Since the late 1970's, it has become well established that the combustion of certain fuels containing both organic material and chlorides can form polychlorinated dibenzo-p-dioxins (CDDs) and poylchlorinated dibenzofurans (CDFs). This discovery has prompted world-wide research to identify combustion sources, to characterize the conditions favoring the formation of CDDs and CDFs within the combustion process, and to characterize the emission of dioxin-like compounds to the air from the stack of the process.

The purpose of this chapter is to provide site-specific procedures for evaluating the emission of dioxin-like compounds from stationary combustion sources. The first step is to characterize stack emissions in terms of mass of TCDD/F congener released, and then to partition that release into a vapor and a particle phase. Using atmospheric transport modeling, these releases are translated to ambient air vapor and particle phase concentrations, and wet and dry particulate deposition amounts, in the vicinity of the release. This chapter demonstrates these procedures on a hypothetical incinerator using an air dispersion model called COMPDEP. A second purpose of this chapter, therefore, is to provide the background and justification for the model inputs and key parameters for COMPDEP. The final results for this model simulation are vapor and particle phase concentrations, and particulate deposition amounts of the specific dioxin-like congeners, which are then used for the demonstration of the stack emission source category in Chapter 5 of this Volume.

This chapter is structured as follows:

• Section 3.2 describes the generation of congener specific emission factors. These factors are defined as the mass of congener emitted per mass of feed material combusted. Subsections within Section 3.2 discuss: 1) a heirachy of preferred options for generating such emission factors, starting with site-specific stack testing for specific congeners and ending with engineering evaluations when no other data is available, 2) an approach to estimating congener-specific emission factors if homologue group emission factors are all that is available, including a presentation of source category homologue group profiles generated with limited data, 3) the emission factors for the example incinerator demonstrated in Chapter 5, and assuming a feed rate into the example incinerator, emissions expressed on a mass per time basis (which is required for transport modeling), 4) partitioning of emissions into a vapor and a particle phase for atmospheric transport modeling, and 5) a procedure to estimate the mass released and concentrations for a related emission of a combustor, that of ash. It is noted that a similar discussion on emission factors is presented in Chapter 3 on Sources in Volume II of this assessment. However, the Volume II discussion derives national average emission factors for various combustor types as a basis for estimating total annual releases to the air in the United States from each combustion source category in comparative units of grams of Toxic Equivalent (TEQs) emissions per year.

• Section 3.3 describes a general air modelling procedure for evaluating the fate and transport of dioxin-like compounds emitted from stacks. The discussion presents a general review of dispersion theory, a general review of dry particle deposition, and a general review of the wet deposition algorithm employed in this analysis. EPA's COMPDEP air dispersion and deposition model is reviewed. Wherever pertinent, Section 3.3 describes the assumptions and parameter values that were used in the demonstration of methodologies in Chapter 5 of this Volume.

• Sections 3.2. and 3.3. summarized input data and assumptions (emission rates, vapor/particle partitioning assumptions, etc.) that were made for the demonstration of the methodologies for evaluating stack emissions in Chapter 5 of this Volume. Section 3.4 supplies all other key assumptions for the stack emission demonstration, such as stack height and exit temperatures, meteorological data, and others. This Section also provides the final results from the COMPDEP modeling, including vapor phase air concentrations at various distances in the predominant wind direction, and dry and wet deposition fluxes, also at various distances in the predominant wind direction.

• Section 3.5 closes out the chapter by summarizing critical aspects for making site specific evaluations of stack emission sources.

Estimating the emission factor is the first step in assessing a specific stack emission source of dioxin-like compound release. For this assessment, an "emission factor" is defined as the total mass (in vapor and particulate forms) of dioxin-like compound emitted per mass of feed material combusted.

This assessment recommends the generation of emission factors for individual dioxin-like congeners for a site-specific assessment. Chapter 3 on Sources in Volume II of this assessment also derives emission factors. However, the Volume II discussion derives national average emission factors for TEQ emissions for various combustor types as a basis for estimating total annual releases to the air in the United States in comparative units of grams TEQ per year.

In the past, EPA converted the concentrations of TCDD/F mixtures into an equivalent concentration of 2,3,7,8-TCDD (Cleverly, et al., 1989, 1991; EPA, 1987a; Mukerjee and Cleverly, 1987) when deriving an emission factor. The fate, transport, and transfer parameters of 2,3,7,8-TCDD were applied to model the environmental fate of this TEQ mixture. This perspective has been changed in the following procedure. In order to increase the level of accuracy, the air dispersion modeling was done separately for each of the toxic congeners. Only at the interface to human exposure, e.g., ingestion, inhalation, dermal absorption, etc., are the individual congeners recombined and converted into the toxic equivalence of 2,3,7,8-TCDD to be factored into the quantitative risk assessment. It is recommended that a site-specific assessment, or even an assessment of a source category, be based on congener-specific emissions rather than TEQ emissions.

