Volume II Chapter 3.0 Pages 6 of 10 page next page 7

3.5.6. Summary of Theories of CDD/CDF Emissions 3-94


3.6.1. Municipal Solid Waste Incineration 3-97

3.6.2. Hazardous Waste Incineration 3-109

3.6.3. Medical Waste Incineration 3-112

3.5.6. Summary of Theories of CDD/CDF Emissions

The above section discussed the likelihood that anthropogenic sources explain the bulk of CDD/CDFs currently in the environment. Still, the considerable research on the complex chemistries of combustion that transpire to ultimately yield CDDs/CDFs remains largely theoretical. The three primary theories being advanced are:

Theory 1:
CDDs/CDFs present as contaminants in the combusted organic materials or that are thermally treated by a combustion process explain the emissions of CDDs/CDFs out of the stack. It is proposed that some quantity of this initial contamination survives the thermal stress imposed by the heat of the incineration or combustion process and is subsequently emitted from the stack.

Theory 2:
CDDs/CDFs are ultimately formed from the thermal breakdown and molecular rearrangement of precursor compounds. Precursor compounds are chlorinated aromatic hydrocarbons having a structural resemblance to the CDD/CDF molecule. Among the precursors that have been identified are polychlorinated biphenyls (PCBs), chlorinated phenols (CPs), and chlorinated benzenes (CBs).

The formation of CDDs/CDFs is believed to occur after the precursor has condensed and adsorbed onto binding sites on the surface of fly ash particles. The active sites on the surface of fly ash particles somehow promote the chemical reactions forming CDDs/CDFs as products of this reaction, which has been observed to be catalyzed by the presence of inorganic chlorides sorbed to the particulate. Heat in a range of 250-450° C has been identified as a necessary condition for these reactions to occur, with either lower or higher temperatures inhibiting the process.

Therefore, the precursor theory focuses on the region of the combustor that is downstream and away from the high temperature zone of the furnace or combustion chamber. This is a location where the gases and smoke derived from combustion of the organic materials have cooled down because of heat losses during conduction through flue ducts; passing through heat exchanger and boiler tubes to recover the heat from combustion for the co-generation of energy; after passing through some air pollution control equipment, and while convected up the stack to be discharged to the atmosphere.

Theory 3:
CDDs/CDFs are synthesized de novo in the same region of the combustion process as described in Theory 2 (i.e., the so-called cool zone). De novo refers to the formation of CDDs/CDFs from organic and inorganic substrates comprised of singular or mixtures of molecules bearing little resemblance to the molecular structure of CDDs and CDFs.

In broad terms, these are nonprecursors and include such diverse substances as petroleum products, chlorinated plastics (PVC), nonchlorinated plastics (polystyrene), cellulose, lignin, coke, coal, particulate carbon, and hydrogen chloride gas. Formation of CDDs/CDFs requires the presence of a chlorine donor, a molecule that partakes a chlorine atom to the predioxin molecule, and the formation and chlorination of a chemical intermediate that is a precursor. The production of a chemical intermediate that is a dioxin precursor does not neatly differentiate Theory 2 from Theory 3, and indeed introduces some confusion into the explanation of the primary chemical events.

The primary distinction is that Theory 3 begins with the combustion of diverse substances that are not defined as precursors, then the nonprecursor reacts on the carbonaceous particulate to eventually yield a chlorinated aromatic hydrocarbon that is a precursor to finally yield CDDs/CDFs. By this distinction, Theory 3 may be viewed as an augmentation to Theory 2 as explained by the initial steps in the synthesis of CDDs/CDFs.

This review has shown that all of these theories are possible and, therefore, taken as a whole, should not be seen as being mutually exclusive. One or more of these theories may act in combination during the combustion of carbonaceous matter. Theory 2 and Theory 3 require the presence of chlorine in the combustible material and in the gaseous state in the combustion plasma.


A summary of the major combustion sources that produce CDDs and CDFs is presented in the following sections. The development of combustor emission estimates has been coordinated with a similar effort ongoing in EPA's Office of Air Quality Planning and Standards (OAQPS). The OAQPS effort is part of a larger EPA effort to inventory air emissions of various toxic substances.

