United States . Solid Waste and EPA530-D-98-002
Environmental Protection Emergency Response August 1998
Agency (5305W) www.epa.gov/osw
Guidance on Collection of
Emissions Data to Support
Site-Specific Risk
Assessments at
Hazardous Waste
Combustion Facilities
PEER REVIEW DRAFT
-------
-------
EPA530-D-98-002
August 1998
www.epa.gov/osw
GUIDANCE ON COLLECTION OF EMISSIONS
DATA TO SUPPORT SITE-SPECIFIC RISK
ASSESSMENTS AT HAZARDOUS WASTE
COMBUSTION FACILITIES
Risk Burn Guidance
PEER REVIEW DRAFT
OFFICE OF SOLID WASTE AND EMERENCY RESPONSE
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
REGION 4
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATLANTA, GEORGIA 30303
-------
-------
DISCLAIMER
This document provides guidance to EPA and states on how best to implement RCRA and EPA's
regulations to facilitate permitting decisions for hazardous waste combustion facilities. It also provides
guidance to the public and the regulated community on how EPA intends to exercise its discretion in
implementing its regulations. The document does not substitute for EPA's regulations, nor is it a
regulation itself. Thus, it cannot impose legally binding requirements on EPA, states, or the regulated
community, and may not apply to a particular situation based upon the circumstances. EPA may change
this guidance in the future, as appropriate.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
-------
ACKNOWLEDGMENTS
This document was developed by the U.S. Environmental Protection Agency (EPA) Region 4 and the
Office of Solid Waste. The technical approach was provided by Beth Antley, EPA Region 4, and
TetraTech EM Inc. under Contract No. 68-W4-007. Andrew O'Palko, Office of Solid Waste, provided
overall direction. Special thanks are extended to the following for their valuable comments and
contributions:
Office of Research and Development,
National Risk Management Research Laboratory
Dr. Paul Lemieux
Dr. Brian Gullett
Dr. William Linak
Jeff Ryan
Office of Research and Development,
National Exposure Research Laboratory
Dr. Larry Johnson
Office of Air Quality Planning and Standards,
Emissions Measurement Center
Gene Riley
Office of Solid Waste,
Permits and State Programs Division
Rosemary Workman
Val de la Fuente
Office of Solid Waste,
Economics, Methods and Risk Analysis Division
Karen Pollard
Barry Lesnik
Larry Rosengrant
Office of General Counsel
Karen Kraus
Georgia Department of Natural Resources
Michele Burgess
North Carolina Department of Environment,
Health and Natural Resources
Katherine O'Neal
EPA Region 2
John Brogard
EPA Region 3
Gary Gross
EPA Region 4
Rick Gillam
Elizabeth Bartlett
Denise Housley
Hugh Hazen
David Langston
Analytical Services Branch
EPA Region 5
Gary Victorine
EPA Region 6
David Weeks
JeffYurk
Ruben Casso
Stan Burger
EPA Region 7
John Smith
EPA Region 8
Carl Daly
EPA Region 10
Catherine Massimino
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draff
IV
-------
CONTENTS
Section
ACRONYMLIST viii
1.0 INTRODUCTION 1
1.1 OBJECTIVES 2
1.2 REVIEW OF RCRA TRIAL BURN PROVISIONS 3
1.3 RISK ASSESSMENTS AT HWC FACILITIES 6
1.4 EPA'S MACT STANDARDS 7
2.0 EXPANDED SCOPE OF THE TRIAL BURN 9
2.1 DIOXINS AND FURANS : 9
2.2 ORGANICS OTHER THAN DIOXINS AND FURANS 10
2.3 METALS 10
2.3.1 Human Health and Ecological Concerns for Specific Metals 11
2.3.2 Issues Related to Indirect Risks from Mercury 14
2.3.3 Issues Related to Indirect Risks from Lead 14
2.4 PARTICLE-SIZE DISTRIBUTION 15
2.5 HYDROGEN CHLORIDE AND CHLORINE 15
3.0 INTEGRATION OF PERFORMANCE TESTING WITH RISK-BASED DATA NEEDS ... 17
3.1 PERFORMANCE TESTING AND DEFINING AN OPERATING ENVELOPE 17
3.2 TEST CONDITIONS FOR RISK-BASED DATA COLLECTION AND RISK BURN 18
3.3 DEFINING AVERAGE OR NORMAL OPERATING CONDITIONS 19
3.4 RISK-BASED DATA COLLECTION, EXAMPLE LOGIC 20
4.0 DIOXIN AND FURAN EMISSIONS 23
4.1 DIOXIN AND FURAN FORMATION MECHANISMS 23
4.1.1 Particulate Hold-Up Temperatures 26
4.1.2 Rapid and Partial Liquid Quench Systems 28
4.1.3 Combustion Conditions 28
4.1.3.1 Transient Conditions 31
4.1.3.2 Containerized or Batch Wastes 32
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities v
-------
CONTENTS (Continued)
Section Page
4.1.3.3 High Carbon Monoxide 33
4.1.4 Feed Composition 33
4.1.4.1 Chlorine 34
4.1.4.2 Metal Catalysts 34
4.1.4.3 D/F Precursors 35
4.1.4.4 D/F Inhibitors 35
4.1.4.5 Other Factors 35
4.1.5 D/F Control Technologies 36
4.2 OPERATING PARAMETERS ASSOCIATED WITH D/F PRODUCTION 36
4.3 D/F EMISSIONS FROM HWIS 37
4.4 D/F EMISSIONS FROM BOILERS 38
4.5 D/F EMISSIONS FROM CEMENT KILNS 44
4.6 D/F EMISSIONS FROM LWAKS 50
5.0 ORGANIC EMISSIONS OTHER THAN DIOXINS AND FURANS 51
5.1 ORGANIC EMISSIONS FROM HWIS AND BOILERS 52
5.2 ORGANIC EMISSIONS FROM CEMENT KILNS AND LWAKS 55
6.0 METAL EMISSIONS 57
6.1 METAL SPECIATION 59
6.2 METAL VOLATILITY GROUPINGS 60'
6.2.1 Volatile Metals (Mercury and Selenium) 64
6.2.2 Semivolatile Metals 65
6.2.3 Low-Volatile Metals 66
6.3 OPERATING CONDITIONS AND PARAMETERS FOR METALS 67
7.0 PARTICLE-SIZE DISTRIBUTION AND HYDROGEN
CHLORIDE AND CHLORINE EMISSIONS 69
7.1 RELATIONSHIP BETWEEN PARTICLE SIZE AND POTENTIAL EMISSIONS .. 69
7.2 MEASURING PARTICLE-SIZE DISTRIBUTION 70
7.3 OPERATING CONDITIONS AND PARAMETERS FOR COLLECTION OF
PARTICLE-SIZE DISTRIBUTION DATA 71
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities vi
-------
CONTENTS (Continued)
Section Page
7.4 HYDROGEN CHLORIDE AND CHLORINE 71
8.0 DATA ANALYSIS AND PERMIT CONDITIONS 72
8.1 PERMIT LIMITS FOR KEY OPERATING PARAMETERS 73
8.2 APPLICATION OF TRIAL BURN DATA IN THE SSRA 75
9.0 REPORTING CHANGES OF WASTE FEED AND OPERATING PARAMETERS 78
REFERENCES 79
Appendix
A TRIAL BURN CONDITIONS AND PERMIT LIMITS FOR AN EXAMPLE HWC FACILITY
B SAMPLING AND ANALYSIS
TABLES
4-1 OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM HWIS AND
BOILERS 39
4-2 OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM CEMENT
KILNS AND LWAKS 45
FIGURES
3-1 SITE SPECIFIC RISK ASSESSMENT DATA COLLECTION FLOW CHART 21
6-1 METALS VOLATILITY 63
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
VII
-------
ACRONYM LIST
Hg/dL Micrograms per deciliter
}lg/m3 Micrograms per cubic meter
APCD Air pollution control device
ASME American Society of Mechanical Engineers
ASTM American Society for Testing and Materials
ATSDR Agency for Toxic Substances and Disease Registry
AWFCS Automatic waste feed cutoff system
BIF Boiler and industrial furnace
Btu British thermal unit
CEMS Continuous emissions monitoring system
CFR Code of Federal Regulations
C12 Molecular chlorine
COPC Compound of potential concern
CSF Carcinogenic slope factor
D/F Dioxins and furans
DNA Deoxyribonucleic acid
DQO Data quality objective
DRE Destruction and removal efficiency
EDL Estimated detection limit
EER Energy and Environmental Research Corporation
ELCR Excess lifetime carcinogenic risk
EMPC Estimated maximum possible concentration
EPA U.S. Environmental Protection Agency
ESP Electrostatic precipitator
GC Gas chromatography
grains/dscf Grains per dry standard cubic foot
HC1 Hydrogen chloride
HEAST Health Effects Assessment Summary Tables
HEPA High efficiency particulate air
HRA Hourly rolling average
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
viii
-------
ACRONYM LIST (Continued)
HWC Hazardous waste combustion
HWI Hazardous waste incinerator
IRIS Integrated Risk Information System
LWAK Light weight aggregate kiln
MACT Maximum achievable control technology
mg/dscm Milligrams per dry standard cubic meter
mg/kg Milligrams per kilogram
MS Mass spectrometry
ng/dscm Nanograms per dry standard cubic meter
NAAQS National Ambient Air Quality Standard
PAH Polycyclic aromatic hydrocarbon
PCB Polychlorinated biphenyl
PCC Primary combustion chamber
PCDD Polychlorinated dibenzo-p-dioxin
PCDF Polychlorinated dibenzofuran
PIC Products of incomplete combustion
PM Particulate matter
POHC Principal organic hazardous constituent
ppm Parts per million
ppmv Parts-per-million volume
RAC Reference air concentration
RCRA Resource Conservation and Recovery Act
RDL Reliable detection limit
RfC Reference concentration
RfD Reference dose
RME Reasonable maximum exposure
SCC Secondary combustion chamber
SRE System removal efficiency
SSRA Site-specific risk assessment
TCDD Tetrachlorodibenzo-p-dioxin
TEQ Toxic equivalents
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
ix
-------
ACRONYM LIST (Continued)
TO Total organic
TRY Toxicity reference value
U/BK Uptake/biokinetic
WESP Wet electrostatic precipitator
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
x
-------
1.0 INTRODUCTION
This document provides guidance on collection of stack emissions data to support multi-pathway, site-
specific risk assessments (SSRA) at hazardous waste combustion (HWC) facilities. HWC facilities include
hazardous waste incinerators (HWI) and boilers and industrial furnaces (BIF) that burn hazardous waste
for energy or material recovery. Multi-pathway human health and ecological SSRAs may be performed for
these facilities to determine potential direct and indirect risks from:
Polychlorinated dibenzo-p-dioxins (PCDD) and polychlorinated dibenzofurans (PCDF),
known collectively as dioxins and furans (D/F)
Organic emissions (volatile, semivolatile, and nonvolatile) other than D/Fs, often referred
to as products of incomplete combustion (PIC)
Metals that are potentially toxic to human or ecological receptors
Hydrogen chloride (HC1) and molecular chlorine (C12)
Direct risks include risks from the inhalation exposure pathway. Indirect risks include risks resulting from
long-term deposition, including contamination of soil and surface water, bioaccumulation, and food chain
effects. The results of an SSRA may be used to establish additional conditions in the Resource
Conservation and Recovery Act (RCRA) permit as necessary to protect human health and the environment.
These additional conditions can include emissions limits for D/Fs, metals, or other site-specific
compounds; additional limits on operating parameters; and expanded waste characterization and waste
tracking requirements." _
This guidance identifies the types of emissions data that are needed to support SSRAs. The guidance
suggests operating and feed conditions that should generally be demonstrated during the testing, and
identifies appropriate stack sampling and analytical techniques for collection of emissions data. In
addition, the relationship between test conditions and potential RCRA permit conditions is discussed. The
guidance presumes that emissions data for SSRAs will typically be collected in conjunction with RCRA
trial burns. However, in some cases emissions data may be generated during other sampling events.
This guidance does not include a methodology for performing the SSRAs. The U.S. Environmental
Protection Agency (EPA) is currently developing protocols for conducting human health and ecological
SSRAs (EPA in press a, in press b). These or other SSRA protocols (e.g., Research Triangle Institute
[RTI] 1996) should be used in conjunction with this guidance. The human health and ecological risk
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
1
-------
protocols, in conjunction with this guidance, supersede previous EPA Office of Solid Waste guidance on
combustion risk assessments (EPA 1994b).
1.1 OBJECTIVES
The objectives of this guidance are to:
Ensure the collection of adequate data to support completion of defensible human health
and ecological SSRAs. Specific data needs of the SSRA include (1) D/F emission rates,
(2) emission rates for organics other than D/Fs, (3) metals emission rates, (4) site-specific
particle-size distribution data, and (5) emission rates for HC1 and C12. These data needs
are discussed in detail in Section 2.0, and sampling and analytical procedures are discussed
in Appendix B.
