Characterization and kinetic evaluation of UV system used to treat commercial kitchen exhaust emissions

Commercial cooking operations are major sources of air-borne particulate matter (particle size mu 2.5 mm) and volatile organic compounds in urban environments. Charbroiling of meat has reportedly produced up to 40g particulate matter (PM) per kg of meat cooked. With the current annual increase rate of 4.1 % in the number of fast-food restaurants (Paeratakul et al., 2003) and also with the tight regulations on release of emissions from commercial kitchen exhausts into the atmosphere, treatment of kitchen exhaust emission prior to the release into atmosphere has become compulsory by RULE 1138 in South Coast Air Quality Management District (SCAQMD), USA. The following research determined the effectiveness of an advanced oxidation process (UV system) developed by Franke, Inc., Food Service Systems for treating kitchen cooking exhaust emissions. While cooking various products, the VOCs and aerosols present in the exhaust emissions were characterized and quantified. A modified EPA Method TO-17 protocol was used to sampling air using sorbent tubes. The samples were analyzed by a thermal desorption-capillary GC/MS analytical procedure. Results were analyzed to determine the effectiveness of the UV treatment system over selected fast-food commercial cooking operations. A surrogate compound (benzene) was selected from the characterized commercial kitchen emissions based on frequency of occurrences, chemical structure, volatility, and toxicity to humans. By varying concentration levels, humidity, and reaction medium, the gas phase photooxidation kinetics of selected benzene was evaluated. A line source spherical emission (LSSE) was used to evaluate the average irradiance of the reactor used in this study. A kinetic model for gas phase photodegradation of benzene was developed based on the literature available on gasphase photodegradation of benzene, reactive species involved in the photochemistry of atmosphere and the current experimental studies. Thus developed model was used to predict concentrations and compare with the experimental concentrations. Initial photooxidation rates at time (t) = 0,from experiment and model followed the order: with 33% RH air > 65%RH air > dry air (5% RH) > 65%RH nitrogen > dry nitrogen (5% RH).