To date, OAQPS has completed emission inventories for about 20 chemicals and has completed a draft document entitled "Locating and Estimating Air Emissions from Sources of Dioxins and Furans" (U.S. EPA, 1993a). This draft OAQPS document summarizes dioxin emissions data for a variety of combustor types. OAQPS is preparing a second report (draft not yet complete) entitled "Emissions Inventory of Section 112 (c)(6) Pollutants: 2,3,7,8-TCDD, 2,3,7,8-TCDF and 2,3,7,8-TCDD Toxic Equivalents."

This report will combine the emission factor information in the "Locating and Estimating" report with production values to develop actual emission rate estimates. Since the OAQPS efforts will not be completed until after this draft report is issued, it is possible that the second OAQPS report may reach somewhat different conclusions. However, the Agency is striving to coordinate these efforts and make the outcomes as consistent as possible. Readers interested in further details about the OAQPS efforts are encouraged to contact OAQPS directly at their location in Research Triangle Park, NC.

3.6.1. Municipal Solid Waste Incineration

Characterization of the Industry

Municipal Solid Waste Incinerators (MSWI) operating in the United States can be classified into four general design categories: mass burn, modular, refuse-derived fuel, and fluidized-bed (U.S. EPA, 1992h). The first type is called mass burn because the waste is combusted without any preprocessing other than removal of items too large to go through the feed system. In a typical mass burn combustor, refuse is placed on a grate that moves through the combustor.

These facilities typically range in combustion capacity from 90 to 2,700 metric tons of MSW per day. Subcategories of mass burn technologies include refractory-walled, rotary kiln, and water-wall facilities. Refractory- walled represent an older class of MSWIs generally built in the late 1970's to early 1980's that were designed to reduce the volume of waste in need of disposal by 70 to 90 percent. These facilities generally lacked boilers to recover the heat of combustion for energy purposes. In the refractory design, the MSW is delivered to the combustion chamber by a traveling grate. Combustion air in excess of stoichiometric amounts is supplied both below and above the grate.

Mass burn water-wall facilities represent substantial design improvements over the refractory-walled incinerators. The water-wall refers to a series of steel tubes running vertically along the walls of the furnace. The tubes contain water, which when heated by combustion, acts as a boiler and transfers energy to produce steam. The steam is then used either to drive an electrical turbine generator or for other industrial needs.

Because a secondary purpose is to generate energy to sell to a customer, significant improvements over refractory-walled MSWIs in terms of increased combustion efficiency have been fostered. The third subcategory of mass burn MSWIs is the rotary kiln. The rotary kiln lacks a traveling or reciprocating grate system to deliver MSW into the furnace.

Rather it employs a water-cooled rotary combustor that is essentially a rotating combustion barrel mounted at a slight angle of decline into which the refuse is pushed by a hydraulic ram (Donnelly, 1992). Preheated combustion air is delivered to the kiln through various portals. The slow rotation of the kiln (i.e., 10 to 20 rotations/hr) causes the MSW to tumble thereby exposing more surface area for complete burn-out. These systems are also equipped with boilers for energy recovery.

As with the mass burn type, modular incinerators also burn waste without preprocessing. Modular MSWIs consist of two combustion chambers (e.g., a primary and secondary chamber mounted in a vertical array). Modular combustors generally range in combustion capacity from 4.5 to 270 metric tons/day.

One of the most common types of modular systems is the starved air (or controlled air system). In these systems, air is supplied to the primary chamber at sub-stoichiometric levels. The incomplete combustion products entrained in the combustion gases from the primary combustion chamber pass into the secondary combustion chamber where excess air is added and combustion is completed by elevated temperatures sustained by auxiliary fuel.

The third major type of MSWI technology is designed to combust refuse-derived fuel (RDF). RDF is a general term describing MSW from which relatively noncombustible items have been removed thereby enhancing the combustibility of the MSW. RDF is commonly prepared by shredding, sorting, and separating metals to create a dense MSW fuel in a pelletized form having a uniform size. RDF fuel is typically burned in a spreader stoker-type combustion chamber (Donnelly, 1992). In the United States, RDF facilities range in total combustion capacity from 227 to 2,720 metric tons/day. These MSWIs are typically steam production facilities that generate salable energy.