Provide flexibility in the trial burn process by allowing (as appropriate) the option for a
separate test conducted exclusively for collection of SSRA data. The separate test (which
can be referred to as a risk burn) may be performed at conditions representing long-term
average, or normal, operations. Selection of appropriate test conditions for collection of
SSRA data is discussed in detail in Section 3.0
Identify key operating parameters that can influence emissions of D/Fs, organics other than
D/Fs, and metals. These operating parameters should be demonstrated during SSRA
emissions testing and may be limited by the final permit. This guidance also recognizes
inherent design differences between HWIs, boilers, cement kilns, and light weight
aggregate kilns (LWAK), and accounts for these differences in recommending operating
parameters to be limited under RCRA. Key operating parameters related to emissions of
D/Fs, organics other than D/Fs, and metals are discussed in Sections 4.0, 5.0, and 6.0,
respectively. The significance of particle-size distribution and HC1 and C12 emissions are
addressed in Section 7.0.
Explain how emissions data may be evaluated in the SSRA, and identify additional
emissions and operating parameters that might be limited in the final RCRA permit based
upon results of the SSRA. These issues are discussed in Section 8.0 and are illustrated by
the example in Appendix A.
Identify ongoing waste feed and operation evaluations that may be needed to ensure that
the SSRA remains representative of existing and future facility operations. These issues
are briefly discussed in Section 9.0.
This guidance focuses on the data needs specified in the first bullet of this section (page 2), above. Other
data needs of the SSRA should be addressed on a site-specific basis, and could include compounds such as
radionuclides or other criteria pollutants not specifically addressed in this guidance. Additional discussion
of some of these other compounds can be found in EPA (in press a).
This guidance supplements, but does not supersede, existing guidance and regulatory requirements
pertaining to performance testing. Although general guidelines are provided, permit writers are
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 2
-------
encouraged to consider facility-specific circumstances that are not fully addressed. Overall, this guidance
should be regarded as a living document and may be updated as additional technical issues arise.
EPA Region 6 has also developed additional guidance for reviewing trial burn plans, RCRA Part B permit
applications, and documents such as quality assurance project plans (EPA 1997f, 1997g, 1997h, 1997i, and
1997J). These documents may also be helpful when reviewing permit applications, trial burn plans, and
related documents. These documents are available from EPA Region 6 on their Internet web site at
http://www.epa.gov/earthlr6/6pd/rcra-c/manual/manual.htm#A.
The remainder of this section provides background information on RCRA performance standards and
existing trial bum guidance. Potential limitations of current EPA regulations are identified, and the use of
SSRAs to supplement current regulatory controls is also discussed. Finally, this section briefly discusses
EPA's proposed Maximum Achievable Control Technology (MACT) standards for HWC facilities (EPA
1996a) and discusses their relevance to this guidance.
1.2
REVIEW OF RCRA TRIAL BURN PROVISIONS
This section reviews RCRA trial burn provisions and it provides the framework for issues discussed later
in the document. The review discussion is minimal, and the cited references should be consulted for more
detail.
Performance standards for HWIs are described in Title 40 of the Code of Federal Regulations (CFR) Part
264, Subpart O. These standards were promulgated on January 23, 1981, and have been subsequently
amended. The performance standards consist of the following: (1) a destruction and removal efficiency
(DRE) of principal organic hazardous constituents (POHC) of 99.99 percent, or 99.9999 percent for
dioxin-listed wastes; (2) particulate matter (PM) emissions not to exceed 180 milligrams per dry standard
cubic meter (mg/dscm) or 0.08 grains per dry standard cubic foot (grains/dscf), corrected to 7 percent
oxygen; and (3) gaseous HC1 emissions not to exceed 1.8 kilograms per hour or a removal efficiency of 99
percent. Compliance with these performance standards is generally established through a carefully
designed trial burn. The trial burn is used to define the operating range (or envelope) that assures the
required DRE of the selected POHCs. Trial burns for a FfWI typically are conducted at extreme "worst-
case" conditions for time, temperature, and turbulence in the combustion chamber. This involves at least
one performance test condition at a minimum combustion temperature to demonstrate DRE. Additional
test conditions may be required to resolve potential conflicts between operating parameters.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
3
------- An error occurred while trying to OCR this image.
-------
Organic emissions from BIFs are controlled using a DRE standard and by limiting carbon monoxide (and
in some cases total hydrocarbon concentrations) in stack gas (EPA 1992b). The BIF rule also recognizes
that HWC facilities with dry APCDs operating between 450 and 750 ฐF may emit higher concentrations of
D/Fs than HWC facilities with other types of APCDs. SSRAs are required for BIFs with dry APCDs
operating at 450 to 750 ฐF to demonstrate that the ELCR from D/Fs is less than 1E-5 for the inhalation
exposure pathway.
The regulation of metals under the BIF rule led to the need for an additional performance test condition,
conducted at a high combustion temperature, in order to maximize metal volatility. Thus, under the BIF
rule, the operational envelope is generally defined by a low-temperature test to demonstrate DRE and a
high-temperature test to demonstrate system removal efficiency (SRE) for up to 12 metals. Operating
limits to control metals are based on SRE conditions that include maximum combustion temperature,
maximum metals feed rates, and appropriate APCD operating parameters (e.g., temperature, pressure, or
power). Maximum metal and chlorine feed rates for BIFs are established using one of the following
approaches: Tier I (assuming zero SRE without site-specific air dispersion modeling), Adjusted Tier I
(assuming zero SRE with site-specific air dispersion modeling), Tier n (based on emissions data without
site-specific air dispersion modeling), or Tier IE (based on emissions data with site-specific air dispersion
modeling). Owners and operators of BIFs may choose to spike the wastes used during the SRE test with
metals and chlorine to establish maximum feed rates. This ensures that sufficiently flexible metal and
chlorine feed rate limits will be incorporated into the facility's operating permit. Similar provisions have
been implemented at HWIs as necessary to protect human health and the environment.
EPA trial burn and technical guidance documents for HWIs and BIFs (EPA 1989, 1992b, 1991 f) describe
control parameters typically monitored during trial burn testing. These operating parameters (monitored
during DRE or SRE conditions) are defined as Group A, B, or C. Short-term limits, such as hourly rolling
averages (HRA) and instantaneous limits, are typically set for these operating parameters in the facility's
RCRA permit. The Group A, B, and C designations are also used in this document, and are discussed
below.
Group A operating parameters are linked with automatic waste feed cutoff system (AWFCS) limits. This
means that waste feed is automatically cut off when specified limits are exceeded. Group A operating
parameters must be continuously monitored while hazardous waste is being fed to the HWC unit.
Examples of Group A parameters include maximum and minimum PCC and SCC temperatures; maximum
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
5
-------
combustion gas velocity, waste feed rate, carbon monoxide concentration, and combustion chamber
pressure; and minimum venturi scrubber differential pressure, scrubber liquid to gas ratio and pH,
baghouse differential pressure, and wet ESP (WESP) power input and liquid flow rate.
Group B operating parameters generally do not require continuous monitoring and are not interlocked with
the AWFCS limits. However, detailed operating records are required to demonstrate compliance with
permitted operating conditions. Examples of Group B parameters include maximum batch size for
containerized waste and minimum scrubber blowdown. Some Group B operating parameters, including
metal and chlorine feed rates, may be continuously monitored once the supporting analytical data have
been entered into the data system.
Group C operating parameters are set independently of trial burn conditions, and limits are generally based
on manufacturers' recommendations and good operating practices. Examples of Group C parameters
include maximum burner turndown and minimum scrubber nozzle pressure. Most Group C parameters do
not require continuous monitoring and interlocks with AWFCS; however, some may (for example,
minimum atomization fluid pressure to a liquid injection burner). Similar to Group B parameters, detailed
operating records for Group C parameters are required to demonstrate compliance with permitted operating
conditions.
The Group A, B, and C designations are used throughout this document. -These designations are applied to
the operating parameters discussed in Sections 4.0 through 7.0, Tables 4-1 and 4-2, and Appendix A.
1.3 RISK ASSESSMENTS AT HWC FACILITIES
The performance standards described in Section 1.2 were developed to control potential inhalation risks
from HWC units. Requirements for HWIs were promulgated during 1981, and requirements for BIFs were
promulgated during 1991. Since these requirements were developed, additional concerns regarding
combustion of hazardous waste have been identified. These concerns relate to potential indirect effects
(including uptake through the food chain) from long-term deposition of metals, D/Fs, and other organic
compounds in soils and surface water. The current regulations do not directly address potentially
significant risks via these indirect exposure pathways.
The U.S. Congress and EPA have recognized that emissions of compounds such as D/Fs and mercury are
especially toxic and have the propensity to cause significant human and ecological effects through
bioaccumulation in the food chain. This includes contamination of beef and dairy cattle, as well as pork
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 6
-------
and poultry, on agricultural lands and fish in aquatic ecosystems. Potential indirect risks from these types
of compounds often exceed potential risks from the direct inhalation pathway. Additional information on
these contaminants can be found in the Mercury Study Report to Congress (EPA 1997e), EPA's proposed
MACT standards (EPA 1996a), and EPA's protocols for human health and ecological risk assessments
(EPA in press a, in press b).
Based on available data summarizing notable D/F and mercury emissions from HWC facilities (EPA
1996a, EPA 1997k), EPA has prioritized the identification of potential human health and ecological risks
due to HWC emissions (EPA in press a, in press b), and implementation of appropriate controls (EPA
1996a). In addition, EPA recognizes that significant uncertainty may exist regarding potential risks due to
organic emissions other than D/Fs. SSRAs provide a mechanism for assessing potential risks and
uncertainties, and for determining protective emission rates. Therefore, EPA recommends that SSRAs be
completed as part of the RCRA permitting process for HWC facilities as necessary to protect human health
and the environment (EPA 1993, 1994c, 1996a).
Section 3005(c)(3) of RCRA and 40 CFR Part 270.32(b)(2) provide EPA with the responsibility to
establish additional permit conditions on a case-by-case basis as necessary to protect human health and the
environment. These provisions are generally referred to as EPA's omnibus authority. Under 40 CFR Part
270.10(k), EPA may use its omnibus authority to require a permit applicant to submit additional
information necessary to establish protective permit conditions. Collection of additional emissions data
and completion of SSRAs can provide the information necessary to determine what, if any, additional
permit conditions are necessary for maintaining risks within acceptable levels. Any decision to add permit
conditions based on an SSRA must be justified in the administrative record for each facility, and the
implementing agency must explain the basis for the additional conditions.
1.4
EPA'S MACT STANDARDS
EPA'S MACT standards for HWC facilities propose technology-based limits for D/Fs; organic emissions
other than D/Fs; low-volatile, semivolatile, and volatile metals; C12 and HC1; and PM (EPA 1996a). These
standards have been proposed under joint authority of RCRA and the Clean Air Act. The MACT standards
will apply to several categories of HWC facilities including HWIs, hazardous waste burning cement kilns,
and hazardous waste burning LWAKs. Emission standards for hazardous waste burning boilers are
anticipated in a future MACT standards rulemaking.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
7
-------
EPA (1996a) recommends that SSRAs continue to be part of the RCRA permitting process, prior to
implementation of final MACT standard requirements. Updated policy regarding the use of SSRAs in
setting permit conditions is expected at the time the MACT standards are promulgated. If, in the future,
MACT standards are promulgated that are more stringent than the existing site-specific risk-based
standards, then the MACT standards would take precedence.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
8
-------
2.0 EXPANDED SCOPE OF THE TRIAL BURN
Specific data needs of the SSRA that should be considered during trial burn planning include (1) D/F
emission rates, (2) organic emission rates other than those for D/Fs, (3) metals emission rates, (4) particle-
size distribution, and (5) emission rates of HC1 and C12. This section summarizes the significance of these
data needs with respect to protection of human health and the environment.
Results of the SSRA for human health will be used to determine whether potential risks exceed acceptable
levels as specified by the regulatory authority. It is important that all toxicological data concerning a
constituent be evaluated for all known effects. Some constituents have effects on multiple organs, and thus
have a collective harmful effect on the receptor.
Prior to data collection, information should be developed for the trial burn plan and associated quality
assurance project plan to ensure that analytical methods will be adequately sensitive so that risks are not
driven by non-detected constituents. Analytical detection limits should be compared with the data needs of
the SSRAs (EPA in press a, in press b), and as appropriate, be compared with those found in EPA (1996g).
Detection limits may need to be reduced to achieve target risk levels. Detection limits and data analysis
are discussed in greater detail in EPA (in press a), and in Section 8.2 and Appendix B of this document.
2.1
DIOXINS AND FURANS
D/Fs should be addressed in SSRAs because they can pose significant risks through both direct and
indirect exposure pathways. Their propensity to partition to adipose (fat) tissue and to bioaccumulate can
make food chain effects particularly significant. According to EPA (1995b) the air-to-plant-to-animal
exposure pathway is a primary exposure route for humans. Other significant exposure pathways include
source-to-surface water-to-fish. Human exposures would then result from ingestion of contaminated milk,
beef, fish, and other foods. D/Fs are believed to promote cancer and other harmful health effects in
humans and other receptors, and the potential carcinogenicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD) is currently being re-evaluated by EPA.