Section 3.2.1 presents a strategy for development of emission factors for conducting a site-specific assessment. Section 3.2.2 describes an approach to estimating congener-specific emission factors when all that is available are homologue group emission factors. Section 3.2.3 summarizes the emission factors for the hypothetical incinerator demonstrated in Chapter 5. Section 3.2.4 presents an in-depth evaluation of the partitioning of emissions between a vapor and a particle phase for further atmospheric transport modeling. This discussion includes subsections on measurements of partitioning at the stack, measurements of partitioning in ambient air, and the theoretical approach used in this assessment for vapor/particle partitioning. Section 3.3.5 closes this section on emission factors by describing procedures to estimate the mass of ash (fly and bottom) produced and the concentration of dioxin-like compounds on ash.

The following is a heirarchal listing of data collection options for emission factors:

A. For facilities that are built and operational, it is preferred that direct stack measurements be used, using EPA recommended congener-specific stack monitoring and analytical protocols. Stack monitoring provides concentrations and mass release rates of the pollutant, actual volume of stack gas and temperature. Care should be taken to ensure that the emissions characterization reflects a wide range of operating conditions and also accounts for deterioration in emissions output of the facility over its useful life. Procedures to convert data expressed in concentrations or mass release rates to an emission factor are as follows:

1. Test data of emissions are first placed into common units of measurement. English units are converted into metric, and the concentration term (mass of pollutant per unit volume of combustion gas emitted from the stack) should be corrected to the standard temperature and pressure on a dry gas basis, and standard percent carbon dioxide or oxygen within the combustion gas (e.g.,12% CO₂). These adjustments may be necessary if more than one test occurred for stack emissions.

2. The next step involves converting the mass emission concentration of the specific dioxin-like congener in units of nanograms per normal cubic meter (at standard temperature and pressure) of combustion gas corrected to 12% carbon dioxide into an equivalent emission factor in units of grams of pollutant emitted from the stack per kilogram of combustable material or feed (g CDD and CDF/kg feed) that was incinerated at the facility during the duration of stack sampling. This is done as follows:

 

where:

E_f = emission factor, µg/kg

C_fg = concentration in flue gas, ng/dscm

V_fg = volume of combustion gas/unit of time, dscm/hr

M_w = mass of waste incinerated/unit of time, kg/hr

0.001 = units conversion factor

3. As a final step, the average emission factor of each congener is derived by summing the emission factors and dividing by the number of data points used. The average should represent an approximation of long-term emissions. Many air dispersion models require emission factors in units of amount of the pollutant emitted per second of time. Therefore the average emission factor must be adjusted accordingly by adjusting the units in Equation (3-1) to a time-scale of one second.

B. For facilities that have been constructed, but not yet operational, or are in the planning stages of development, the following procedure is recommended:

1. Collect and review stack test reports which have measured the emissions of specific dioxin-like congeners from facilities that are most similar in technology, design, operation, capacity, fuel, waste feed composition, and pollution control as the facility under consideration.

2. Determine if the stack test reports used EPA recommended stack monitoring and analytical protocols specific to dioxin-like compounds and discard those data not in conformance.

3. When combining data from a test results of a number of facilities, care should be taken to convert emissions, process feed rates and stack gas parameters to consistent units of measurement.

4. Ranges and average values should be developed for purposes of exposure analysis.

C. If no congener-specific data exists for a specific facility or similar facilities, then use homologue profile emissions from similar facilities. Steps 1-4 in B above pertain here as well. Estimating congener-specific emission factors given homologue emission factors is described in Section 3.2.2 below.

D. If no data exist relevant to a specific facility, then the Compilation of Air Pollution Emission Factors (EPA, 1985; and subsequent updates), should be used. This compilation was put together and is periodically updated by EPA's Office of Air Quality Planning and Standards (OAQPS), and is commonly referred to as AP-42. Care should be taken to select emission factors which were developed for technologies that best match the facility under consideration. The basic limitation of these of these data is the fact that emission factors are not usually reflective of specific emission control equipment.