The fourth type of MSWI is the fluidized-bed design. In this design, the waste burns in a turbulent bed of noncombustible material, usually sand. The MSW may be fed into the incinerator either as unprocessed waste or as a form of RDF. There are two basic design concepts to the technology: (1) a bubbling-bed incineration unit and (2) a circulating-bed incineration unit. Fluidized-bed MSWIs typically have capacities ranging from 184 to 920 metric tons/day. These systems are usually equipped with boilers to produce steam.

Currently, there are about 170 to 190 MSWI facilities located in 37 states in the United States. (Berenyi and Gould, 1993; Burton and Kiser, 1993). This range in number of facilities reflects the fact that, for any given point in time, the exact population of operating facilities is unknown. However, the best estimate is that 171 MSWI facilities are in operation (Berenyi, 1993; Berenyi and Gould, 1993). About one-half of the operating MSWIs were built since 1988 (Berenyi and Gould, 1993).

The states with the greatest number of facilities are: New York (16), Florida (14), Minnesota (14), Massachusetts (8), Virginia (8), and Connecticut (7) (Berenyi and Gould, 1993). In the most recent reporting year, 1991, EPA estimated that approximately 29.35 million metric tons of MSW were combusted by all operating MSWIs; this represents approximately 17 percent of the annual generation of MSW in the United States (U.S. EPA, 1992c).

Gould (1991) estimated the average annual utilization capacity of typical MSWI designs. Utilization capacity is defined as the percentage of days a facility operates during the course of the year (U.S. EPA, 1992h). Gould (1991) estimated that existing mass burn, modular, and RDF MSWIs had average annual utilization capacities of 87.5, 84.2, and 83.3, respectively.

An estimated 85 percent of existing MSWIs are equipped with one or more air pollution control devices (APCD) to remove some class of pollutants prior to release from the stack (e.g., particulate matter, heavy metals, acid gases, and/or organic constituents) (U.S. EPA, 1992h). These APCDs include electrostatic precipitators (ESPs), fabric filters (FFs), dry sorbent injection (DSI), spray dryer adsorption (SDA), and wet scrubbers (WS).

The ESP is generally used to collect and control particulate matter derived from combustion. This is accomplished by introducing a strong electrical field in the flue gas stream, which, in turn, imparts a charge to the particles entrained in the combustion gases (Donnelly, 1992). Large collection plates are given an opposite charge to attract and collect the particles.

Fabric filters are also particulate matter control devices. Six- to eight-inch diameter bags made from woven fiberglass material are arranged in series. The combustion gases are forced through the tightly woven fabric. The porosity of the fabric is such that the bags act as a filter medium and retain small particles comprising the particulate matter.

Dry sorbent injection is designed for the control of MSWI acid gases. DSI involves the injection of hydrated lime or soda ash into the gas stream to react with and neutralize the acid gas constituents (Donnelly, 1992). Spray dryer adsorption involves both acid gas and particulate matter control. In a typical SDA system, hot combustion gases enter a reactor where atomized hydrated lime slurry is introduced at a controlled velocity (Donnelly, 1992).

The flue gas temperature is significantly decreased, and the acid gas constituents quickly react with the reagent. The reaction evaporates the moisture to produce a dried product that is removed from the bottom of the spray dryer. In general, SDAs are used in combination with either ESPs or FFs. Greater than 95 percent reduction and control of CDDs/CDFs in MSWI emissions has routinely been achieved with FF/SDA systems (U.S. EPA, 1992h). Wet scrubber devices (WS) are designed for acid gas removal, and are more common to MSWIs in Europe than in the United States.

Wet scrubber devices consist of two-stage scrubbers whereby the first stage removes HCl and the second stage removes SO2 (Donnelly, 1992). Water is used to remove the HCl, and either caustic or hydrated lime is added to remove SO2 from the combustion gases. Table 3-27 summarizes the current estimated distribution of operating MSWIs by design category and installed APCDs.

Estimation of MSWI Dioxin Emissions Using an Emission Factor Approach

The approach used here to estimate emissions is based on an emission factor. Emission factors are estimates of the mass of CDD/CDF emitted from the stack per kg of waste combusted. As shown in Table 3-28, these factors were estimated for each design category, multiplied by the amount of waste burned within the design category and then summed to get the total emissions.The first step in this process is to collect emission test data representative of each design category. EPA's Office of Air Quality Planning and Standards (OAQPS) has already collected such a data set (U.S. EPA, 1993a).

table Table 3-27 Estimated Number of Operating MSWI Facilities in the United States by Design Category and Type of Air Pollution Control Device
... This summary presents emission testing for dioxin-like compounds for 30 existing MSWI facilities.