D/Fs are a group of anthropogenic chemical compounds created as unintended by-products during
combustion and industrial activities (EPA 1995b). As described in Section 4.0, the formation of D/Fs in
HWC facilities is highly dependent on temperature, time, and the presence of a reaction surface. In
general, D/Fs addressed in this document contain at least four chlorine atoms and can assume a planar
configuration that allows for specific biological effects. D/Fs are generally quantified in terms of 2,3,7,8-
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
9
-------
TCDD toxic equivalents (TEQ). For the purpose of this guidance, TEQs are calculated using the
procedure in 40 CFR Part 266, Appendix IX.
2.2
ORGANICS OTHER THAN DIOXINS AND FURANS
A number of organic compounds other than D/Fs can result in increased risks from both direct and indirect
exposures and should therefore be evaluated in SSRAs. The treatment of hazardous wastes in HWC
facilities results in emissions of organics that include halogenated and nonhalogenated organic compounds.
Chapter 2.0 of EPA (in press a) states that HWC organic emissions can include classes of compounds such
as polycyclic aromatic hydrocarbons (PAH) and polychlorinated biphenyls (PCB). Both PAHs and PCBs
are considered by EPA (1998a) as potential human carcinogens, and are considered to be toxic, persistent,
and bioaccumulative. PAHs are frequently associated with PM from combustion facilities. EPA (in press
a) cites data from EPA Region 10, whereby PCBs were formed as PICs during combustion of hazardous
wastes. As discussed in Section 5.0 of this document, organic PICs from HWC facilities typically include
compounds such as methane, propane, chlorinated alkanes and alkenes, phenols, and chlorinated aromatics
(Ryan and others 1996,1997; Midwest Research Institute and AT Kearney, Inc. 1997).
Target analyte lists for specific organics to be identified as part of SSRA data collection are provided in
Appendix B. However, the total identification and quantification of all organics is frequently not possible,
which adds to the uncertainty of the risk assessment process. Therefore, Section 5.0 and Appendix B of
this document describe a method for total organic (TO) analysis for quantifying the total mass of organic
PICs based on boiling point ranges. This information is useful in indicating the level of uncertainty
associated with the risk assessment.
2.3
METALS
Metal emissions can pose potential human and ecological risks and should be evaluated in SSRAs. EPA's
BIF rule (40 CFR 266, Subpart H) and supporting documentation (EPA 1992a) lists the metals antimony,
arsenic, barium, beryllium, cadmium, chromium, lead, mercury, nickel, selenium, silver, and thallium for
establishing metal feed rates. Table A-l, "Chemicals for Consideration as Compounds of Potential
Concern," of Appendix A of EPA (in press a, in press b) identifies the additional metals aluminum, copper,
cobalt, manganese, vanadium, and zinc. Thus, these additional metals would also be considered as
compounds of potential concern (COPC) during the COPC selection process. As appropriate, these metals
may be carried through the risk assessment process.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
10
-------
Potential human cancer risks (based on the inhalation exposure pathway) are based on carcinogenic slope
factors (CSF) for arsenic, beryllium, cadmium, hexavalent chromium, and certain species of nickel. Health
effects other than cancer are based on RfDs and/or RfCs for antimony, barium, trivalent chromium, copper,
manganese, mercury, selenium, silver, thallium, and vanadium. As explained in Section 2.3.3 of this
document, potential health effects from lead are modeled using an alternative approach (EPA in press a).
EPA's ecological risk assessment protocols (EPA in press b) include ecological toxicity reference values
(TRY) for a number of ecosystems and receptors. This includes freshwater quality, freshwater sediment,
marine and estuarine water quality, marine and estuarine sediment, and specific receptors including
earthworms, terrestrial plants, mammals, and birds. TRVs have been developed for aluminum, antimony,
arsenic, barium, beryllium, cadmium, copper, hexavalent chromium, total chromium, lead, mercury,
methyl-mercury, nickel, selenium, silver, thallium, and zinc. These metals should be considered during the
COPC selection process for ecological SSRAs.
2.3.1
Human Health and Ecological Concerns for Specific Metals
The fpllowing paragraphs discuss human health and ecological significance of the metals aluminum,
cobalt, copper, manganese, nickel, selenium, vanadium, and zinc. This information is provided because
these metals were not addressed in the original BIF rule (40 CFR 266, Subpart H) or supporting
documentation (EPA 1992b). All of these metals (except aluminum and cobalt) have published toxicity
values (RfDs and/or RfCs) in EPA's Integrated Risk Information System (IRIS) (EPA 1998a) and/or EPA's
Health Effects Assessment Summary Tables (HEAST).
Aluminum, zinc, and copper are commonly occurring metals that bioaccumulate through the food chain in
their ionic form (Paasivirta 1991). These metals can accumulate in soils and/or aquatic ecosystems.
Selenium has been demonstrated to cause significant ecological impacts to wildlife, primarily because of
the effects from irrigation return water. Nickel has caused both human heath as well as ecological effects,
and it is commonly associated with industrial operations such as plating.
Aluminum
Aluminum bioconcentrates in aquatic species and vegetation and can cause toxic effects (Agency for Toxic
Substances and Disease Registry [ATSDR] 1992; Paasivirta 1991). In mammals, aluminum is poorly
absorbed and is not readily metabolized (Ganrot 1986). From a human health perspective, aluminum
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
11
-------
emissions do not appear to pose any toxicological concern. Therefore, aluminum is considered as a COPC
only for ecological SSRAs and TRVs have been developed for aluminum.
Cobalt
Cobalt is a relatively rare metal that is produced primarily as a by-product during refining of other metals,
primarily copper. Industrial exposures to airborne cobalt appears to result in sensitization followed by
irritation and produced pulmonary effects (Doull and others 1991). TRVs for cobalt have not been
developed.
Copper
Copper is an essential trace element required by numerous oxidative enzymes and is found in all mammals
and several other classes of organisms; however, high doses can result in adverse effects to humans and
ecological receptors (Doull and others 1991). In general, humans are less sensitive to copper toxicity than
other mammals (such as sheep and cattle), and healthy adults usually are not affected by exposure to low,
chronic doses of copper. Infants and children are susceptible to chronic copper toxicity. Once absorbed,
copper is distributed to the liver and can bind to deoxyribonucleic acid (DNA) or generate free radicals.
Organisms lacking effective biological barriers to absorption are the most susceptible to copper toxicity.
Aquatic organisms such as algae, fungi, certain invertebrates, and fish represent examples of this type of
organism. Thus, aquatic ecosystems would be the most sensitive to copper loading; however, copper
released to surface waters generally precipitates out or adsorbs to sediments (ATSDR 1990). TRVs have
been developed for copper.
Manganese
Manganese is an essential trace element, is present in all living organisms, and is a cofactor in a number of
enzymatic reactions. Industrial exposures to large doses of airborne manganese dioxide (or other forms of
manganese) can result in manganese pneumonitis, which can result in permanent lung damage. Chronic
exposure to airborne manganese can result in central nervous system effects including psychiatric disorders
and a Parkinson-like syndrome. Liver cirrhosis has also been observed. It is significant that these effects
have only been observed in higher primates such as humans and monkeys and not in experimental animals
such as rats (Doull and others 1991). IRIS (EPA 1998a) bases the chronic reference dose for inhalation on
data collected during occupational exposures to manganese dioxide; therefore, any potential risks from
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
12
-
-------
manganese will be driven by the direct inhalation exposure route. TRVs for manganese have not been
developed.
Nickel
Nickel is a commonly used industrial metal, and it is frequently associated with iron and copper ores.
Significant loading of nickel to the atmosphere and surface waters has been observed over the past several
decades (Doull and others 1991). EPA (in press a, 1998a) states that nickel refinery dust is a potential
human carcinogen. Based on this information, EPA (in press a) recommends that inhalation risks from
nickel be evaluated using the CSF for nickel refinery dust. This represents a change in EPA's previous
policies under the BIF rule whereby nickel was evaluated as a noncarcinogen. This may result in lower
allowable nickel feed rates. Additional information on nickel speciation is found in Section 2.3.8.4 of EPA
(in press a) and Section 6.1 of this document. TRVs have been developed for nickel.
Selenium
Selenium occurs naturally in soils, is associated with copper refining and several industrial processes, and
has been used in pesticides. It is an essential element, bioaccumulates in certain species of plants, and has
been associated with toxic effects in livestock (blind staggers syndrome). Soils containing high levels of
selenium (seleniferous soils) can lead to high concentrations of selenium in certain plants, and pose a
hazard to livestock and other species. In humans, selenium partitions to the kidneys and liver and is
excreted though the urine and feces. Toxic effects in humans include central nervous system and
gastrointestinal effects (Doull and others 1991). Aquatic birds are extremely sensitive to selenium; toxic
effects include teratogenesis. Based on available data, aquatic birds are potential sensitive receptors.
TRVs have been developed for selenium.
Vanadium
Vanadium is found in several ores, is associated with petroleum products, and is commonly found in food
oils. It is unclear whether or not vanadium has any essential role in human metabolism. Toxic effects of
vanadium are usually associated with the inhalation exposure route. Effects of chronic exposure to
airborne vanadium compounds in workers include bronchitis, bronchopneumonia, and kidney damage
(Doull and others 1991). Animal experiments have confirmed effects on the lungs and kidney. TRVs for
vanadium have not been developed.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
13
-------
Zinc
Zinc is an essential trace element that is involved with many enzymatic reactions in humans and other
species (Doull and others 1991). It generally occurs in the divalent oxidation state. Zinc does
bioaccumulate, specifically through the indirect soil-to-vegetable exposure pathway. The most common
observed human health effect from zinc would be depression of enzyme production, resulting in copper
deficiency. Ecological effects include a tendency to bioaccumulate and toxic effects to vegetation and
aquatic organisms (Paasivirta 1991). TRVs have been developed for zinc.
2.3.2
Issues Related to Indirect Risks from Mercury
Control of mercury emission and feed rate limits to protect against potential indirect risks will generally
result in lower limits compared to those established under the BIF rule, 40 CFR 266, Subpart H and EPA
(1992b). Mercury has significant potential to biotransfer up the food chain through water and sediments to
fish to human receptors. Thus, indirect exposure pathways may drive potential risks from mercury to
human and ecological receptors.
Default exposure assumptions for speciated mercury are found in EPA (in press a, in press b). However, a
facility may wish to collect site-specific data to address mercury speciation. Additional information on
mercury speciation is found in Section 2.3.8.3 of EPA (in press a), and in Section 6.1 and Appendix B of
this document.
2.3.3
Issues Related to Indirect Risks from Lead
Under the BIF regulations, emissions and feed rate limits for lead are based on a RAC equivalent to 10
percent of the NAAQS for lead converted to an annual average, or 0.09 micrograms per cubic meter
(jig/m3). However, both air and soil lead levels should be evaluated in a SSRA. Air concentrations can be
modeled, and soil levels can be calculated based on modeled results.
Information on estimating threshold levels of lead exposure is provided in Sections 2.3.8.2 and 7.5 of EPA
(in press a). Toxicity factors (CSFs and RfDs) are not available for lead. Therefore, adverse health effects
for lead will be characterized through a direct comparison with media-specific health-based levels,
adjusted for background exposure.
The neurobiological effects observed in children are used as the sensitive endpoint for evaluating lead
toxicity from ingestion. EPA recommends a maximum lead concentration in blood of 10 micrograms per
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
14
-------
deciliter (ng/dL), which is at the low end of the range of concern for adverse health effects in children.
Potential risks from lead are evaluated based on the application of an uptake/biokinetic (U/BK) modeling
approach that evaluates potential risks by predicting blood lead levels associated with exposure to lead.
When run with standard recommended default values, the U/BK model predicts that no more than five
percent of children exposed to a lead concentration in soil of 400 milligrams per kilogram (mg/kg) will
have blood lead concentrations exceeding 10 jig/dL. In lieu of direct comparison to a target soil level,
EPA (in press a) also provides the option of running the U/BK model.
2.4
PARTICLE-SIZE DISTRIBUTION
Information on particle-size distribution (presented as particle diameter in micrometers, referred to as
microns) is needed for the air dispersion and deposition modeling that supports the SSRAs (EPA in press
a). Because particle dispersion and subsequent deposition are directly related to particle size, potential
risks from the indirect exposure route are directly dependent on particle-size distribution. In general, most
metals and a few organic COPCs with very low volatility are assumed to occur only in the particle phase.
An exception is mercury which partitions between particle and vapor phases.
Default assumptions for particle-size distribution for sources equipped with ESPs and fabric filters are
provided in EPA (in press a). However, recent measurements show that the default assumptions may be
overly biased towards larger particles (EPA 1998b). In addition, default assumptions are not available for
facilities equipped with wet APCDs, or for facilities with no APCDs. Thus, site-specific particle-size
measurements are recommended for all HWC facilities completing SSRAs.
