. The highest mean number concentrations were produced by what was termed "complex cooking" on a gas stove that generated between 35,000 and 50,000 [cm.sup.-3] compared to less than 1% of this when no indoor sources were observed. A related study was performed using personal exposure monitors on 37 participants in their own homes. The highest mean source strengths were identified from burned food (470 mg/min), grilling (173 mg/min), and using a fry pan (60 mg/min). A group at the University of Aberdeen Medical School in Aberdeen, Scotland (Dennekamp et al. 2001), performed a study in which emissions from both electric and gas-operated cooking appliances were conducted in a small (70 [m.sup.3]) unventilated chamber that simulated a small residence. Particles from 10 to 500 nm NO and [NO.sub.2] were analyzed. Conclusions drawn from this study were that potentially toxic concentrations of small particles and very high concentrations of oxides of nitrogen can be generated without adequate exhaust ventilation.
Effluents from commercial cooking processes were also investigated. Particulate (Kleeman et al. 1999) and organic compounds (Schauer et al. 1999) emitted from meat charbroiling were measured at a commercial kitchen field site. Samples were obtained in the exhaust duct downstream of the grease filters installed in the exhaust hood. Frozen and thawed hamburger patties (20% fat) were cooked on a natural gas-fired charbroiler. Total mass emission rates were determined to be approximately 30 g/kg meat cooked. Peak particle mass concentration was determined using a micro-orifice uniform deposit impactor (MOUDI) and was found to be near 0.2 [micro]m. Although inorganic composition was investigated, the vast majority of the effluent consisted of organic compounds. Several nonmethane volatile organic compounds and semi-volatile organic compounds were identified including n-alkanoic acids, n-alkenoic acids, and carbonyls.
Personal exposure of the cook staff in 19 commercial kitchens was made in Norway (Svendsen et al. 2002) using glass fiber filter cassettes and a sampling device for aldehydes. Results showed that in some types of kitchens the concentration of fat aerosols can reach 6.6 mg/[m.sup.3], and the sum of aldehydes can attain 185 [micro]g/[m.sup.3]. Exposures were found to be greater in small kitchens than in chain restaurants or hotel kitchens.
A more recent series of measurements conducted at the University of California Riverside CE-CERT facility was reported (McDonald et al. 2003). Emissions from several appliances were measured, all in the exhaust duct, after which they were sent through a dilution tunnel and then a residence chamber to approximate the conditions expected after release to the ambient. Particulate mass emission factors were considerably less than measured in earlier studies and ranged from 4.5 g/kg for 21% fat hamburger cooked on a chain-driven char-broiler to 15 g/kg for 25% fat hamburger cooked on an under-fired charbroiler. The majority of the mass was determined to be organic carbon with significant amounts of PAHs (primarily naphthalene), lactones, and cholesterol.
Cooking exhaust from 15 commercial kitchens was sampled and analyzed for 13 carbonyl compounds in Hong Kong (Ho et al. 2006). Various restaurant types were included in the survey that represented different cooking processes and food styles. The authors concluded that, on a total mass emissions basis, the top four carbonyls (formaldehyde, acetaldehyde, acrolein and nonanal) contribute 72% of the carbonyl emissions from commercial kitchens in Hong Kong.
The main goal of this investigation is to extend the data on particulate and condensable vapor emissions from commonly used commercial kitchen appliances and food products that were previously documented in the ASHRAE RP-745 final report dated February 9, 1999 (Gerstler et al. 1999a). As with the previous study, appropriate food products were selected for each appliance to provide significant grease emissions and to be in accordance with corresponding ASTM International test protocol requirements. Table 1 provides the seven appliances that were tested and the food product used for each.
The main particle sampling instrument, the personal cascade impactor, and the same grease vapor sampling instrument, the EPA Method 5 sampling train, were used in this study so the results could be compared directly with the results obtained in ASHRAE RP-745. Data were obtained both in the plume from each appliance, as in the earlier study, and also in the exhaust duct with no grease filters installed in the hood.
One of the observations from the earlier study was that a significant fraction of the particulate mass emissions can occur in the submicron size range. This was especially true for broilers. In the previous study, these small particles were captured primarily by the after filter in the personal cascade impactor and thus their size distribution was not determined. Aerosol measurement instrumentation has evolved since the earlier study, and now several instruments are available to measure particles down to a few nanometers in size. A driving factor in the characterization of ultrafine particle emissions is their impact on human health, as documented in several recent studies (Oberdorster et al. 2005, 2007). Because of the advancements in instrumentation capability and the perception that human health issues are associated with ultrafine particles, two scanning mobility particle sizers (SMPS) were added to the instrumentation package that cover the range from 20 nm to 0.8 [micro]m. Thus, particles from 20 nm to 15 [micro]m in diameter are captured and quantified.