These tests have been reviewed by OAQPS and determined to have used appropriate stack testing and laboratory protocols and to have been conducted under normal operating conditions.

These 30 facilities represent a mix of MSWI designs and technologies as well as air pollution control devices (APCDs) in actual use, providing a basis for extrapolating to all U.S. facilities.

For design categories where more than one test was available, the concentrations for each congener were averaged across tests.
expand table Table V2 3-27
Table 3-28 Estimated MSW Incineration Emission Factors (EF) and Annual Emissions of Total CDD/CDFs
table page 1 of 2 table page 2 of 2
expand Table V2 3-28 page 1-2 expand Table V2 3-28 page 2-2
The emissions data (U.S. EPA, 1993a) are presented in concentration units of nanogram of CDD/CDF per dry standard cubic meter of combustion gas (ng/dscm) corrected to 7% oxygen. Emission factors were computed for each MSWI design category by multiplying the average CDD/CDF concentration by the volume of combustion gas that is produced per kg of waste incinerated. The gas production factor was derived considering the typical heat content of the refuse as follows (Federal Register, 1987c):

1. Assume the heat content of typical MSW = 4500 B.t.u./lb of MSW.

2. Assume that 2.57E-7 dscm are produced per joule value of the MSW.

3. One joule = 9.47E-04 B.t.u.

4. One pound = 0.4536 kg


dscm/kg of MSW = (4500 B.t.u./lb) x (1 joule/9.47E-04 B.t.u) x (lb/.4536 kg) x (2.57E-07 dscm/joule)

dscm/kg of MSW = 2.69

As indicated above an emission factor was estimated for each design class and multiplied by the amount of waste burned to get the emission rate. The emission rate estimates shown in Table 3-28 reflect all congeners of CDD and CDF. These values can be converted to TEQs by applying a ratio of total CDD/CDF to TEQ. EPA has reviewed the congener-specific emissions profiles of twelve MSWI technologies and has determined that, although variable, the average ratio appears to be about 60:1 (i.e., the total mass of CDD/CDF is roughly 60 times greater than the computed TEQ) (Radian, 1994).

As a measure of variability, the standard deviation from this analysis was +/- 20 from the mean ratio (i.e., a ratio of 40:1 to 80:1). As noted in Table 3-28, the total CDD/CDF mass emission from all operating 171 MSWIs is 1.8E+05 grams. The TEQ mass emission is estimated to be 3,000 grams TEQ/yr (assuming the average ratio of 60:1 for the TEQ conversion from the total CDDs/CDFs).

Discussion of Uncertainties

The procedure used to estimate national emissions of dioxin from the MSWI industry involves uncertainties that could cause the estimate to be lower or higher than the true value. The emission estimates were derived on the basis of emissions testing from 30 facilities. As discussed below much of the uncertainty revolves around the representativeness of these facilities.

. How well do the 30 facilities represent the whole population of 171 facilities in terms of technologies? The 30 facilities were selected to be representative of the range of MSWI designs and air pollution control systems. As indicated in Table 3-28, only one of the 13 design classes was not represented. Facilities of particular concern are those that use ESPs which operate in a temperature range of 200° - 400° C.

As discussed in Section 3.5 these conditions can promote the formation of CDDs/CDFs. Over the past few years, some of the facilities with "hot-sided ESPs" have made changes in operating conditions or equipment to address this problem. Although, the 30 tested facilities do include some with "hot-sided ESPs" it is not clear if they are representative of current conditions at all such facilities.

. How well do the 30 facilities represent the whole population of 171 facilities in terms of timing? The emissions were largely derived from stack tests conducted during the period 1988 to 1991. Since 1991, new facilities may have become operational or changes may have been made to existing ones. Therefore emissions today may be somewhat lower, reflecting continued improvements in combustor design.

. For individual facilities, how representative are emission tests of long term performance? The average emissions from a single facility are typically derived from 3 - 4 days of testing over the year. It is not known to what extent such short-term testing may truly reflect long-term emissions, e.g., through the life of the facility. Most stack testing data were collected while the MSWI was operating according to design specifications, e.g., under normal operating conditions.