Sampling and analysis for particle-size distribution will reflect site-specific combustion characteristics and
the efficiency of the APCD; significant particle fractions with particle sizes less than one micron are not
uncommon (EPA 1998b). Because a widely accepted method for determining particle-size distribution is
not yet available, permit writers should work with facilities to obtain the best data possible when
considering facility-specific circumstances. This issue is discussed in more detail in Section 7.0 and in
Appendix B of this document.
2.5
HYDROGEN CHLORIDE AND CHLORINE
Potential risks from HC1 and C12 are limited to the inhalation pathway. Thus, completion of SSRAs is not
expected to affect the current regulatory approach for HC1 and C12 from BIFs. However, EPA recommends
that HWIs test their stack emissions for both HC1 and C12. Risk-based limits may be established in the
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
15
-------
final permit for HWIs in lieu of the 40 CFR Part 264 technology-based limits, if the risk-based limits are
more stringent. Detailed information on control of HCI and C12 has been provided in previous guidance
(EPA 1989,1992a, 1996a). Therefore, this document does not address control measures for HCI and C12 in
detail.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
16
-------
3.0 INTEGRATION OF PERFORMANCE TESTING WITH RISK-BASED DATA NEEDS
As discussed in Sections 1.0 and 2.0, collection of SSRA data at HWC facilities involves sampling and
analysis for D/Fs, non-D/F organics, metals, particle-size distribution, and HC1 and C12. Based on the
results of the SSRA, EPA may establish risk-based permit limits under omnibus authority of RCRA as
described in 40 CFR 270.32(b)(2). Therefore, collection of emissions data for use in the SSRAs should be
integrated with performance testing as necessary to produce a consistent set of enforceable permit
conditions.
This section provides general guidelines regarding appropriate test conditions for collection of emissions
data to be used in SSRAs. Data collection may be associated with (1) a low-temperature test conducted to
demonstrate DRE of POHCs, (2) a high-temperature SRE test conducted to establish metal feed rates and
to account for metal emissions and partitioning, and/or (3) a specially designed risk burn test (which may
be conducted at normal operating conditions) specifically for the collection of risk assessment data. A
flow chart (Figure 3-1) is presented to help guide permit writers and facility personnel through the
decision-making process. Sections 4.0 through 6.0 build on these concepts by identifying operating
parameters associated with emissions of specific contaminants (D/Fs, non-D/F organics, and metals) and
specific types of HWC facilities (HWIs, boilers, cement kilns, and LWAKs). These operating parameters
include many of the same Group A, B, and C parameters typically measured during DRE and SRE
performance testing.
3.1
PERFORMANCE TESTING AND DEFINING AN OPERATING ENVELOPE
Section 1.0 explains that performance testing of HWC facilities involves the identification of an
operational envelope based on DRE (low temperature) and SRE (high temperature) conditions. Defining a
facility's operational envelope has been regarded as a way of challenging the facility's operating systems to
perform under "worst-case" conditions while meeting EPA performance standards for DRE (of organic
POHCs) and emission rates (of PM, metals, HC1, C12, carbon monoxide, and total hydrocarbons). In some
cases, wastes synthesized from pure compounds are burned during performance tests in lieu of actual
wastes to help achieve these "worst-case" conditions. Synthetic wastes may also be burned during
performance testing because of analytical interferences between the waste constituents and the POHCs,
because of incompatibilities between the wastes and spiking materials, or because waste volumes are
insufficient for completion of three runs per test condition.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
17
-------
These considerations influence whether collection of SSRA data can be integrated with DRE and SRE
performance testing, or whether data collection during a separate risk burn test is preferred. For example, a
facility may need to consider collection of SSRA emissions data during a separate risk burn test if actual
wastes are not burned during the DRE and SRE tests. As discussed further in Sections 4.0 and 5.0, actual
representative wastes should be burned to the extent possible during collection of SSRA emissions data for
D/F and non-D/F organics. If circumstances preclude use of actual wastes during the DRE or SRE test
conditions, then integration of risk-based data collection with performance testing may not be an option.
Additional information on balancing the use of real wastes versus synthetic wastes is found in Section 5.1
of this document.
3.2 TEST CONDITIONS FOR RISK-BASED DATA COLLECTION AND RISK BURN
Data for the SSRA may be collected during DRE or SRE test conditions if issues regarding the use of
actual representative wastes can be resolved. However, estimating potential risks based on emissions from
worst-case DRE and/or SRE conditions could be overly conservative, given that a facility will not operate
at the edges of its operating envelope all of the time. For example, the SRE test condition may involve
spiking to increase metal feed rates above normal levels, resulting in high metal emissions. From a risk
assessment standpoint, a case can be made for using emissions data generated during average, or normal
operation of the HWC unit. SSRAs assess potential risks of operations over 30 or more years. Therefore,
for contaminants such as metals, emissions during normal operations (as compared to emissions during
DRE or SRE conditions) may be more directly related to the risk posed by the HWC facility over its
operating life.
If a facility requests that emissions data be collected under normal or average operating conditions (for
consideration in the SSRA), then an additional test condition is specified in the trial burn plan. This
additional risk burn is optional and is performed by the facility on a voluntary basis. However, the facility
may feel that a test condition at average operating conditions is more representative of long-term emission
rates. Evaluation of emissions data generated under normal operating conditions in an SSRA (in
conjunction with or instead of data generated during a DRE or SRE test) may be considered by the
permitting authority on a case-by-case basis. If an agreement with the facility can be reached, then an extra
layer of permit conditions may be needed to ensure that conditions represented as normal during the test
are, in fact, normal over the long-term operation of the facility.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 18
-------
3.3
DEFINING AVERAGE OR NORMAL OPERATING CONDITIONS
Collecting emissions data under normal or average operating conditions may result in lower emission rates
and lower, but more representative, potential risks. However, every HWC facility may not be eligible for
using risk data collected under normal operating conditions in the SSRA. Relevant criteria include the
following:
1. Can the facility provide sufficient information to define normal feed and operating
conditions?
EPA (1997f) states that data collection for SSRAs under normal operating conditions will
only be considered if the facility burns wastes that have little temporal variation in
chemical and physical properties, at nearly constant rates, and under operating conditions
that do not fluctuate widely. Facilities that cannot meet these criteria, including facilities
that burn highly variable wastes, should perform the risk testing under DRE and SRE
conditions. Similar logic also should be applied to fuels and raw materials used in cement
kilns and LWAKs. For example, mercury in coal can contribute to potential risks and
should be considered in the feed characterization.
2. Can the facility identify additional permit limitations and record keeping requirements
to ensure that the facility does not operate in excess of the normal conditions over the
long term?
Facilities choosing to use SSRA data collected under normal operating conditions will
likely be subject to additional permit conditions. The purpose of the additional conditions
is to ensure that ongoing feeds and operating conditions remain within the boundaries
represented by the risk assessment (feeds include hazardous waste and, as appropriate, raw
materials and auxiliary fuels such as coal). These permit conditions may be based on long-
term (monthly, quarterly, semiannual, and/or annual) averages for operating parameters, or
they might involve waste tracking and record keeping. The facility should propose a
permitting approach to ensure that normal conditions are maintained. The permit may
require that significant changes be reported to the permitting authority. These changes
may be subject to review and determination of the need for possible retesting.
3. Are there additional site-specific considerations?
A facility may be required to use emissions data collected under DRE and/or SRE
conditions in the SSRA because of other circumstances, such as acute risk concerns.
An example of a situation meeting the criteria for use of emissions data collected under normal operating
conditions would be a BEF facility that can provide historical data to establish average metal feed rates, and
that proposes to comply with both ERA and monthly-average metal feed rate permit limits. The testing
and permitting approach would be as follows:
SRE Test: The SRE test would be conducted at maximum metals feed rates, and could
involve metals spiking. Emissions from the SRE test would need to be below the BIF
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
19
-------
allowable emissions limits (which are based only on inhalation risks). HRA metal feed
rate limits would be established based upon this test.
Normal Test: This test would be conducted at normal metals feed rates, without metals
spiking. Emissions data from the normal test would be evaluated in the SSRA to
determine potential risks from both direct and indirect exposures. Monthly-average metal
feed rate limits would be established based upon this test.
An example of a situation that may not meet the criteria for data collection under normal operating
conditions would be a commercial HWC facility with a diverse customer list that randomly burns any
hazardous waste available on the test day. This facility may not be able to make a case as to why the
specific waste represents normal operations, or may not commit to a waste inventory tracking scheme and
long-term averaging to assess whether the test waste remains representative. In this case, emissions data
from "worst-case" DRE and SRE tests should be evaluated in the SSRA.
Emission rates (and related operating parameters and conditions) that are evaluated in the SSRA should
clearly correspond to the permit terms and conditions. These risk-based terms and conditions should serve
as a constraint to prohibit operations outside of the operating boundaries evaluated in the SSRA. The
permit writer should work closely with the facility to determine the appropriate test condition(s) for
collection of the SSRA data.
3.4
RISK-BASED DATA COLLECTION, EXAMPLE LOGIC
Figure 3-1 provides example logic for permit writers and facility personnel to use in determining
appropriate test conditions for risk-based data collection. The first step involves determining whether data
collected under DRE and SRE conditions are anticipated to meet target risk levels, as specified by the
regulatory agency. This question can be answered by evaluating existing emissions data, by conducting
"miniburns" (as allowed by existing permit or interim status conditions), or by evaluating data from similar
facilities in a preliminary SSRA. Miniburns involve collection of emissions data prior to trial burn testing.
Miniburns can be used for determining potential compliance with risk-based limits, and for establishing
risk-based detection limits.
Based on preliminary data analysis, a facility may anticipate that emissions data collected during
performance testing (DRE and SRE conditions) will meet target risk levels. The permit writer would then
work with the facility to determine whether the operating parameters identified in Sections 4.0 through 6.0
of this document can be demonstrated under DRE and SRE test conditions. If so, then the SSRA
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
20
-------
FIGURE 3-1
SITE SPECIFIC RISK ASSESSMENT DATA COLLECTION FLOW CHART
NO. _.
Are emissions from DRE/SRE
performance tests likely to pass
the site-specific risk assessment?
Yes
Yes
Can normal operating conditions
be defined and maintained?
Implement process changes
to reduce emissions.
i No
i
Yes
Can adjustments be made to define
normal operating conditions?
Yes
^ซI
Adjust operating parameters and
conduct engineering analysis to
achieve normal operating
conditions.
Testing:
Perform additional test conditions at
normal conditions for collection of
D/F, non-D/F organic, and metals
data for evaluation in the site-
specific risk assessment.
Are the appropriate operating
parameters (Section 4.0)
demonstrated during
performance tests?
No
Add test conditions as necessary to
establish limits for appropriate
operating parameters.
Zesting:
Collect D/F, non-D/F organic, and
metals data for evaluation in the site-
specific risk assessment in
conjuction with
performance tests.
Permit Limits:
- Operating Establish long-term average permit limits and/or
Limits
- Emission
- Limits
a compliance monitoring plan for important
operating parameters for:
i D/F
i Non-D/F organics
i Metals
Establish HRA or instantaneous limits from
DRE/SRE tests.
Normal Test
Emission
Rates
Allowable
Permit Limits^
forD/Fsand
Metals
Lower of
Risk-Based
Limit or
Existing
Regulatory
Standard
Permit Limits:
- Operating No long-term average permit limits needed.
Establish HRA or instantaneous limits per
existing guidance from DRE/SRE tests.
Limits
- Emission Perf. Test
Limits Emission
Rates
Allowable
Permit Limits^
for D/Fs and
Metals
Lower of
Risk-Based
Limit or
Existing
Regulatory
Standard
Normal or Average Operating Conditions
Performance Test Conditions
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
21
-------
emissions data may be collected in conjunction with the DRE and SRE test conditions. If not, then extra
test conditions may be needed. For example, a trial burn plan for a boiler equipped with a dry APCD
would not include a maximum temperature SRE test condition if the waste feed did not contain metals.
Section 4.1.1 identifies maximum inlet temperature to a dry APCD as a primary operating parameter
related to D/F emissions. Therefore, the facility would evaluate whether maximum APCD temperature
could be achieved during the DRE test. If the boiler could not establish a maximum APCD inlet
temperature during the DRE test (because of an inability to control APCD temperature independently of
combustion temperature), then an extra test condition would be added to the test plan specifically to
demonstrate maximum APCD temperature while sampling for D/Fs.
In some cases, the preliminary data analysis might suggest that emissions data collected during
performance testing (DRE and SRE conditions) will not meet target risk levels. Facilities that cannot meet
target risk levels for emissions data collected during performance testing at DRE and SRE operating
conditions have two options (1) demonstrate that average conditions can be defined and maintained and
collect data at normal or average operating conditions, or (2) implement process changes to reduce
emissions and meet risk-based requirements under DRE and SRE conditions. Process changes can include
changes to improve combustion efficiency (including burner design and methods of feeding wastes) or the
addition of control systems such as carbon injection.