Table 1. List of Appliances Tested and Corresponding Food Products Type Brand And Model Food Product Conveyor Broiler Nieco model 980 Frozen 1/8 lb hamburger patties 10% fat, 55% moisture Clamshell Griddle Garland model Frozen 1/4 lb MWG-9501 hamburger patties 15% fat, 48% moisture Conveyor Pizza Oven Middleby Marshall Frozen 17.75 oz 12" model PS360-WB diameter thin crust pepperoni pizzas Overfired Broiler Vulcan model 5 oz sirloin steaks, Sunglow IR-71P 2.0% fat, 72.3% moisture Electric Steamer Stellar model 5 oz boneless Steam Altair II skinless chicken breasts Mesquite Solid Fuel Holstein Frozen 1/4 lb Broiler Manufacturing hamburger patties 10% model Charcoal fat, 55% moisture Country Club Custom 36" length Gas Chinese Wok Jade model JCR-1 5 oz boneless skinless chicken breasts diced into 1" cubes with peanut oil
The test kitchen is located within the Mechanical Engineering Building at the University of Minnesota and was constructed and used in the previous ASHRAE RP-745 (Gerstler et al. 1999a) study. A schematic drawing of the facility and associated instrumentation is provided in Figure 1 and a list of instrumentation is provided in Appendix A. The construction is of steel frame with floor dimensions 10 x 10 ft (3.05 x 3.05 m) with an inside height of 9 ft (2.74 m). Wood 2 x 4 studs are attached to the steel frame and support 5/8 in. (1.59 cm) fire-rated wallboard. Twenty-gauge stainless steel panels cover the entire back wall behind the appliances. The floor is covered with a single sheet of linoleum cut to fit the interior dimensions of the space. The wall opposite the appliance location is constructed of a two-part removable wood frame covered by window screen. The frames are attached to the chamber walls by hinges with removable pins for moving the appliances in and out of the facility. The screen allows makeup air to be taken from the remaining laboratory space at low velocity so as not to interfere with the thermal plume above the appliance or performance of the exhaust hood.
[FIGURE 1 OMITTED]
An 8 ft (2.44 m) long by 4 ft (1.22 m) wide type 1 wall-mounted canopy listed ventilation hood is bolted to the steel frame that supports the ceiling of the chamber. The hood has an internal depth of 2 ft (0.61 m) and is placed against the back wall of the chamber with the opening at a height of 6.5 ft (1.98 m) above the floor. A 16 in. (40.64 cm) diameter collar connects the hood to a horizontal round stainless steel exhaust duct of the same diameter. The duct runs horizontally to a 16 in. (40.64 cm) 1 hp centrifugal exhaust fan modified by the manufacturer to operate in a horizontal position. The fan exhausts to the outside air. Tempered makeup air is provided by a separate air-handling unit installed in the same room and by drawing additional air from the large building.
Appliance connections along the back wall include one three-phase 208 V electrical outlet, a 4-plug 115 V grounded outlet, and a 1 in. (2.54 cm) natural gas line. In addition, a 4-plug 115 V grounded outlet is mounted inside the right hand wall for instrumentation power as is the variable speed drive for the exhaust fan. A fire extinguisher and a personnel access door are on the left wall near the removable screen panel.
Each appliance was located at the center of the exhaust hood and at least 6 in. (15.2 cm) back from the front of the hood for appropriate capture and containment of the effluent plume. Whenever possible, trained personnel who were familiar with operation of each appliance were present to supervise the connections and to ensure the appliance was calibrated and operating normally before any cooking was initiated. Figure 2 shows the solid-fuel broiler positioned under the exhaust hood before the EPA Method 5 sampling train and the attached personal cascade impactors (PCIs) were moved into position.
[FIGURE 2 OMITTED]
Particles were captured and classified in the size range of 0.5 to 15 [micro]m using Marple model 298 PCIs with model 290 IA in-line adaptors. Short sections of copper tubing, 0.183 in. (4.65 mm) I.D., were given a tapered inlet and a 2.48 in. (63 mm) bend radius to form the 90[degrees] sampling inlets. Substrates were 1.34 in. (34 mm) mylar. Final filters were either 1.34 in. (34 mm) PVC membrane with 5 [micro]m pore size or 1.34 in. (34 mm) glass fiber filters. The air drawn through the impactors was then sent either to the EPA Method 5 sampling train to remove the condensable grease or to a vacuum pump.
Particles between 20 nm and 0.8 [micro]m in size were analyzed by two SMPS, one sampling in the plume and one in the exhaust duct. Sampling inlets were fabricated from copper tubing similar to those for the impactors. Approximately 6 in. (15 cm) from the inlet, the sample was diluted by a factor of 10 with filtered dry air. The diluted sample was then sent to the SMPS for analysis of the particle size distribution. Each scan over the particle size range required approximately four minutes. For appliances with variable emissions, the scanning results must be averaged over several cooking cycles to provide data comparable to appliances with more steady emissions.