Using these data would not reflect any additional emissions that may occur during upsets in the combustion zone, poor operations, equipment malfunctions, or degradation in the effectiveness of the pollution control systems.

. How accurate is the approach used to convert stack concentrations to emission factors? As discussed earlier in this section, the approach used to convert concentration in the combustion gas to an emission factor is based on an assumption that 2.69 dscm of combustion gas are produced per kg of MSW burned. This quantity is variable among facilities and is dependent on such factors as the temperature of combustion, the amount of air supplied to combustion chamber in excess of stoichiometric requirements, the moisture of the feed material being burned, and the heat value of the feed material being combusted.

For some technologies with relatively high amounts of excess air delivered to the combustion chamber, the gas volume may be as high as 5.0 - 6.0 dscm/kg.

. How accurate is the procedure used to convert total CDD/CDF emissions to Toxic Equivalents (TEQs)? The conversion ratio was based on a review of emissions from 12 MSWIs. In actuality, the ratio of total CDD/CDF to TEQ is variable from one facility to another. It is influenced by the composition of the MSW and the operating conditions of the combustor. It is not known how representative the generic ratio of 60:1 is of dioxin emissions from all existing MSWIs.

Although MSWIs have the strongest emission data base of all combustion sources evaluated in this document, it still must be considered uncertain for the reasons stated above. Therefore, the estimated emission factors are given a "medium" confidence rating. The amount of MSW that is annually combusted by various MSWI technologies (see Table 3-28) is given a "high" confidence rating.

These estimates are based on a recently conducted and comprehensive survey (U.S. EPA, 1992h). Based on these confidence ratings, the estimated range of potential annual emissions is assumed to vary by a factor of 5 between the low and high ends of the range. Assuming that the best estimate of annual emissions (3,000 g TEQ/yr) is the geometric mean of this range, then the range is calculated to be 1,300 to 6,700 g TEQ/yr.

EPA Regulatory Activities

EPA will soon propose revised emission standards for all existing and new MSWIs with unit capacities greater than 30 metric tons per day. Once these standards have been promulgated, and the States have fully enforced the emission limits, then EPA expects to reduce the national emissions of dioxin emissions from all existing MSWIs by about 99 %.

All existing facilities combined should then be emitting about 30 g TEQ/yr. Full implementation and enforcement of the rules should be achieved by the year 2000. As the compliance date approaches and facilities are upgraded, EPA expects that emissions from these facilities will decline significantly from current levels.

Estimated CDD/CDFs in MSWI Ash

An estimated 7 million metric tons of total ash (bottom ash plus fly ash) are generated annually by MSWIs (telephone conversation between J. Loundsberry, U.S. EPA Office of Solid Waste, and L. Brown, Versar Inc., on February 24, 1993). U.S. EPA (1991b) indicates that 2.8 to 5.5 million tons of total ash are produced from MSWIs with fly ash comprising 5 to 15 percent of the total. U.S. EPA (1990c) recently reported the results of analyses of MSWI ash samples for CDDs and CDFs.

Ashes from five state-of-the-art facilities located in different regions of the United States were analyzed for all 2,3,7,8-substituted CDDs and CDFs. The TEQ levels in the ash (fly ash mixed with bottom ash) ranged from 106 ng/kg to 466 ng/kg with a mean value of 258 ng/kg. CDD/CDF levels in fly ash are generally much higher than in bottom ash.

For example, Fiedler and Hutzinger (1992) report levels of 13,000 ng TEQ/kg in fly ash. Multiplying the mean TEQ total ash concentration by the estimated volume of MSWI ash generated annually (7 million metric tons) yields an estimated annual TEQ in MSWI ash of 1,800 g TEQ/yr.

The total ash generation estimate is given a "medium" confidence rating since it is based on an expert opinion and is about twice as high as earlier published estimates. The emission factor is given a "medium" confidence rating because it is based on direct measurements at five facilities, although these five facilities may not be representative of all technologies in the United States.

Based on these confidence ratings, the estimated range of potential annual emissions is assumed to vary by a factor of 5 between the low and high ends of the range. Assuming that the best estimate of annual emissions (1,800 g TEQ/yr) is the geometric mean of this range, then the range is calculated to be 810 to 4,000 g TEQ/yr.