Facilities that demonstrate the ability to define and maintain normal operating conditions will likely be
subject to additional permit conditions based on long-term averages of appropriate waste feed rates,
temperatures, and other appropriate operating parameters. These facilities may also be required to submit
a compliance monitoring plan that demonstrates how long-term averages will be maintained and tracked.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
22
-------
4.0 DIOXIN AND FURAN EMISSIONS
This section summarizes specific operating and waste feed parameters to be considered for collection of
D/F emissions data to support human health and ecological SSRAs. COPC emission rates are dependent
on several operating parameters, most of which are monitored during DRE and SRE tests. Operating
parameters may also vary between types of HWC facilities (HWIs, boilers, cement kilns, and LWAKs).
Separate subsections are included to further discuss the relevance of the operating parameters as they relate
to each type of HWC facility. The MACT database (EPA 1996a, 1996b, 1996c, 1996d, 1996e, 1997a, and
19971) is used as a reference to describe operations and APCD performance in HWIs, cement kilns, and
LWAKs. MACT data on boilers are limited and are not considered fully representative of the entire boiler
universe within the United States.
This guidance relies on available research and emissions databases to draw general conclusions and
provide recommendations. However, it is important to note that this guidance cannot encompass every
potential situation. Permit writers should always evaluate facility-specific operating trends and
information against the underlying principles of the recommendations in this document.
The subject of D/F formation is both complex and extensive, and this section starts with general
information and becomes progressively more specific. Formation mechanisms are discussed in Section
4.1, key operating and waste feed parameters are reviewed and summarized in Section 4.2, and the
relevance of the parameters for each industry category are discussed in Sections 4.3 through 4.6. Finally,
Tables 4-1 and 4-2 summarize the recommendations by industry category.
4.1
DIOXIN AND FURAN FORMATION MECHANISMS
D/F formation mechanisms, emission rates, and potential control measures in combustion systems have
been studied since the late 1970s with increased efforts in the United States over the past 10 years. D/Fs
are formed as the result of many complex side reactions that occur in a combustion system (Townsend and
others 1995). These side reactions occur primarily in the post-furnace (downstream) regions of the HWC
facility. D/Fs can result from a combination of formation mechanisms depending on combustion
conditions, the type of APCD, and waste feed characteristics.
D/F formation in HWC facilities is believed to include three possible mechanisms. Depending on waste
feed, design, APCD, and operating characteristics, one or more of the following mechanisms may
predominate:
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
23
-------
1. Homogeneous gas-phase formation was one of the earliest D/F formation mechanisms
observed in combustion systems (Sidhu and others 1994). However, gas phase formation
is believed to play a relatively minor role in D/F formation in HWC facilities.
2. The term de novo synthesis is commonly used for heterogeneous, surface-catalyzed D/F
formation from flyash-based organic material coupled with flyash-based metal catalysts
(such as copper). This mechanism is likely to occur in HWC facilities.
3. Heterogeneous D/F formation from gas-phase precursors and flyash-based metal catalysts
is also considered a likely formation mechanism in HWC facilities.
Gas-phase D/F formation from trichlorinated phenols was observed to occur at temperatures of 570 to
1,475 ฐF by Sidhu and others (1994). Their data indicated that the kinetic model developed by Shaub and
Tsang (1983) underestimated potential D/F emissions by a factor of approximately 50. The model
developed by Sidhu and others (1994) is dependent on the presence of halogenated phenols which are
recognized as D/F precursors. Sidhu and others (1994) concluded that pure gas-phase formation of D/Fs in
combustion systems is possible given the presence of halogenated hydrocarbons that form halophenols.
The kinetic model for gas-phase formation developed by Shaub and Tsang (1983) failed to account for all
the D/F emissions from a municipal waste incinerator, and subsequent work focused more on
heterogeneous, surface-catalyzed reactions. Subsequently, over the past 15 years, research has focused on
de novo synthesis of D/Fs and synthesis from gas-phase precursors.
Early studies on municipal waste incinerators indicated that organic compounds in the gas coupled with
high flyash concentrations promote chlorination reactions and subsequent synthesis of D/Fs (Bruce 1993;
Townsend and others 1995). Bruce (1993) and Griffin (1986) theorized that this synthesis involves the
Deacon reaction:
2HC1 + '/a O2 <==> C12 + H2O, with copper or other metals serving as catalysts (Equation 1)
where:
HC1
02
Cl,
H2O
= hydrogen chloride
= oxygen
= chlorine
= water
The free chlorine formed by the reaction then chlorinates D/F precursors, including halogenated aromatics,
through substitution reactions. Sulfur has been shown to interfere with the Deacon reaction, and thereby
decreases D/F formation (Griffin 1986; Bruce 1993; Raghunathan and Gullet 1994). Researchers have
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
24
-------
theorized that sulfur may affect these results by (1) reducing the C12 to HC1 (Equation 2), and (2) altering
the copper in the Deacon reaction (Equation 3) (Bruce 1993):
C12 + S02 + H20<=>2HC1 + S03 (Equation 2)
where:
C12 = chlorine
SO2 = sulfur dioxide
H2O = water
HC1 = hydrogen chloride
SO3 = sulfur trioxide
and:
CuO + SO2+ l/2O2 <=>CuSO4 (Equations)
where:
CuO = cupric oxide
SO2 = sulfur dioxide
H2O = water
CuSO4 = cupric sulfate
De novo synthesis of D/Fs involves many complex reactions that can occur at several stages in the
combustion process. However, all de novo formation mechanisms appear to depend on solid phase
chemistry (Townsend and others 1995). Historically, D/F emissions were believed to be controlled by
ensuring good combustion and by controlling temperature, oxygen, and PM (carbon monoxide
concentration has been used as a surrogate for good combustion). Recent studies indicate that even in
systems achieving good combustion (with low carbon monoxide concentrations), D/F reformation may
occur in cooler zones downstream of combustion chambers (Santoleri 1995). Critical operating parameters
related to D/F formation in downstream zones include (1) presence of particulates, which allow for solid-
phase, metal-catalyzed reactions, (2) appropriate temperature window (approximately 400 to 750 ฐF), (3)
presence of C12 and other precursors, including chlorinated aromatics, and (4) particulate residence time.
Poor combustion can increase D/F formation through increased PM (which serves as the reaction site for
D/F formation), increased formation of PICs (which could serve as D/F precursors), and increased gas-
phase formation of D/Fs.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 25
-------
Santoleri (1995) summarizes several operating conditions and parameters that are relevant to D/F
formation and control as follows:
Combustion temperatures lower than approximately 1,800 ฐF or higher than 2,250 ฐF can
lead to higher free C12 emissions and subsequent D/F formation. A rapid quench is
recommended to quickly lower the temperature and improve the conversion of C12 to HC1.
Sulfur and sulfur dioxide have been observed to be effective in reducing C12 to HC1,
thereby reducing D/F emissions.
Downstream zones that potentially collect PM (including boiler tubes, ESP plates, and
fabric filters) provide reaction sites that promote D/F formation. More rapid cycling of
cleaning processes can shorten the residence time for D/F formation, and decrease D/F
emissions.
Overall, researchers have concluded that D/F formation mechanisms in HWC facilities are extremely
complex and cannot be predicted accurately with kinetic models or surrogate monitoring parameters such
as carbon monoxide or total hydrocarbons (Santoleri 1995). Almost any combination of carbon,
hydrogen, oxygen, and chlorine can yield some D/Fs, given the proper time and temperature (Altwicker
and others 1990; Santoleri 1995). Factors such as non-detect levels of chlorine in feed streams, lack of dry
APCD systems, presence of D/F inhibitors (such as sulfur), lack of D/F catalysts (such as copper), and lack.
of D/F precursors (such as chlorinated phenols) may lead to reduced or low emissions of D/Fs. However,
because mechanisms of D/F formation are extremely complex and are not well understood, it is not
possible to predict with certainty whether or not a given HWC facility will have significant D/F emissions.
Therefore, it is anticipated that all HWC facilities will need to test for D/Fs. The remainder of this section
discusses key operating parameters that should be considered for D/F testing.
4.1.1
Participate Hold-TJp Temperatures
Several studies have demonstrated the importance of identifying critical operating parameters associated
with D/F emissions. Data described in Altwicker and others (1990), Harris and others (1994), Lanier and
others (1996), and EPA (1994a, 1996a, 1997a) indicate the importance of inlet temperatures for HWC
units equipped with dry APCDs (such as ESPs, fabric filters, or possibly high efficiency particulate air
[HEPA] filters). In general, these data indicate that, within the D/F formation window of approximately
400 to 750 ฐF, D/F formation can increase exponentially with increases in temperature. Thus, dry APCD
inlet temperature is a critical operating parameter. The lower temperature of 400 ฐF, versus 450 ฐF as
prescribed by the current BIF regulations, has been emphasized in evaluations conducted for the MACT
standards (EPA 1996a, 1997a).
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
26
-------
Additional data indicate that any particulate holdup areas (including boiler tubes and long runs of
ductwork) can serve as reaction sites for D/F formation if the temperature profile falls within the D/F
formation window. Santoleri (1995), citing numerous studies in Germany and the United States, notes that
facilities with heat recovery boilers have been found to have higher emissions of D/Fs than facilities
without heat recovery. The proposed mechanism is a result of boiler tube corrosion as the tubes trap ash
and form deposits. As HC1 gas passes over these deposits, the deposits and iron within the tubes react to
form C12 and iron chlorides, resulting in conditions conducive for D/F formation. The D/F emissions trend
for waste heat recovery boilers is further supported by EPA (1997a), who found that incinerators equipped
with recovery boilers have significantly higher D/F emissions than other incinerators. EPA (1997a) noted
that the heat recovery boilers preclude rapid temperature quench of combustion gases to a temperature of
less than 400 ฐF. Acharya and others (1991) hypothesized that D/Fs in a boiler could be minimized by
only cooling combustion gases to about 800 ฐF. Although energy recovery might be reduced, this would
keep the gases outside of the 400 to 750 ฐF range.
EPA (1997a) also found elevated D/F emission rates at some LWAKs where formation apparently
occurred in extensive runs of ductwork connecting the kilns to the fabric filters. EPA noted that reductions
of D/F emission rates could likely be achieved simply by rapidly quenching gases at the exit of the kiln to
less than 400 ฐF and insulating the ductwork to maintain gas temperatures above the dewpoint prior to the
fabric filter.
Results of these studies indicate that, for D/F testing, the relatively low temperature (approximately 400 to
750 ฐF) areas of particulate holdup downstream of the combustion zone should be emphasized. These
areas are conducive to surface-catalyzed D/F formation through mechanisms such as de novo synthesis.
Available data indicate that PM provides the substrate to act as a chemical reactor, given the appropriate
temperature, time, and presence of C12. Thus, any particulate holdup area (including fabric filters, ESPs,
HEPA filters, heat recovery boilers, and extensive runs of ductwork) can serve as a reactor for D/F
formation.
Particulate holdup temperatures should be considered very carefully in determining the appropriate test
condition for D/F testing. Unless the temperature fluctuation across the PM holdup device is negligible,
D/F testing should not be performed at normal or average holdup temperatures. D/F formation has been
observed to increase exponentially with increases in temperature over the range of approximately 400 to
750 ฐF (EPA 1994a, 1996a; Lanier and others 1996). Thus, a long-term average temperature limit will not
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
27
-------
necessarily ensure that D/Fs remain below the levels observed during a normal temperature test (i.e., the
D/F emissions from one minute of operation at 100 ฐF above normal could not be offset by one minute of
operation at 100 ฐF below normal). Unless a facility can provide a monitoring scheme that will reliably
ensure that D/Fs can be maintained below the levels observed during testing at average holdup
temperatures, then D/F emissions data should be collected while the facility is operating under maximum
particulate holdup temperatures.
4.12 Rapid and Partial Liquid Quench Systems
Ullrich and others (1996) describe the reduction of D/F emissions through the use of a rapid liquid quench,
which decreases residence time in the D/F formation window. A liquid quench involves rapid quenching
(on the order of milliseconds) from combustion temperatures to saturation temperatures of approximately
170 to 185 ฐF. HWC facilities that provide for rapid flue gas quenching to below saturation temperatures
generally have low D/F emissions. However, this may not necessarily be the case for facilities that
perform only a partial quench. Waterland and Ghorishi (1997) observed significant increases in D/F levels
in the flue gas as post-partial-quench temperatures increased from 711 to 795 ฐF (prior to the full quench).
The observed residence time between the partial quench and full quench chamber was approximately 0.5
seconds. This phenomenon, termed rapid high-temperature D/F formation, appears to be active in a
post-partial-quench temperature range of 570 to 800 ฐF.
Based on this information, it appears that operating limits on rapid quench systems are unnecessary for the
control of D/Fs. However, limits on post-quench temperatures from partial-quench systems are potentially
important.