A third sample of the effluent in the plume was taken through 1.85 in. (47 mm) open-face quartz filters for subsequent chemical composition analysis. A second filter was placed downstream of the primary filter to capture anything that passed through it. Aerosol samples in the exhaust duct were captured in a separate copper sampling line and sent to an aerosol time-of-flight mass spectrometer (ATOFMS) for determination of particle chemistry by size. The chemistry results will be provided in a separate paper.
A sketch of the instrumentation layout in the test facility is shown in Figure 1. All the instrumentation associated with sampling in the plume was located within the test kitchen. The instrumentation used to sample and characterize effluent in the exhaust duct was located in the surrounding laboratory space. The dashed lines in Figure 1 indicate the instruments that were used for only some of the tests. A photograph of a representative effluent sampling setup for the conveyor broiler is shown in Figure 3. All three sampling inlets (PCI, SMPS, filter holder) were located as close to the center of the effluent plume as possible.
[FIGURE 3 OMITTED]
Other instrumentation included type T thermocouples for air temperature measurement, type K thermocouples for appliance temperature measurements, a portable hot film anemometer for air velocity measurements in the exhaust duct, an optical particle counter for checking the uniformity of aerosol distribution in the exhaust duct, and an electrical power data logger and a natural gas flowmeter to measure the energy input rates into the appliances.
After each appliance was installed and operating correctly, several preliminary cooking runs were made following the appropriate ASTM procedure developed for that appliance. The results from these preliminary tests were used to determine the amount of food product to use, the cooking time to achieve the desired weight loss and/or internal temperature, and the number of batches or length of cooking needed to capture sufficient effluent for analysis.
Each of the impactor substrates was desiccated and weighed prior to installation into the substrate holders and then into the impactor assembly. Clean after filters were installed and the impactors connected to the end of the Method 5 heated probe or the vacuum pump. The completed Method 5 assembly with the impactor at the front end was then positioned so that the inlet was as close to the center of the effluent plume as possible.
The center of the plume was determined by mapping the air temperature near the bottom of the exhaust hood and determining the point with the highest temperature. Clean quartz filters were installed into the filter holders, and the filter assemblies were positioned under the hood and connected to the vacuum pump. An SMPS, dilution air system, and computer were all located inside the test kitchen opposite the personnel door. The sampling inlet was repositioned slightly from appliance to appliance as the plume center changed location.
Sampling from the exhaust duct was accomplished using an isokinetic probe positioned in the centerline of the duct. A second PCI assembly was connected to a flow-calibrated laminar flowmeter to determine the correct sampling airflow rate. This was connected to a vacuum pump except the runs where the method 5 sampling train was used to determine grease vapor concentration in the exhaust duct. The SMPS shared the same sampling line with the PCI but no dilution air was used as the concentration in the exhaust duct was sufficiently low so that dilution was not necessary.
Once all the instrumentation was positioned, the appliance was turned on and allowed to come to operating temperature. At the beginning of the cooking, all vacuum pumps were turned on simultaneously so that all of the instruments sampled during the same time interval. The SMPS instruments were restarted several times during a run to take sweeps over the particle size range. Once sufficient time had elapsed, the cooking was terminated and all pumps shut off. The impactors were then disassembled and the stage substrates allowed to dry in a desiccator for at least 24 hours. Once the substrates were dry, the final weight measurements were made that were used to determine the mass of particles collected for the size range that corresponded to each impactor stage. The Method 5 impingers were washed out with acetone and placed into preweighed beakers. The solvent was allowed to evaporate, and the final weight was measured to determine the amount of condensable vapor collected. The quartz filters were removed from the filter holders, put into sealed bags, and placed in a freezer for subsequent chemical analysis. The data from the SMPS measurements were curve fitted using a commercial software package that assumed a log-normal distribution and determined the mean particle size and geometric standard deviation.
A minimum of three tests were run as described above for each appliance, with the Method 5 sampling train used to determine the condensable grease effluent in the plume under the hood. At least one run was then made with the Method 5 used to sample effluent in the exhaust duct.
RESULTS AND DISCUSSION
Results are separated into total grease mass particulate and vapor levels measured in the plume and exhaust duct and particle size distributions measured by both the personal cascade impactor and the scanning mobility particle sizer in the plume and duct.