OAQPS's AP-42 document provides TCDD/F emission factors for municipal waste combustors, sewage sludge incinerators, and medical waste incinerators. At this time, emissions from hazardous waste incinerators are not addressed in AP-42. Emission factors presented in AP-42 are designed for estimating emissions from a large number of sources over a wide area. They are averages of values determined at one or more individual facilities. The individual values which are used to develop the average may vary considerably. The use of AP-42 emission factors to estimate emissions from any one facility should be done with great care.

E. In the absence of suitable AP-42 emission factors, clearly documented engineering evaluations may be used. Documentation should include copies of emission test reports used to derive the emission estimates, any assumptions made and the rationale for the conclusions that were made.

This section describes emission factors for homologue groups of dioxin-like compounds from various stack emission sources. These emission factors are described in units of m g homologue emitted/kg feed material combusted. These are presented for comparative purposes only, and should not be interpreted as representative of the sources described. Most of the profiles are based on very limited data generated under limited emission controls.

When only homologue emission factors are available, then rough estimates of congener specific factors can be made. First, an equal probability of occurrence of the specific congener is assumed based on relative proportionality. For example, 2,3,7,8-TCDD is one congener out of 22 possible congeners in the TCDD homologue. Therefore, the probability of occurrence is assumed to be the ratio of 1/22 or 0.045. Multiplication of a total TCDD emission factor by 0.045 gives an approximation of the emission factor for 2,3,7,8-TCDD. Table 3-1 lists the number of dioxin-like congeners within a homologue group and the total number of congeners within that homologue group.

The use of this procedure may be the only way to evaluate source or site-specific emissions of dioxin-like polychlorinated biphenyls (PCBs), known as the coplanar PCBs. No congener specific test data of coplanar PCBs from incinerators or combustion sources could be found for this assessment. The data that are available usually have reported PCBs as the sum total of all PCBs present in the sample without further speciation of toxic congeners or congener groups (EPA, 1987b). The greatest level of detail in any test report is a further breakdown of total PCBs into homologue groups, e.g., mono - deca-chlorobiphenyl.

Figure 3-1 displays the homologue profiles for 11 specific source categories. Following now are brief summaries of the reference materials for these homologue profiles. Again, it is emphasized that these homologue profiles are not offered as source generalities; they are mostly generated on a small number of different facilities not including a range of emission controls. It is not known whether tested facilities represent the average, or are higher or lower than typical facilities in each source category.

1. Municipal Solid Waste Incineration (MSWI): Municipal incinerators can be classified into four general design categories: mass burn, modular, refuse-derive fuel (RDF), and fluidized-bed combustors (EPA, 1991). Figure 3-1 depicts the homologue profile of TCDD/Fs from MSWIs. It was constructed by merging emissions data from ten modern mass burn technologies (EPA, 1983; EPA, 1988a; Knisley, et al., 1986; Seelinger, et al., 1986; EPA, 1988b; Marklund, et al., 1985; Siebert, et al., 1991; Entropy, 1987).

Table 3-1. The number of dioxin-like and total congeners within dioxin, furan, and coplanar PCB homologue groups.

n: number of N: number of

Homologue Group dioxin-like congeners total congeners 1/N

1. Dioxins

Tetra-CDD 1 22 0.022

Penta-CDD 1 14 0.071

Hexa-CDD 3 10 0.100

Hepta-CDD 1 2 0.500

Octa-CDD 1 1 1.000

2. Furans

Tetra-CDF 1 38 0.026

Penta-CDF 2 28 0.036

Hexa-CDF 4 16 0.063

Hepta-CDF 2 4 0.250

Octa-CDF 1 1 1.000

3. Coplanar PCBs

Tetra-Chloro PCBs 2 42 0.024

Penta-Chloro PCBs 4 46 0.022

Hexa-Chloro PCBs 4 42 0.024

Hepta-Chloro PCBs 1 24 0.042

2. Hazardous Waste Incineration: Hazardous waste incinerators have not been extensively evaluated for stack emissions of dioxin-like compounds. Only a few reports appear in the published literature from which homologue emission factors can be developed (NATO, 1988; EPA, 1992), and these do not give complete inventories of emissions. Therefore, homologue emission factors were estimated based on a series of emission tests at a rotary kiln waste incinerator in Biebesheim, German (EPA, 1992). Figure 3-1 depicts the homologue profile for this single hazardous waste incinerator.

3. Drum and Barrel Reclamation Furnace: Dioxin-like compounds were measured by EPA in the stack gas emissions of a drum and barrel reclamation furnace as part of the