4.1.3
Combustion Conditions
This section provides general information regarding the impact of combustion conditions on D//F
emissions. Further industry-specific discussion is provided in Sections 4.3 through 4.6. These discussions
are based on the underlying assumption that HWC facilities must operate under combustion conditions that
meet or exceed 99.99 percent ORE.
Combustion conditions and associated quality can play a key role in minimizing the formation of D/F
precursors, and thus, in potentially minimizing D/F emissions (EPA 1994a, 1996a). Berger and others
(1996) describe an increase in D/F, carbon monoxide, and total hydrocarbon emissions through poor
combustion in HWIs. High D/F emissions were observed only during the same incineration processes that
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
28
-------
included high total hydrocarbon emissions. Gullett and Raghunathan (1997) observed substantial increases
in D/F emissions under conditions of poor combustion and carbon monoxide levels greater than 2,000 parts
per million (ppm).
In order to assure combustion quality, EPA (1996a) has indicated that the following combustion
parameters should be demonstrated during D/F testing and controlled (during facility operation) to
minimize D/F precursors:
Minimum PCC and SCC combustion temperatures
Maximum combustion gas velocity
Maximum waste feed rates
For batch feeds,
maximum feeding frequency
maximum batch size
minimum oxygen concentration
Maximum carbon monoxide
Maximum total hydrocarbons
Unfortunately, it is often difficult to determine a direct correlation between an individual combustion
parameter and D/F emissions. Combustion processes involve complex physical and chemical interactions.
A change in a single independent variable can simultaneously impact several dependent variables. These
changes may or may not impact D/F emissions, and the most influential combustion parameters may not
always be the ones listed above. These points are demonstrated by the following two examples.
The first example involves minimum combustion temperature. Operating conditions associated with DRE
testing, including minimum combustion temperature, are generally believed to result in higher PIC
formation (and thus, potentially higher D/F emissions). This should be the case for most systems.
However, the opposite has been shown for incinerators feeding containerized wastes. For these units, pilot
testing shows that PIC emissions can be minimized by operating at lower PCC temperatures (Lemieux and
others 1990). Higher PCC temperatures and higher kiln rotation speeds result in rapid heating and
rupturing of the containers. Evolution of waste gases from the containers can exceed the rate at which the
stoichiometric amount of oxygen can be supplied, resulting in increased organic emissions rates. Lower
temperatures may lead to more gradual rupture of waste containers, and less disruptive transients. (The
tern "transient" refers to frequent changes in combustion conditions. These changes may be indicated by
recurring temperature, carbon monoxide, or total hydrocarbon spikes, or by frequent changes in
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
29
-------
combustion pressure.) The impact of this phenomenon on D/F emissions has been confirmed during at
least one trial burn at a HWC facility burning containerized wastes. At this facility, dioxin yields were
higher at maximum PCC temperatures than at minimum temperatures (EPA 1998c).
The second example involves oxygen concentration. Oxygen concentration is not specifically addressed
during many trial burns. In fact, it often varies considerably between test conditions when excess air is
used to simultaneously achieve minimum combustion temperature and maximum combustion gas velocity.
However, D/F emission rates may be impacted by oxygen levels. Gullett and Lemieux (1994) performed a
pilot study to investigate the impact of oxygen concentrations (as well as several downstream parameters)
on dioxin yields. Intermediate levels of oxygen (4.7 percent) were found to produce greater dioxin yields
than extreme levels (1.7 and 8.9 percent). In addition, oxygen significantly affected the partitioning
between dioxins and furans. Increases in oxygen favored formation of dioxins over furans.
These examples illustrate that the relationship between individual combustion parameters and D/F
emissions is not necessarily intuitive or readily demonstrated. Key parameters are likely to vary by
facility, and the facility-specific key parameters may or may not be those identified in EPA (1996a).
Because of these uncertainties, it is recommended that D/F emissions be determined during all of the
planned test conditions (e.g., DRE and SRE) at a HWC facility whenever possible. By characterizing D/Fs
over the entire range of combustion conditions, a facility can minimize the possibility of inadvertently
omitting combustion situations that may play a key role in D/F formation. In addition, the data collected
during multiple conditions can be analyzed for trends to determine the combustion parameters that should
be limited in the RCRA permit to control D/F emissions.
The recommendation for D/F sampling during all test conditions is a general guideline. However, some
facilities and permit writers may be faced with situations that are not addressed by this general guideline.
For example, DRE and SRE testing may have been conducted in advance of the sampling effort to collect
SSRA data, or stack sampling ports may not accommodate all of the necessary sampling trains for
consolidated testing. These and other situations call for decisions regarding the specific combustion
conditions to be demonstrated. Therefore, this guidance recommends that the following combustion
situations (if applicable) be preferentially targeted for D/F testing:
Transient conditions
Combustion of containerized or batch wastes
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
30
-------
Operation at high carbon monoxide levels, for units with carbon monoxide limits above
lOOppm
As appropriate, permit conditions for the combustion parameters listed in Tables 4-1 and 4-2 should be
established based on testing under the conditions indicated above. In addition, a facility-specific review of
trial burn and historical operating data should be performed to determine whether transient operations
correlate with other operating or feed parameters. If so, then the correlating parameters may be limited in
the permit in addition to, or in lieu of, the specific parameters listed in Tables 4-1 and 4-2.
Some HWC units do not operate under the scenarios identified above. For example, a liquid injection
incinerator feeding a single high-British thermal unit (Btu) waste stream may sustain very constant
temperatures and extremely low carbon monoxide concentrations. Ideally, D/F testing performed in
conjunction with the DRE test will demonstrate the combustion parameters indicated in Tables 4-1 and
4-2. However, if this is not possible then historical operating data for the appropriate combustion
parameters should be reviewed. Demonstration of absolute maximum or minimum values for combustion
parameters during D/F testing may be less critical if the review indicates steady-state operations with very
few fluctuations. For this situation, consideration may be given to testing under normal combustion
conditions. Periodic reporting to confirm continued absence of transients may be appropriate in lieu of
specific permit limits for the parameters listed in Table 4-1. When D/F testing is not performed in
conjunction with the DRE test, caution should be exercised to ensure that combustion parameters are not
substantially different from levels demonstrated during the DRE test.
The remainder of this section provides additional information on transient conditions, combustion of
containerized or batch wastes, and operating at high carbon monoxide levels for units with carbon
monoxide limits above 100 ppm.
4.1.3.1
Transient Conditions
The permit writer should review historical operating data to determine whether a facility experiences
routine transients, and, if so, the waste feed or operating conditions that cause the spikes should be
determined. The feeds or operating conditions causing transients represent candidate conditions for D/F
testing. Particular attention should be given to data indicating transients for combustion temperatures,
combustion chamber pressure, carbon monoxide, and total hydrocarbons. Instantaneous data may be more
useful in defining transients than rolling average data, which inherently dampen spikes.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
31
-------
During D/F testing, the facility should treat difficult-to-burn wastes under operating extremes that may
challenge combustion quality. Actual wastes (and not surrogate wastes synthesized from pure compounds)
should be used whenever possible. Candidate wastes should be selected based upon a review of the wastes
handled at a particular facility. Special consideration should be given to those wastes burned at
commercial facilities due to their variation and complexity. Examples of wastes that can cause transients
include:
Stratified or highly viscous liquids and sludges
Aqueous or low heating value liquids
Liquids with a high percentage of solids
Highly chlorinated wastes
Low heating value solids and sludges
Wastes with a high moisture content
Batch feeds with high moisture, volatility, or instantaneous oxygen demand
4.1.3.2
Containerized or Batch Wastes
Transient operations due to batch waste feeds are fairly common. D/F testing during batch feed conditions
should be performed regardless of carbon monoxide concentrations (which are generally measured
downstream of the SCC and which may or may not reflect the transients experienced in the PCC). Based
upon EPA (1996a) and Lemieux and others (1990), the following batch feed parameters should be
demonstrated during D/F testing:
Maximum feeding frequency
Maximum batch size
Maximum PCC combustion temperature
Maximum kiln rotation speed
Minimum oxygen concentration
A trial burn plan for a batch-fed facility should include a description of the procedures used to maintain
adequate oxygen while feeding batch or containerized wastes. Unless the oxygen demand from the batch
waste is insignificant compared to the oxygen demand of other fuels (e.g., 1-gallon containers fed to the
hot end of a cement kiln), EPA (1996a) suggests establishing a minimum oxygen limit at the end of the
combustion chamber into which the batch is fed, at the time the batch is fed. Implementation of minimum
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
32
-------
oxygen limits at the exit of the PCC on rotary kilns can sometimes be difficult, due to potentially
significant gas-phase stratification (Cundy and others 1991). If this is a problem, alternate monitoring
locations may need to be considered. Minimum oxygen limits for HWC facilities other than batch-fed
units are generally not necessary because emission limits for carbon monoxide will ensure that wastes are
not fed to the unit while excess air is at too low a level. However, if a HWC facility operates at conditions
that frequently exceed the carbon monoxide limits, the permit writer may consider establishing either a
minimum oxygen limit from the trial burn, or requiring an automatic control system to maintain fuel-to-air
ratios. Carbon monoxide may not always be a good indicator of combustion efficiency for cement kilns, as
discussed later in Section 4.5.
The physical and chemical composition of the batch waste is also important. Key characteristics include
volatility, instantaneous oxygen demand, moisture content, and heating value. Historical information on
operating trends and AWFCS events should be reviewed in an effort to determine which batch
characteristics are most likely to cause transients for a particular HWC facility. Some batch-charged and
containerized wastes can volatilize rapidly, causing an instantaneous release of heat and gases that
completely consume the available oxygen. This results in a momentary oxygen-deficient condition that
can result in poor combustion. Conversely, if too large a batch of aqueous waste or wet soil is fed, there is
danger that the batch can instantaneously quench temperature.
4.1.3.3
High Carbon Monoxide
Units with carbon monoxide limits above 100 ppm should perform D/F emissions testing while carbon
monoxide levels are maximized. EPA (1994a) evaluated D/F emissions data by normalizing the data for
APCD inlet temperature and carbon monoxide. Low carbon monoxide levels (less than 100 ppm) were
associated with very low D/F emissions (less than 1 nanogram per dry standard cubic meter [ng/dscm] on a
total basis). For carbon monoxide levels greater than 100 ppm, temperature-normalized dioxin emissions
were significantly higher (in the range of 10 to 100 ng/dscm on a total basis).
4.1.4
Feed Composition
In addition to the physical waste characteristics that can cause poor combustion, there are several chemical
characteristics that can potentially influence D/F emissions. These include chlorine concentration, the
presence of metals (such as copper, iron, and nickel) that can act as catalysts in D/F production
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
33
-------
mechanisms, the presence of D/F precursors (such as chlorobenzenes and chlorophenols), and the presence
of D/F inhibitors (such as sulfur and ammonia). Each of these is discussed below.
4.1.4.1
Chlorine
While the presence of chlorine is necessary for the formation of D/Fs, there does not appear to be a direct
correlation between the level of chlorine in the feed and the level of D/Fs in the flue gas in full-scale HWC
facilities. The American Society of Mechanical Engineers (ASME) (Rigo and others 1995) analyzed over
1,700 test results with chlorine feed concentrations ranging from less than 0.1 percent up to 80 percent, and
found no statistically significant relationship between D/F emission rates and chlorine concentration.
Obviously, no D/Fs could be formed without the presence of chlorine. However, other parameters, such as
APCD inlet temperature, are more statistically significant and any potential effect of chlorine feed input is
effectively masked.
EPA (1996a) is not proposing to limit the amount of chlorine fed to the HWC facility to ensure compliance
with the proposed D/F MACT standards. For D/F testing, chlorine feed rates should be maintained at
normal levels (i.e., chlorine should not be biased low). For purposes of this guidance, the term chlorine
feed rate refers to total chlorine from all sources, including both organic and inorganic forms. Chlorinated
wastes are preferred over non-chlorinated wastes, where the choice exists. However, specific HRA limits
on total chlorine are not anticipated based upon the D/F testing.
4.1.4.2
Metal Catalysts
Abundant pilot-scale and fundamental research has shown that certain metals, such as copper, may catalyze
the formation of D/Fs. This phenomenon has not been observed during full-scale testing (Lanier and
others 1996); however, the testing may have been conducted in a system that was influenced by other,
more dominant factors. EPA (1996a) is not proposing to limit the amount of catalytic metals to ensure
compliance with the future D/F MACT standards. Wastes or other feed materials containing copper are
preferred over feeds without copper during the D/F testing, where the choice exists. However, specific
limits on copper (or other catalytic metals) are not anticipated based upon the D/F testing.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
34
-------
4.1.4.3
D/F Precursors
Some HWIs that burn D/F precursors, including chlorobenzenes, chlorophenols, and PCBs, have been
shown to have high D/F emissions. EPA (1996a) compared a limited number of facilities that feed known
D/F precursors to those that do not feed D/F precursors. This limited study suggested no strong correlation
between the level of precursors and D/F formation; however, the issue has not been examined in detail. If
a facility burns wastes with significant quantities of D/F precursors, these wastes are preferred over wastes
without precursors for D/F testing. Although specific permit limits on D/F precursors are not anticipated,
the permit writer may require waste profile tracking to determine whether increased quantities of precursor
wastes warrant retesting.