Fraction of the Plume that was Sampled
In the ASHRAE RP-745 (Gerstler et al. 1999a) study, small particles collected on the lower stages of an electrical low pressure impactor (ELPI) or MOUDI in the plume and in the exhaust duct were used to determine the dilution ratio. This is needed to determine the fraction of the plume that was sampled and to calculate the total mass emissions in the plume. However, for this study there were three possible methods for determining the dilution ratio: (1) SMPS integrated number concentrations in the plume and duct, (2) particle mass collected on the PCI after filters in the plume and duct, or (3) mass of condensable grease vapor collected using the EPA Method 5 apparatus in the plume and duct. The assumption is made that negligible transfer of mass from particle-to-vapor phase and vice versa occurs between the plume sampling point and the sampling point in the exhaust duct. This is a reasonable approximation, as the effluent time interval between the two sampling locations is approximately one second, and the effluent is already diluted and cooled by the time it reaches the sampling point in the plume. A similar observation was made by Schauer (Schauer et al. 1999), where the transfer of semivolatile organic compounds from meat cooking appeared to be prohibited from transferring between the gas phase and particle phase. The authors attributed this to the solid character of meat fat at room temperature, which has a relatively low vapor pressure. When looking at each of these possibilities, it was found that the SMPS raw number concentrations were sometimes greater in the duct than in the plume. This occurred when the plume was not well defined-for example with the clamshell griddle, where the center of the plume varied from side to side, and with the conveyor pizza oven, where the effluent was emitted out of both ends of the oven. In these cases the results led to dilution ratios less than one and indicated that normalized mass concentrations in the duct would be greater than in the plume, which is unrealistic. The second possibility was to use the mass collected on the PCI after filters. Here it was found that, in some cases, the dilution ratios were large enough that calculated normalized condensable grease concentrations in the duct would be greater than in the plume, which is also unrealistic. The third possibility of using dilution ratios calculated from the condensable grease vapor collected in the plume and duct provided the most realistic results, where the normalized particle concentrations in the plume were always greater than or equal to those in the duct. Determining dilution ratios from collected grease vapor in the plume and duct seems to be reasonable since there is no significant grease vapor removal mechanism between the two sampling locations. Therefore, the dilution ratio for each appliance was calculated by dividing the concentration of condensable grease vapor collected in the plume by the concentration of condensable grease vapor collected in the duct.
Grease Particulate and Vapor Levels in Plume and Exhaust Duct
The total normalized grease mass emissions measured in the plume and exhaust duct from each appliance are shown in Figures 4 and 5, respectively, with the results broken down into vapor, PM2.5, PM10-PM2.5 and particles larger than 10 [micro]m. Numerical values are provided in Appendix B--Table B1 for the results in the plume and Table B2 for the duct. Note that no grease filters were used in the hood during these tests, so the emissions in the exhaust duct are different than what would be expected in a real cooking scenario where grease filters or extractors are present. This was a requirement of the initial RFP and provided information on natural effluent elimination by the exhaust system between the plume and exhaust duct without the use of grease filters. Results from the gas conveyor broiler cooking hamburger have an emission distribution that is comparable to the previous results for the gas and electric underfired broilers cooking hamburger (Kuehn et al. 1999). The total grease mass emissions from the clamshell griddle, conveyor pizza oven, and overfired broiler cooking steak are relatively low and comparable to several of the lower-emitting appliances tested previously. Grease vapor is the predominant emission from all these appliances. No reliable PCI or SMPS data were obtained in the plume of the steamer, as the high water content precluded accurate measurements there. The emissions from the solid fuel broiler cooking hamburger are similar in mass distribution but more than twice the quantity of those from the gas underfired broiler cooking hamburger from the RP-745 study. The wok generated huge amounts of large particles in the plume primarily caused by spatter or mechanical particle generation that caused the total emission to be a factor or two larger than the total from any other appliance.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
The total particulate mass emissions in the plume for three of the appliances--the clamshell griddle, conveyor pizza oven, and overfired broiler--are very similar to those measured in the exhaust duct. The particulate mass emissions for the conveyor broiler and the solid-fuel broiler are significantly higher--twice as high in the plume as in the exhaust duct. This can be explained by the higher proportion of mass in the largest particle size measured that indicates that some of these particles do not make it to the exhaust duct sampling probe and may be lost by impaction or settling. The wok has much higher particulate mass emission in the plume than in the exhaust duct by a factor of about five. Again, this can be explained by the overwhelming amount of large grease particles in the plume that are not found in the exhaust duct.
McDonald et al. (2003) compared their PM 2.5 mass emission results with those of previous investigators. The results closest to the present study are the emissions using a conveyor broiler with hamburger. They obtained a value of 4.5 lb (kg)/1000 lb (kg) food product, whereas Norbeck et al. (1997) measured 7.4 lb (kg)/1000 lb (kg) food product. Both used hamburger with 21% fat. The present results were obtained with 10% fat hamburger and give a PM 2.5 value of 8.0 lb (kg)/1000 lb (kg) food product. Both of these previous studies used considerably different methods than the particle/ vapor sampling procedures used here, as they were focused on emissions into the environment. McDonald obtained the sample from the exhaust duct and passed it into a dilution tunnel with a dilution factor of 30:1 followed by a mixing chamber with a 90 s residence time to simulate the conditions expected after release into the atmosphere. In the present study, samples were passed as quickly as possible into the particle and vapor sampling instruments to characterize the nature of the effluent stream in the plume and duct, which does not necessarily correspond to the ultimate fate in the atmosphere. When the masses associated with other particle sizes from this study are added (2.8 lb [kg]/1000 lb [kg] food product) and the condensable vapor is added (16.9 lb [kg]/ 1000 lb [kg] food product), the total normalized mass emissions measured here become 27.6 lb (kg)/1000 lb (kg) food product.