Each of the five facilities sampled in U.S. EPA (1990c) had companion ash disposal facilities equipped with leachate collection systems or some means of collecting leachate samples. Leachate samples were collected and analyzed for each of these systems. Detectable levels were only found in the leachate at one facility (TEQ = 3 ng/l); the only detectable congeners were HpCDDs, OCDD, and HpCDF.

3.6.2. Hazardous Waste Incineration

EPA estimates that there are 190 Hazardous Waste Incinerators (HWIs) in the United States. This total includes both operating facilities and facilities that are not operating but have filed an application with EPA (Helble, 1993). The four principal technologies employed for the combustion of hazardous waste in the United States are:

  • liquid injection
  • rotary kiln
  • fixed hearth
  • and fluidized-bed incinerators (Dempsey and Oppelt, 1993).

Liquid injection incinerators are designed to burn pumpable liquid hazardous waste. These incinerators are typically simple refractory-lined cylinders (either horizontally or vertically aligned) equipped with one or more waste burners. The liquid waste is injected into the combustion chamber through an atomizer, and the liquid droplets are exposed to high temperatures in suspension.

Rotary kiln incinerators are the more common design. They have the added versatility of being able to combust hazardous waste in any physical form (i.e., liquid, semi-solid, or solid). The rotary kiln is a horizontal cylinder lined with refractory material. Rotation of the cylinder on a slight slope provides for transport of the waste through the kiln, as well as enhanced mixing and exposure to the heat of combustion. The combustion gases emanating from the kiln are usually passed through a high temperature afterburner chamber to more completely destroy organic pollutants arising from combustion.

Fixed hearths, the third principal hazardous waste incineration technology, are starved air or pyrolytic incinerators. These are two-stage combustion units. Waste is ram-fed into the primary chamber and incinerated at about 50 to 80 percent of stoichiometric air requirements. The resulting smoke and pyrolytic combustion products are then passed though a secondary combustion chamber where relatively high temperatures are maintained by the combustion of auxiliary fuel. Oxygen is introduced into the secondary chamber to promote complete thermal oxidation of the organic molecules entrained in the gases. T

he fourth hazardous waste incineration technology is the fluidized-bed incinerator. This technology is similar in design to that employed in MSW incineration. (See Section 3.6.1).

Dempsey and Oppelt (1993) summarized the results of EPA-sponsored stack testing at six full-scale HWIs, three PCB incinerators, and one incinerator burning PCP waste. CDD/CDFs were detected at all three PCB incinerators with TEQ emission rates ranging from 0.3 to 1.63 ng TEQ/dscm (@ 7 percent oxygen). CDD/CDFs were detected at three of the HWIs with TEQ emission rates ranging from 0.57 to 17.7 ng TEQ/dscm (@ 7 percent oxygen).

Helble (1993) reviewed recent data from trial burn reports on CDD/CDF emissions from 15 HWIs. CDD/CDFs were detected in the stack emissions of 11 of the 15 facilities at total CDD/CDF emission rates ranging from 0.1 to 1,600 ng/dscm (@ 7 percent oxygen) with most facilities between 1 and 100 ng/dscm. Based on his evaluation of the emissions data, Helble (1993) concluded that the CDD/CDFs observed in emissions from HWIs are formed catalytically under low temperature conditions either through catalytic chlorination or through catalytic condensation of dioxin-like precursors such as chlorobenzenes and PCBs.

Emission factors are estimated based on the results of the emission tests reported by Helble (1993). Homologue-specific emissions data, waste feed rates, and stack flow rates (dscm @ 7 percent oxygen) were available for six of the HWIs evaluated by Helble (1993). From these data, total CDD/CDF emission factors were calculated for each facility (range: 10 to 6,830 ng/kg of waste feed; mean: 1,550 ng/kg of waste feed).

For those facilities with more than one test run reported, the total CDD/CDF emission rates for the individual runs were averaged to obtain a facility average emission rate. These total CDD/CDF emission factors were converted to TEQ emission factors using a conversion factor of 1.75 ng TEQ/ng of total CDD/CDF that was developed by EER, Inc. for EPA's Office of Solid Waste (EER, 1993). The resulting mean TEQ emission factor is 27.2 ng TEQ/kg waste feed (range = 0.18 to 119 ng TEQ/kg waste feed).