4.1.4.4
D/F Inhibitors
D/F inhibitors, such as sulfur, have been commercially marketed as feed stream additives to control D/F
emissions. These same compounds may naturally be present in fossil fuels (such as coal) or hazardous
waste fuels. Raghunathan and Gullett (1994) and Raghunathan and others (1997) conducted bench and
pilot-scale tests of municipal solid waste combustion facilities and concluded that co-firing with coal can
effectively reduce D/F emission rates. Significant decreases in D/F emission rates were observed at a
sulfur to chlorine ratio of 0.64 (Raghunathan and Gullett 1994). Depletion of active chlorine by sulfur
dioxide through a gas-phase reaction appears to be a significant inhibition mechanism, in addition to sulfur
dioxide deactivation of copper catalysts. In reviewing the D/F test protocol, the permit writer should
ensure that the facility will not burn a high sulfur waste or fuel in greater quantities than during normal
operation. The permit writer may require waste and fossil fuel tracking to determine whether burning
decreased quantities of sulfur warrant retesting.
4.1.4.5
Other Factors
Other waste feed components may also potentially affect D/F emissions. The presence of bromine, in
particular, has been found to affect emissions of chlorinated organic PICs and D/Fs in pilot-scale
experiments (Lemieux and Ryan 1998; Lemieux and Ryan in press). Although the effects of the presence
of bromine has not been clearly established during full-scale testing, permit writers should be aware of its
potential when selecting waste feeds for trial burns, particularly if the facility burns brominated waste
during normal operations.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
35
-------
4.1.5
D/F Control Technologies
Some facilities may install specific D/F control technologies. These include carbon injection, carbon beds,
catalytic oxidizers, and D/F inhibitor technologies. If a facility uses one of these technologies, then permit
limits on key operating parameters should be established during D/F testing. Relevant operating
parameters are identified in EPA (1996a, 1996d).
4.2 OPERATING PARAMETERS ASSOCIATED WITH D/F PRODUCTION
Based on a review of existing information, this guidance prioritizes operating parameters and conditions
associated with D/F formation as primary, secondary, or tertiary. These hierarchial designations should
not be considered absolute, but are intended to emphasize the relative importance of demonstrating various
operating parameters during D/F testing and limiting those parameters in the final RCRA permit.
Parameters related to combustion conditions are categorized as primary; however, this designation should
be tempered by the previous discussion for steady-state systems. A description of primary, secondary, and
tertiary operating parameters follows:
Primary operating parameters are those that have shown the highest correlation with D/F
emission rates during full-scale testing, and are expected to dominate D/F formation.
These parameters should always be demonstrated during the D/F test, and should be
limited in the permit by specific quantitative limits. These operating parameters relate to
either surface-catalyzed D/F formation, or the use of specific D/F control technologies and
include:
Inlet temperature to dry APCDs
Temperature profiles over particulate holdup areas (including long runs of
ductwork, economizers, and boiler tubes)
Key operating parameters for specific D/F control technologies
Combustion parameters listed in Tables 4-1 and 4-2
Secondary operating parameters are those that may influence D/F emissions under certain
circumstances. However, there is less information indicating a direct correlation between
these parameters and D/F emission rates. These parameters may or may not need to be
demonstrated during the D/F test and limited in the permit, depending on the significance
of these parameters for a given system configuration and the presence or absence of
dominant primary parameters. Secondary parameters include:
Conditions other than combustion quality that could lead to the formation
of organic precursors (such as organics from raw materials in cement kilns
and LWAKs)
Flue gas temperatures due to partial quenching
Tertiary operating parameters are those that relate to feed composition. These operating
parameters have been the subject of fundamental and pilot-scale research on D/F
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
36
-------
formation, but have not routinely been correlated with D/F emissions during full-scale
testing. These parameters may influence the selection of feeds for D/F testing and
subsequent waste profile tracking, but are not expected to be limited in the permit by
specific feed rate limits. Tertiary parameters include:
Chlorine feed rates
Presence of D/F catalysts (such as copper)
Presence of D/F precursors (such as chlorinated aromatics)
Presence of naturally-occurring D/F inhibitors (such as sulfur)
The following subsections discuss critical D/F operating parameters in more detail as they relate to specific
types of HWC facilities.
4.3
D/F EMISSIONS FROM HWIS
HWIs include rotary kiln, liquid injection, fluidized bed, and fixed hearth designs. Commercial HWIs
typically accept hazardous waste from generators throughout the United States. Waste feeds to these units
can be highly variable, for example waste feed material may include low- and high-Btu liquids, as well as
solids from laboratory packs and soils contaminated with low levels of RCRA hazardous wastes. Large
chemical complexes may operate captive HWIs that treat waste feeds generated on site and from corporate
affiliates off site. These wastes may also be highly variable, especially if the facility burns a number of
wastes from different production operations and does not have the capability to blend the wastes to a
consistent specification. Small chemical companies may generate only one or two waste streams. These
wastes are typically more predictable and homogeneous.
HWIs are generally associated with two-stage APCDs (EPA 1996a) that first cool hot flue gases and then
remove PM, metals, and organics. Most HWIs use wet APCDs (three were cited that use dry scrubbers).
Typical APCDs include (1) packed towers, spray dryers, or dry scrubbers for temperature reduction and
acid gas control and (2) venturi scrubbers, wet or dry ESPs, or fabric filters for PM, metal, and organics
control. Some new technologies are being developed, and several facilities are injecting activated carbon
in the spray dryers for control of D/Fs, non-D/F organics, and mercury (EPA 1996a). Some HWIs may
have heat recovery boilers that affect D/F emissions.
The level of D/F emissions from HWIs may be dependent on incinerator design, APCD type, particulate
hold-up temperatures, type of quench or presence of a heat recovery unit, combustion conditions, and feed
composition. In summary, all of the considerations discussed previously in Section 4.1 apply to HWIs.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
37
-------
Table 4-1 summarizes operating parameters associated with D/F emissions from HWIs. Recommended
averaging periods are discussed further in Section 8.0. Depending on the system configuration,
demonstration of operating parameters associated with D/F formation may coincide with both the DRE and
SRE test conditions. If dry APCD equipment or heat recovery devices are present in the HWI system, the
temperature profile across these systems is recognized as a primary operating parameter directly related to
D/F formation. Therefore, for these systems, D/F data collection may be performed in conjunction with
SRE testing (unless the facility can adjust inlet temperature to obtain the requisite temperature profiles
during DRE testing). Demonstration of operating parameters affecting combustion efficiency (especially
for transient operations, units burning containerized wastes, or high carbon monoxide situations) will most
likely coincide with the DRE test condition.
Facilities with more predictable, homogeneous waste feeds, few operating fluctuations, and no particulate
holdup devices may opt to collect D/F emissions data during a risk burn conducted under normal operating
conditions. Waste feed selection is based on a representative waste stream, with a preference for D/F
precursors such as chlorophenols and minimal amounts of D/F inhibitors (such as sulfur).
4.4
D/F EMISSIONS FROM BOILERS
General boiler designs are discussed by EPA (1994a), and requirements for boilers burning hazardous
waste are defined in 40 CFR Part 266.100 et seq. Boilers recover the heat from hazardous waste
combustion to pressurize water. The three most common boiler designs used for treating hazardous waste
include firetube boilers, watertube boilers, and stoker-fired boilers. Most boilers treating hazardous waste
are on-site units at chemical production facilities. Most boilers do not have APCDs. Historically,
emissions tests from boilers have focused on metals and PM, and the database for D/F emissions from
boilers is not as extensive as it is for D/F emissions from HWIs and cement kilns.
D/F emissions from boilers are expected to be dependent on boiler design, APCD type, particulate hold-up
temperatures, combustion conditions, and feed composition. Table 4-1 summarizes operating parameters
associated with D/F and other organic emissions from boilers. Recommended averaging periods are
discussed further in Section 8.0. Depending on the system configuration, demonstration of operating
parameters associated with D/F formation in boilers may coincide with both the DRE and SRE test
conditions.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
38
-------
TABLE 4-1
OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM HWIS AND BOILERS
*t ' * >\ ( v *
Operating
Parameters
Most Likely
Achieved
During
Parameter
, Type/Suggested
Averaging Periods* %t
How Limit is
Established
Other Considerations
=" tt ' > '
PRIMARY OPERATING PARAMETERS
Surface-Catalyzed Formation:
(DryAPCD)
Surface-catalyzed formation is a
predominant D/F formation mechanism
for post-combustion dry APCD
particulate holdup areas operating at
temperatures between 400-750 ฐF.
D/F-Specific Control Technology:
D/F-specific control technologies
include carbon injection, carbon bed,
and inhibitor technologies.
Maximum dry ESP inlet
temperature
Maximum FF inlet
temperature
Maximum HEPA filter
inlet temperature
Boiler exit temperature
SRE test,
unless a
variable
quench is used
Any test that
achieves the
critical
temperature
window
Group A:
Dual 10 minute/ 1 hour
Group A:
Dual 10 minute/ 1 hour
Average of three maximum
10-minute RAs/Average of
three maximum HRAs
Average of three minimum
or maximum 10-minute
RAs/Average of three
minimum or maximum
HRAs (depending on
which edge of the boiler
operating range is in the
critical temperature
window)
Paniculate loading should
not be biased low during
the test, based upon a
review of:
- ash feed rate
- combustion gas
velocity
- APCD operation
Ongoing PM control is
assured by limits on APCD
operating parameters
established during the PM
test.
If a specific control technology is used to limit D/F emissions, operating, limits should be established per EPA (1996a, 1996d).
Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments
at Hazardous Waste combustion Facilities
August 1998 Peer Review Draft
39
-------
TABLE 4-1
OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM HWIS AND BOILERS (Continued)
Operating
Parameters
Most Likely
Achieved
During
Parameter
Type/Suggested
Averaging Periods*
How Limit is
Established
Other Considerations
PRIMARY OPERATING PARAMETERS (Continued)
Combustion Conditions Related to
Formation of D/F Precursors:
(These parameters should also be
limited to control non-D/F organics, as
discussed in Section 5. 1)
Operating parameters to limit D/F
precursors from poor combustion are
most critical for transient operations.
Transient operations may be identified
by frequent temperature, carbon
monoxide, oxygen, or total hydrocarbon
spikes.
Operating parameters related to good
combustion may be less critical for
steady-state operations. Although
demonstration of these operating
parameters during DRE conditions is
preferred whenever possible, D/F
testing at normal combustion conditions
may be considered for some steady-
state units. Record keeping and
periodic reporting to confirm continued
absence of transients may be considered
in lieu of HRAs or 10-minute averages.
Minimum combustion
temperature, each chamber
Maximum PCC
temperatures should be
demonstrated for units
burning containerized
wastes
Maximum combustion gas
velocity
Maximum waste feed rate,
each location
DRE
DRE/
SRE
DRE
Group A:
Dual 10-minute/l hour
Group A:
1 hour
Group A:
1 hour
Average of three minimum
1 0-minute RAs/Average of
three minimum HRAs
Average of three maximum
HRAs
Average of three maximum
HRAs
Limits should be
established for:
- maximum organic
liquids to PCC
- maximum aqueous
liquids to PCC
- maximum sludges to
PCC
- maximum solids to PCC
- maximum organic
liquids to SCC
- maximum aqueous
liquids to SCC
Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments
at Hazardous Waste combustion Facilities
August 1998 Peer Review Draft
40
-------
TABLE 4-1
OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM HWIS AND BOILERS (Continued)
\ t
Operating
Parameters
Most Likely
Achieved
During
, Parameter
Type/Suggested
' Averaging Periods*
How. Limit is- r~
Established
f
Other Considerations
PRIMARY OPERATING PARAMETERS (Continued)
Combustion Conditions Related to
Formation of D/F Precursors:
(Continued):
Waste variability that
could cause transients
Batch feed conditions:
- batch size
- batch frequency
- minimum oxygen level
- maximum PCC
temperature
- maximum kiln
rotation speed
Maximum carbon
monoxide and total
hydrocarbons
DRE
ORE
DRE
This is not a continuously monitored parameter, but
pertains to selection of wastes for testing. Conditions
for waste profile tracking may be specified by the
permit writer.
Group B:
Per batch
Group A:
1 hour
Group A:
1 hour
Batch:
- size demonstrated
during test
- frequency demonstrated
during test
- oxygen level
demonstrated during
test
Average of three maximum
HRAs
Average of three maximum
HRAs, or 100 ppm carbon
monoxide, whichever is
higher
Wastes with physical
properties that can cause
combustion transients (as
discussed in Section 4.1)
should be selected.
Test wastes with high
volatility and oxygen
demand.