Particle Size Distributions
The particle size distributions in the plume and exhaust duct were determined for each appliance using the PCI for the larger sizes and the SMPS for the smaller sizes. Particles that impacted inside the PCI sampling inlet tube were removed by swabbing with acetone, and the results were added to the mass removed from the impactor inlet and Stage 1. Thus the smallest particles included in this total were larger than the cut size for Stage 1. However, the largest size is unknown. As in the previous study, the maximum particle size was assumed to be 100 [micro]m. The results for Stages 2 through 8 are well characterized, as the cut sizes for each of these stages is known as a function of airflow rate. The largest particles captured on the final filter are equal to the cut size for Stage 8 of the impactor. However, the smallest particle size is unknown. The smallest size is assumed to be 10 nm (0.01 [micro]m), except where noted. The results from the SMPS are used to provide more detailed information on these small particles. It is also important to note that the impactor provides mass concentration data, whereas the SMPS provides number concentration data. These results can be combined, if the particle density is assumed, so that the number concentration results can be converted to mass concentration for direct comparison. All the particle size results are included in the project final report (Kuehn et al. 2008). Results from the conveyor broiler are presented here as representative.
Figure 6 shows the particle size data from three runs with the personal cascade impactors sampling from the plume and exhaust duct for the conveyor broiler cooking hamburger. The results are given in terms of particle mass per unit volume. The sampling airflow rate changed slightly from run to run, so the cut sizes of the impactor stages are slightly different for each of the three runs. The general size distribution is similar between the plume and exhaust duct, although there are fewer large particles in the duct. Note that the concentration of the smaller particle sizes decreases dramatically between the plume and the duct by about a factor of four. This is to be expected because of the large amount of dilution air brought in by the hood.
[FIGURE 6 OMITTED]
The impactor and its sampling inlet capture particles that range in size from 0.5 to about 100 [micro]m. Particles smaller than 0.5 [micro]m are captured on an after filter that does not provide resolved particle size information. To resolve the particle size distribution in the size range, SMPSs were used. Figure 7 shows the corresponding results from the SMPS measurements taken in the plume and in the exhaust duct. The plots show number concentration mean and standard deviation for each SMPS channel calculated from a minimum of three scans per run with three runs in the plume and four in the duct. Therefore, at least nine scans were taken in the plume and 12 in the duct that were used in calculating the means and standard deviations.
[FIGURE 7 OMITTED]
The size distributions measured with the SMPS instruments show that the maximum particle number concentrations occur near 100 nm (0.1 [micro]m) in size. This is the result of heterogeneous nucleation, where vapor condenses onto small nuclei, and these droplets continue to grow until they reach about 100 nm in size. Similar size distributions have been measured in many previous aerosol studies. This physical phenomenon is described in more detail in the text by Hinds (Hinds 1999).
To offer additional information about the SMPS data, a commercial data reduction program was used to fit a curve to the mean values under the assumption that the data are log-normally distributed. The results of the fit indicate that the geometric mean diameters are 0.138 and 0.144 [micro]m, and the geometric standard deviations are 1.73 and 1.65 for the average results in the plume and the duct, respectively. This indicates that the size distributions are nearly identical between the plume and duct sampling points and corroborates the assumption made earlier that negligible grease transfer between particle and vapor phases occurs between the two locations. However, the concentrations in the duct are reduced significantly because of dilution. Geometric mean particle diameters and corresponding geometric standard deviations from the SMPS data for all appliances tested are provided in Table 2. The conveyor broiler and the solid fuel broiler have mean diameters of approximately 0.15 [micro]m, whereas all the others are less than 100 [micro]m.
The particle mass concentration results obtained with the PCI shown in Figure 6 and the small particle number concentration data obtained using the SMPS shown in Figure 7 can be compared if the particles are given an assumed density so that the particle number concentration data can be converted into mass concentration. Using an assumed grease particle density of 0.90 g/c[m.sup.3], the SMPS results are overlayed onto the PCI results for the plume and exhaust duct in Figure 8.