Dempsey and Oppelt (1993) estimate that between 216 and 249 million metric tons of hazardous waste were generated in 1987 (the year for which the most comprehensive data on waste management are available). Of this total amount, Dempsey and Oppelt (1993) estimate that between 1.0 and 1.3 million metric tons of hazardous waste were incinerated. Based on an estimated 1.3 million metric tons of hazardous waste incinerated per year in the United States and the mean emission factor derived above, it is estimated that 2,000 grams CDD/CDF per year and 35 grams TEQ/yr are emitted from HWIs.

A "low" confidence rating is ascribed to the emission factors derived above because stack test data were available for only 6 of the 190 HWIs in the United States and the stack test data used represent only one hazardous waste technology (rotary kiln).

The "production" estimate has been assigned a "medium" confidence rating because it is based on a thorough review of the various studies and surveys which have been conducted in recent years to assess the quantity and types of hazardous waste generated in the United States, as well as the quantities and types of wastes being managed by various treatment, storage and disposal facilities.

Based on these confidence ratings, the estimated range of potential annual emissions is assumed to vary by a factor of 10 between the low and high ends of the range. Assuming that the best estimate of annual TEQ emissions (35 g TEQ/yr) is the geometric mean of this range, then the range is calculated to be 11 to 110 g TEQ/yr.

3.6.3. Medical Waste Incineration

Buonicore (1992b) has reviewed the primary incineration technologies used to burn medical and pathological wastes in the United States. These medical waste incinerations (MWI) fall within three broad technology categories: retort, controlled-air, and rotary kiln. Retort incinerators are multiple chamber combustors characteristic of the "older" existing technology. The medical waste is fed into a primary combustion chamber, and the gases from combustion are passed into a secondary chamber.

In the secondary chamber, secondary auxiliary fuel is burned to sustain higher temperatures and to more completely burn the organic pollutants entrained in the combustion gas from primary combustion. Combustion air, 100 to 300 percent in excess of stoichiometric requirements, is usually added to the secondary chamber. Gases exiting the secondary chamber are directed to an incinerator stack. The second principal technology is the controlled-air incinerator.

This is the most common technology used to incinerate hospital/medical waste and is often referred to as modular incineration. Like retort incinerators, combustion occurs in two stages. Medical waste is fed into a primary combustion chamber where air is delivered at less than stoichiometric requirements. Under these conditions, the waste is pyrolyzed and volatile compounds are released. A secondary chamber is located on top of the primary unit.

Auxiliary fuel is added to sustain high temperatures in a controlled-air environment. These systems are usually automated with computer-directed controllers that are integrated with a thermocouple. Thus, the quality of combustion is superior to the retort technology. The third type of MWI is the rotary kiln. This is the same technology as employed in both municipal and hazardous waste incineration. (See Sections 3.6.1 and 3.6.2).

EPA has estimated that about 4.3 million metric tons (4.76 million short tons) of hospital/medical wastes are generated annually in the United States (U.S. EPA, 1991d). Table 3-29 summarizes the types and number of facilities that generate medical waste, and their corresponding annual generation rate of medical wastes. There are about 6,700 MWIs operating nationwide combusting approximately 3.72 million metric tons of medical waste annually (U.S. EPA, 1991d). Table 3-30 summarizes the estimated population of MWI currently operating in the United States.

CDDs and CDFs have been identified in the stack gas emissions of MWIs located at hospitals in the United States (U.S. EPA, 1993a). Although operating on a smaller-scale, the mechanism of CCD/CDF formation in hospital waste incineration is similar to that described for MSWI in Section 3.6.1. To support future rulemaking, EPA has developed a summary of annual emissions of dioxin from all existing hospital waste incinerators operating in the United States (U.S. EPA, 1991e).