[FIGURE 8 OMITTED]
Table 2 Particle Mean Diameters and Geometric Standard Deviations from the Scanning Mobility Particle Sizer Data from all the Appliances Tested Plume Exhaust Duct Geometric Geometric Appliance Mean, Standard Mean, Standard [micro]m Deviation [micro]m Deviation Conveyor Broiler 0.138 1.73 0.144 1.65 Clamshell Griddle 0.086 2.14 0.098 1.76 Conveyor Pizza Oven 0.036 1.53 0.038 1.53 Overfired Broiler 0.044 1.57 0.049 1.54 Steamer <0.02 Solid Fuel Broiler 0.173 1.79 0.145 1.66 Wok 0.054 2.14 0.055 1.88
A lower limit of 0.01 [micro]m (10 nm) has been assumed for particles captured by the final filter used on the impactor, as was assumed in the previous ASHRAE study (Gerstler et al. 1999b). However the results from the scanning mobility particle sizer now provide much better information on the small particle size range. Changing the assumed lower particle size limit from 0.01 to 0.15 [micro]m for the final filter used with the conveyor broiler and more consistent with the results from the SMPS and replotting the results in the plume given in Figure 8a provides the results shown in Figure 9. The mass concentration results provided by combining the data from the two instruments are now more consistent. Similar adjustments can be made for the other appliances that were tested. However, two assumptions need to be made: (1) a grease particle density and (2) a lower particle size limit for the final filter in the personal cascade impactor.
[FIGURE 9 OMITTED]
SUMMARY AND CONCLUSIONS
The total grease mass emissions in the plume were found to be comparable to the emissions documented in the ASHRAE RP-745 (Gerstler et al. 1999a) study for some of the appliances. Results from the conveyor broiler agree well with previous results from the underfired broiler cooking hamburger. Emissions from the conveyor pizza oven were the lowest at approximately 2.5 lb (kg)/1000 lb (kg) food product and agreed with the results from the previous pizza ovens. Emissions from the clamshell griddle were similar to those from the electric griddle tested previously. The overfired broiler cooking steak generated between 10 and 15 lb (kg)/ 1000 lb (kg) food cooked that consisted primarily of grease vapor and compared well with the results from the underfired broilers cooking chicken in the earlier study.
Both the solid-fuel broiler cooking hamburger and the wok cooking diced chicken breast in peanut oil generated huge amounts of large particles in the plume. However very little of this was found in the exhaust duct, although no filters were installed in the hood.
Although large amounts of grease mass corresponding to particles larger than 10 [micro]m were found in the plume from the solid fuel broiler and the wok, nearly all of the grease mass emission in the exhaust duct was found to be in the vapor phase or associated with particles smaller than 1 [micro]m in size. Neither of these effluent components can be easily removed by inertial impaction utilized by most current technology grease filters. Thus, novel grease removal technologies should be developed and implemented to better control grease emissions from commercial kitchens using the appliances tested here. High concentrations of ultrafine particles similar to what were measured in this study have been shown to increase human health risk associated with respiratory exposure. This issue should be investigated further to determine what levels of emission control are needed based on a variety of factors, including economic impacts on the food service industry and the risks to worker health.
We would like to acknowledge several companies and individuals who helped make this research possible:
* Metal-Fab for donating the exhaust duct
* Greenheck for the use of the optical particle counter, exhaust hood, and fan
* Burger King for supplying the conveyor broiler, a technician to assist in the proper setup, and the food product used (hamburger patties)
* McDonalds, Garland, Martin Brower, and OSI for supplying the clamshell griddle and the food product (hamburger patties)
* PG&E Food Service Technology Center for supplying the conveyor pizza oven, overfired broiler, steamer, solid-fuel broiler and wok
* Kevin Stockman for cooking during the wok tests
* Fisher Nickel, Inc. for coordinating the cofunding needed for the chemical analyses
Finally, we would like to thank the members of ASHRAE TC 5.10 and the members of the RP-1375 Project Monitoring Subcommittee for their suggestions, guidance, and interest in this investigation.
ASTM Standard Test Method for Performance of Steam Cookers, F 1484-05
ASTM Standard Test Method for Performance of Double-Sided Griddles, F 1605-95
ASTM Standard Test Method for Performance of Underfired Broilers, F 1695-03
ASTM Standard Test Method for Performance of Conveyor Ovens, F 1817-97
ASTM Standard Test Method for Performance of Chinese (Wok) Ranges, F 1991-99
ASTM Standard Test Method for Performance of Upright Overfired Broilers, F 2237-03
ASTM Standard Test Method for Performance of Conveyor Broilers, F 2239-03
BAAQMD. 2007. Regulation 6: Particulate matter, Rule 2: Commercial cooking equipment. Bay Area Air Quality Management District, San Francisco, CA.
Dennekamp, M., S. Howarth, C.A.J. Dick, J.W. Cherrie, K. Donaldson, and A. Seaton. 2001. Ultrafine particles and nitrogen oxides generated by gas and electric cooking. Journal of Occupational and Environmental Medicine 58:511-16.
Gerstler, W.D., T.H. Kuehn, D.Y.H. Pui, J.W. Ramsey, and M.P. Doerr. 1996. Identification and characterization of effluents from various cooking appliances and processes as related to optimum design of kitchen ventilation systems. ASHRAE RP-745, Phase 1 Final Report. Atlanta: American Society of Heating, Refrigerating and Air Conditioning Engineers, Inc.
Gerstler, W.D., T.H. Kuehn, D.Y.H. Pui, J.W. Ramsey, M. Rosen, R.R. Carlson, and S.D. Petersen. 1999. Identification and characterization of effluents from various cooking appliances and processes as related to optimum design of kitchen ventilation systems. ASHRAE RP-745, Phase 2 Final Report. Atlanta: American Society of Heating, Refrigerating and Air Conditioning Engineers, Inc.
Gerstler, W.D., T.H. Kuehn, D.Y.H. Pui, and J.W. Ramsey. 1999. Measurements of the effluent from hamburger cooked on a gas underfired broiler. ASHRAE Transactions 105(2):303-15.
Hinds, W.C. 1999. Aerosol Technology, 2nd ed. Hoboken, NJ: John Wiley & Sons.
Ho, S.S.H., J.Z. Yu, K.W. Chu, and L.L. Yeung. 2006. Car-bonyl emissions from commercial cooking sources in Hong Kong. Journal of Air & Waste Management Association 56:1091-98.
Kleeman, M.J., J.J. Schauer, and G.R. Cass. 1999. Size and composition distribution of fine particulate matter emitted from wood burning, meat charbroiling and cigarettes. Environmental Scince Technology 33:3516-23.
Kuehn, T.H., W.D. Gerstler, D.Y.H. Pui, and J.W. Ramsey. 1999. Comparison of emissions from selected commercial kitchen appliances and food products. ASHRAE Transactions 105(2):128-41.
Kuehn, T.H., B.A. Olson, J.W. Ramsey, J. Friell, and J.M. Rocklage. 2004. Development of a standard method of test for commercial kitchen grease removal systems. Final Report, Department of Mechanical Engineering, University of Minnesota, Minneapolis, MN.
Kuehn, T.H., B.A. Olson, J.W. Ramsey, J. Rocklage. 2008. Characterization of effluents from additional cooking appliances. Final Report, ASHRAE 1375-RP, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., Atlanta.
McDonald, J.D., B. Zielinska, E.M. Fujita, J.C. Sagebiel, J.C. Chow, and J.G. Watson. 2003. Emissions from charbroiling and grilling of chicken and beef. Journal of Air & Waste Management Association. 53:185-94.
Oberdorster, G., E. Oberdorster, and J. Oberdorster. 2005. Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles. Environmental Health Perspectives 113:823-39.
Oberdorster, G., V. Stone, and K. Donaldson. 2007. Toxicity of nanoparticles: A historical perspective. Nanotoxicol-ogy 1(1):2-25.
Olson, D., and J.M. Burke. 2006. Distributions of PM2.5 source strengths for cooking from the Research Triangle Park particulate matter panel study. Environmental Science and Technology 40:163-69.
E.H. Pechan & Associates, Inc. 2003. Technical Memorandum: Methods for developing a national emission inventory for commercial cooking processes. Report prepared for the Emission Factor and Inventory Group, Emissions Monitoring and Analysis Division, Office of Air Quality Planning and Standards, U.S. EPA, Research triangle Park, NC.
SCAQMD. 1997. Rule 1138: Control of emissions from restaurant operations. South Coast Air Quality Management District, Diamond Bar, CA.
Schauer, J.J., M.J. Kleeman, G.R. Cass, and B.R.T. Simonet. 1999. Measurement of emissions from air pollution sources. 1. C1 through C29 organic compounds from meat charbroiling. Environmental Science and Technology 33:1566-77.
Svendsen, K., H.N. Jensen, I. Sivertsen, and A.K. Sjaastad. 2002. Exposure to cooking fumes in restaurant kitchens in Norway. Annual of Occupational Hygiene 46(4):395-400.
Wallace, L. 2000. Real-time monitoring of particles, PAH, and CO in an occupied townhouse. Applied Occupational and Environmental Hygiene 15(1):39-47.
Wallace, L.A., S.J. Emmerich, and C. Howard-Reed. 2004. Source strengths of ultrafine and fine particles due to cooking with a gas stove. Environmental Science and Technology 38:2304-11.
Wallace, L. 2006. Indoor sources of ultrafine and accumulation mode particles: size distributions, size resolved concentrations, and source strengths. Aerosol Science and Technology 40:348-60.