This summary represents an analysis and review of dioxin emissions measured at the stack from six MWIs (Radian, 1991a; 1991b; 1991c; McCormack, 1990; Lew et al., 1988; Lew et al., 1989). From these reports, EPA derived an average emission factor of total amount of dioxin released to the air per kg of medical waste combusted in a typical MWI (g total CDD/CDF per kg waste), based largely on uncontrolled emissions (U.S. EPA, 1993a).

table Table 3-29 Estimated Number and Type of Facilities and Quantities of Medical Waste Generated Annually in the United States table Table 3-30 Medical Waste Incineration Facilities Operating in the United States
expand Table V2 3-29 expand Table V2 3-30

The average uncontrolled emission factor of total CDD/CDF is 8.53E-05 g/kg (U.S. EPA, 1993a). This factor can be compared with the average controlled emission factor of total CDD/CDF of 4.46E-06 g/kg from one facility equipped with acid gas controls and a fabric filter (U.S. EPA, 1993a). The ratio of controlled to uncontrolled emissions is a factor of 1:20. Table 3-31 summarizes the emission factors developed for this analysis.

In computing an estimate of national emissions of dioxin from 6,700 existing MWIs, EPA applied an average emission factor developed for uncontrolled MWIs. Uncontrolled emissions are defined as emissions from a MWI facility not equipped with add-on air pollution control devices (APCD) (e.g., electrostatic precipitator, scrubber, fabric filter, etc.).

However, MWIs are modular designs consisting of both a primary and secondary combustion chamber. The purpose of the secondary combustion chamber is the continued destruction of organic compounds emanating from the primary chamber. Therefore MWIs are not actually totally uncontrolled. EPA expects that the majority of the existing MWIs are uncontrolled with respect to dioxin control measures (U.S. EPA, 1991e). EPA believes that the selection of an average emission factor derived from uncontrolled emissions
currently represents the most accurate means of estimating the magnitude of potential dioxin release from all 6,700 operating MWIs (U.S. EPA, 1991e).

In order to estimate national emission of total dioxin to the air from all operating facilities, EPA categorized the population of MWIs according to the operating duty, the size of the combustor, and the amount of medical or pathological waste combusted per year. Table 3-32 summarizes the estimate of total dioxin emitted (g/yr) from all operating MWIs in the United States according to this disaggregation.

Emissions to air of total CDD/CDF (i.e., tetra-chlorinated through octa-chlorinated compounds) from approximately 6,700 existing medical waste incinerators are estimated to be 3.18E+05 grams/yr. This emission estimate was derived from tests conducted at six facilities (considered to be representative of the major design types), extrapolating average emissions nationwide using the amount of waste burned in each of the design classes.
table Table 3-31 CDD/CDF Emission Factors for Controlled-Air Medical Waste Incinerators Operating in the United States table Table 3-32 Estimated Annual Emission of Total CDD/CDFs (g/yr) from Incineration of Medical Waste
expand Table V2 3-31 expand Table V2 3-32

For purposes of deriving an estimate of emissions in terms of TEQ, it is necessary to convert the total CDD/CDF emission into an estimated TEQ emission. This is done by assuming a ratio of TEQ to total CDD/CDF using existing data on emissions from existing facilities. The State of California Air Resources Board (CARB) has stack tested a number of hospital waste incinerators in southern California (CARB, 1990a).

Congener-specific emissions of CDD/CDFs were measured in the stack gas emissions of seven facilities. From these data, the ratio of TEQ to total CDD/CDF is 0.016 as an overall average of five tested facilities. Multiplication of the annual emissions of CDD/CDF (in grams per year) by this ratio yields an estimate of 5,100 g TEQ emitted (grams per year) for all existing MWIs in the United States.

U.S. EPA (1993a) reports emissions testing at a number of controlled-air medical waste incinerators with a variety of emission controls. These tests yielded a lower range of emission factors. Based on these data, it appears possible that the national releases from medical waste incinerators could be much lower than the "average" value identified above. It is difficult to say how much lower, since it is unknown how representative these tested facilities are of all 6,700 facilities in the United States.

A "medium" confidence rating is assigned to the estimate of amount of hospital waste burned since it is based on a detailed study specific to the United States; however, the large number of these facilities makes it difficult to estimate precisely. The emission factor used to extrapolate to a national basis is given a "low" confidence rating, because the average was derived from the stack sampling at a small sample of the large numbers of MWI facilities (6 of 6,700).

Based on these confidence ratings, the estimated range of potential annual emissions is assumed to vary by a factor of 10 between the low and high ends of the range. Assuming that the best estimate of annual emissions (5,100 g TEQ/yr) is the geometric mean of this range, then the range is calculated to be 1,600 to 16,000 g TEQ/yr.