1. Personal Cascade Impactor (PCI)
a. Marple model 298 with model 290 IA in-line adaptor
b. 90[degrees] sampling probe: copper tubing, machine tapered inlet, 2.48 in. (63 mm) bend radius, 0.183 in. (4.65 mm) inside diameter
c. 1.34 in. (34 mm) mylar substrates: model c-290-MY
d. 1.34 in. (34 mm) PVC membrane after-filters: 5 um pore size, model F-290-P5
e. 1.34 in. (34 mm) glass fiber after-filters: Spectro grade, type A, Gelman filtration media
2. Total Filter Sample
a. 1.85 in. (47 mm) Millipore open-face filter holder
b. 1.85 in (47 mm) quartz filters.
3. Scanning mobility particle sizer (SMPS) used for plume measurements
a. Differential mobility analyzer, TSI model 3071
b. Condensation particle counter, TSI model 3025
4. Scanning mobility particle sizer (SMPS) used for duct measurements
a. Differential mobility analyzer, manufactured in house
b. Condensation particle counter, TSI model 3010
5. Aerosol time-of-flight mass spectrometer (ATOFMS), TSI series 3800
6. Optical particle counter (OPC), Climet model Spectro 0.3 airborne
Condensable Grease Vapor
1. EPA Method 5 sampling train, Graseby Andersen universal stack sampler
2. Condensation vapor monitor constructed in house
1. Air: Type T thermocouple, 24 gauge wire, Teflon PFA insulation (500[degrees]F, 260[degrees]C).
2. Appliance: Type K thermocouple, 20 gauge wire, glass braid insulation (900[degrees]F, 482[degrees]C).
3. General digital multimeter, Keithley 132F TRMS Multimeter
1. Portable hot-wire anemometer, TSI model 8330 VelociCheck
Appliance Energy Measurement
1. Electrical power data logger, Synergistic model C 180
2. Natural gas flowmeter with 1/8 [ft.aup.3] digital resolution
1. Analytic microbalance, Cahn C-31, 0-25 mg, 0.1 mg resolution
2. Fine analytic balance, Sartorius B-120S, 0-110g, 0.1 mg resolution
3. Course analytic balance, Sartorius L-610, 0-610 g, 0.01 g resolution
4. Kitchen scale, Pelouze model Y32R, 0-32 oz resolution
Table B1. Normalized Grease Mass Emission Results for Six Appliances Measured in the Plume Below the Hood lb. Emissions/1000 lb. Food Product Appliance Dp > 10 2.5 um < Dp Dp [less than Condensable Total [micro]m [less than or or equal to] Vapor equal to] 10 2.5 [micro]m [micro]m Conveyor 22.1 0.69 10.5 16.9 50.2 Broiler (Hamburger) Clamshell 2.94 0.58 0.70 9.15 13.4 Griddle (Hamburger) Conveyor 0.19 0.02 0.03 2.39 2.64 Oven (Pizza) Overfired 2.01 0.19 1.16 7.44 10.8 Broiler (Steak) Solid-Fuel 55.2 2.68 35.4 48.9 142 Broiler using Mesquite (Hamburger) Chinese Wok 188 4.89 11.2 43.0 247 (Cubed Chicken Breast in Peanut Oil) Table B2. Normalized Grease Mass Emission Results for All Seven Appliances Measured in the Exhaust Duct with No Grease Filters Present lb. Emissions/1000 lb. Food Product Appliance Dp > 10 2.5 um < Dp Dp [less than Condensable Total [micro]m [less than or or equal to] Vapor equal to] 10 2.5 [micro]m [micro]m Conveyor 2.35 0.41 7.97 16.9 27.6 Broiler (Hamburger) Clamshell 0.81 0.41 0.37 9.15 10.7 Griddle (Hamburger) Conveyor 0.17 0.03 0.05 2.39 2.64 Oven (Pizza) Overfired 1.56 0.14 0.78 7.44 9.93 Broiler (Steak) Steamer 1.16 0.07 0.06 13.7 15.0 (Chicken Breast) Solid-Fuel 0.46 1.69 21.4 48.9 72.5 Broiler Using Mesquite (Hamburger) Chinese Wok 4.03 3.00 5.62 43.0 55.6 (Cubed Chicken Breast in Peanut Oil)
Thomas H. Kuehn, PhD, PE
James W. Ramsey, PhD
Bernard A. Olson, PhD
Joshua M. Rocklage
Student Member ASHRAE
This paper is based on findings resulting from ASHRAE Research Project RP-1375.
Thomas H. Kuehn is a professor and director of the Environmental Division, Bernard A. Olson is a research associate, James W. Ramsey is a professor and associate department head, and Joshua M. Rocklage is a research assistant in the Department of Mechanical Engineering, University of Minnesota, Minneapolis, MN.
|Printer friendly Cite/link Email Feedback|
|Author:||Kuehn, Thomas H.; Ramsey, James W.; Olson, Bernard A.; Rocklage, Joshua M.|
|Date:||Jan 1, 2009|
|Previous Article:||Calibration of a building energy model using measured data.|
|Next Article:||Comparisons of HVAC simulations between EnergyPlus and DOE-2.2 for data centers.|