None
Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments
at Hazardous Waste combustion Facilities
August 1998 Peer Review Draft
41
-------
TABLE 4-1
OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM HWIS AND BOILERS (Continued)
Operating
Parameters
Most Likely
Achieved
During
Parameter
Type/Suggested
Averaging Periods*
How Limit is
Established
Other Considerations
SECONDARY OPERATING PARAMETERS
Rapid High Temperature Formation:
(WetAPCD)
May be a concern for partial quench
situations with post-partial quench gas
temperatures between 570-800 ฐF.
Not a concern for rapid wet quench
systems that cool gases to saturation
temperatures within milliseconds.
Maximum post-partial
quench gas temperature
Any test
condition that
achieves the
critical
temperature
window for
D/F formation
Group A:
Dual 10-minute/l hour
Average of three maximum
10-minute RAs/Average of
three maximum HRAs
None
TERTIARY OPERATING PARAMETERS
Feed Composition:
Wastes should be chosen based on
consideration of chlorine and D/F
precursors, catalysts, and inhibitors.
Total Chlorine
D/F Precursors
D/F Catalysts
D/F Inhibitors
These are not continuously monitored parameters, but pertain to selection
of wastes and fuels for testing. Conditions for waste profile tracking may
be specified by the permit writer.
Considerations are
discussed in Section 4.1.
Notes: APCD = air pollution control device
D/F = dioxins and furans
DRE = destruction and removal efficiency
EPA = U.S. Environmental Protection Agency
ESP = electrostatic precipitator
FF = fabric filter
HEPA = high efficiency paniculate air
HRA = hourly rolling average
PCC = primary combustion chamber
PM = paniculate matter
RA = rolling average
SCC = secondary combustion chamber
SRE = system removal efficiency
Hourly and 10-minute rolling averages are specified as examples, but other averaging periods and techniques may be considered. When establishing permit
limits that are based on the average of the three highest (or lowest) rolling averages, it is important to ensure that the test is conducted in a manner that only
allows for normal variability about a central value. For example, it would not be acceptable to conduct the test at 15 minutes of artificially high carbon
monoxide concentrations, with the remainder of the test at normal levels. One way to avoid this is to establish the permit limit as the time-weighted average
over all runs. Averaging periods are also discussed in Section 8.0
Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments
at Hazardous Waste combustion Facilities
August 1998 Peer Review Draft
42
-------
As explained in Section 4.1, boiler tubes may serve as particulate holdup areas and lead to D/F emissions.
D/Fs may form when boiler flue gases are within the D/F formation temperature window. Because boilers
typically have no rapid quench, the time and temperature window for D/F formation may be large.
Therefore, boiler exit temperature (which can include temperatures at heat exchangers and economizers) is
considered a primary operating parameter for D/F formation and control. Collection of D/F emissions data
for boilers is recommended during conditions that achieve boiler exit temperatures in the upper end of (but
well within) the 400 to 750 ฐF range. For example, for a facility with boiler exit temperatures ranging
from 350 to 550 ฐF, D/F testing at the boiler exit temperature of 550 ฐF would be preferred over testing at
the exit temperature of 350 ฐF. Boiler exit temperatures may fall in the upper end of the D/F formation
window during either ORE or SRE conditions, depending on the facility-specific operating envelope.
Demonstration of parameters related to combustion quality can also be a consideration, especially for
boilers that burn wastes resulting in combustion transients. Some boilers at chemical facilities burn
different production run wastes in campaigns. These conditions should be evaluated by the permit writer
prior to trial burn to determine the potential for transients. Demonstration of operating parameters
affecting combustion efficiency will most likely coincide with the DRE test condition.
Demonstrating key operating parameters related to combustion quality can sometimes be problematic for
boilers based on potential test condition conflicts (Schofield and others 1997). For example, a facility with
a fixed combustion air flow rate burning a single high-Btu waste stream will not be able to demonstrate
minimum combustion temperature and maximum feed rate simultaneously. Thus, two test conditions may
be needed to demonstrate all of the key control parameters related to combustion. However, if combustion
air can be controlled, then temperature could be minimized and feed rate could be maximized
simultaneously by adjusting the amount of combustion air.
In some cases, D/F testing during the DRE condition may not be possible for reasons discussed in Section
4.1 (e.g., because of sampling port limitations, or because the risk testing is being performed separately
from performance testing). In these situations, a facility with predictable, homogeneous waste feeds and
few combustion transients may opt to test during a test condition that represents normal combustion
conditions. The facility would still need to demonstrate boiler exit temperatures in the upper end of the
400 to 750 ฐF range.
In general, facilities with highly variable operations should collect D/F emission samples during DRE
conditions and any other condition that is necessary to achieve boiler exit temperatures in the upper end of
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 43
-------
the 400 to 750 ฐF window. This could result in multiple test conditions. Facilities with more predictable,
homogeneous waste feeds and few combustion transients may need to test only during the test condition
achieving the requisite boiler exit temperatures.
Permit writers should also be aware of soot blowing practices at boilers because high particulate loading
due to this practice could affect D/F emissions. The permit writer should determine normal sootblowing
procedures from the facility's operating record. Sootblowing should be performed during D/F testing to
capture the potential impact of higher particulate loading on D/F emissions. However, sootblowing
should not be performed on a more rapid cycle than normal, because this could potentially shorten the
residence time for D/F formation, and decrease D/F emissions (Santoleri 1995). EPA (1992b) provides
guidance on structuring test runs to reflect sootblowing practices.
4.5
D/F EMISSIONS FROM CEMENT KILNS
Background information on potential D/F emissions from cement kilns is summarized by EPA (1994a,
1996a). Cement kilns may use hazardous waste as a supplementary fuel while producing a salable product.
In general, the operating envelope of cement kilns is dictated in large part by the American Society for
Testing and Materials (ASTM) requirements for their final product. Cement kilns also have regions that
operate at high temperatures approaching 3,000 ฐF. Based on these characteristics, issues related to good
combustion and minimum combustion temperatures are less relevant, as compared to HWIs and boilers.
Also, because of the chemical composition of the raw materials, carbon monoxide and total hydrocarbon
concentrations may not always serve as indicators of good combustion. According to EPA (1996a) all
hazardous waste burning cement kilns use either fabric filters or ESPs as APCDs.
Table 4-2 summarizes operating parameters associated with D/F and other organic emissions from cement
kilns and LWAKs. Data presented by Harris and others (1994) and Lanier and others (1996) demonstrate
that D/F emissions from cement kilns increase exponentially with increases in inlet temperatures to the dry
APCD while within the D/F formation window (400 to 750 ฐF). Given these conditions, maximum inlet
temperature to the dry APCD system is the primary operating parameter related to D/F emissions for
cement kilns. Collection of D/F emission data should occur during conditions that achieve maximum
APCD inlet temperatures. These conditions may coincide with the SRE test if the APCD inlet temperature
cannot be independently controlled from combustion temperature.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
44
-------
TABLE 4-2
OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM CEMENT KBLNS AND LWAKS
>
Operating
Parameters
Most Likely
, Achieved
Daring
i
Parameter
Type/Suggested
Averaging
Periods*
How Limit is
: ; Established
' J-
, Other Considerations
PRIMARY OPERATING PARAMETERS
Surface-Catalyzed Formation:
(DryAPCD)
Surface-catalyzed formation is a
predominant mechanism for post-
combustion dry particulate holdup areas
operating at temperatures between 400-
750 ฐF.
D/F-Specific Control Technology:
D/F-specific control technologies include
carbon injection, carbon bed, and inhibitor
technologies.
Maximum dry ESP
inlet temperature
Maximum FF inlet
temperature
LWAKS: Maximum
inlet temperature to
extensive runs of
ductwork
SRE
SRE
Group A:
Dual 10 minute/
1 hour
Group A:
Dual 10 minute/
1 hour
Average of three maximum
10-minute RAs/Average of
three maximum HRAs
Average of three maximum
10-minute RAs/Average of
three maximum HRAs
Ongoing PM control is
assured by limits on APCD
operating parameters
established during the PM
test.
If a specific control technology is used to limit D/F emissions, operating limits should be established per EPA (1996a and
1996d).
Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments
at Hazardous Waste combustion Facilities
August 1998 Peer Review Draft
45
-------
TABLE 4-2
OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM CEMENT KILNS AND LWAKS (Continued)
Operating
Parameters
Most Likely
Achieved
During
Parameter
Type/Suggested
Averaging
Periods*
How Limit is
Established
Other Considerations
SECONDARY OPERATING PARAMETERS
Good Combustion to Control D/F
Precursors:
Note: These parameters should also be
limited to control non-D/F organics, as
discussed in Section 5.2
Applicable only to kilns that feed wastes at
locations other than the hot end of the kiln.
Control of Precursors from Raw Material
Organics:
Note: These parameters should also be
limited to control non-D/F organics, as
discussed in Section 5.2
Total hydrocarbons originating from raw
materials may lead to formation of
chlorinated organics that could potentially
serve as D/F precursors.
Batch feed
conditions:
- batch size
- batch frequency
- feed location
- minimum oxygen
- maximum
temperature at
feed location
Maximum total
hydrocarbons, as
measured at both the
main and bypass
stacks, not to exceed
20 ppmv per BIF
Any test
SRE
Group B:
Per batch
Group A:
,
1 nour
Batch:
- size demonstrated in test
- frequency demonstrated in
test
- location demonstrated in
test
- oxygen level demonstrated
in test
- temperature demonstrated
in test
20 ppmv, regulatory limit, at
the monitoring location used
for BIF compliance.
Limits for other locations
will be considered based on
the results of the SSRA.
Test wastes with high
volatility/oxygen demand.
Kiln rotation speed is
generally limited by the
production process and need
not be limited for cement
kilns.
Temporary total hydrocarbon
monitors may be needed if
the facility does not normally
measure total hydrocarbons.
Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments
at Hazardous Waste combustion Facilities
August 1998 Peer Review Draft
46
-------
TABLE 4-2
OPERATING PARAMETERS ASSOCIATED WITH D/F EMISSIONS FROM CEMENT KILNS AND LWAKS (Continued)
'Operating
Parameters
Most Likely
Achieved
During '
Parameter
Type/Suggested
Averaging *
Periods* *
How Limit is
Established
Other Considerations,
TERTIARY OPERATING PARAMETERS
Feed Composition:
Total chlorine and the presence of D/F
inhibitors such as sulfur in coal should be
considered during selection of wastes and
other fuels.
Total Chlorine
D/F Inhibitors
These are not continuously monitored parameters, but pertain to
selection of wastes and fuels for the testing. Conditions for waste profile
tracking may be specified by the permit writer.
Normal to high levels of total
chlorine should be
maintained during the D/F
testing.
Coal should not be fed at
higher than normal rates
during the D/F testing, and
low-sulfur coal is preferable
if the facility uses coal with
varying sulfur content.
Notes: APCD = air pollution control device
BIF = boiler and industrial fur
D/F = dioxins and furans
EPA = U.S. Environmental Protection Agency
ESP = electrostatic precipitator
FF = fabric filter
HRA = hourly rolling average
PM = participate matter
ppmv = parts per million volume
RA = rolling average
SRE = system removal efficiency
Hourly and 10-minute rolling averages are specified as examples, but other averaging periods and techniques may be considered. When establishing permit
limits that are based on the average of the three highest (or lowest) rolling averages, it is important to ensure that the test is conducted in a manner that only
allows for normal variability about a central value. For example, it would not be acceptable to conduct the test at 15 minutes of artificially high carbon
monoxide concentrations, with the remainder of the test at normal levels. One way to avoid this is to establish the permit limit as the time-weighted average
over all runs. Averaging periods are also discussed in Section 8.0.
Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments
at Hazardous Waste combustion Facilities
August 1998 Peer Review Draft
47
-------
The operating parameters in Table 4-2 related to combustion conditions are limited to situations where
kilns feed hazardous waste at locations other than the hot end of the kiln. Controls on waste charging rate
and kiln oxygen concentration are recommended because wastes injected at mid- or feed-end locations may
not experience the same elevated temperatures and long residence times as those wastes injected at the hot
end. In a worst-case scenario, volatile compounds may be released from the charge so rapidly that they are
not able to mix with oxygen and ignite before they cool below a critical temperature forming PICs
(Dellinger and others 1993).
Table 4-2 does not establish control parameters related to combustion of hazardous wastes introduced to
the hot end of kilns. Results from both kinetic modeling and field studies suggest that organics are
efficiently destroyed when fed at -the hot end of cement kilns (Dellinger and others 1993). DRE failures at
cement kilns are extremely limited, and can generally be explained by high blank or baseline (non-
hazardous waste) levels of POHCs. In one instance, DRE failure has been attributed to poor atomizer
design. However, facility-specific DRE testing should be sufficient to reveal design problems.
In cement kilns, main stack emissions of total hydrocarbons are dominated by organics that are volatilized
from the raw materials prior to entering the high temperature regions of the kiln. The chlorination of these
hydrocarbons is a potential source of chlorinated hydrocarbon emissions, including D/F precursors such as
monochlorobenzene (Dellinger and others 1993). Therefore, D/F testing should be performed at the upper
end of the operating range for total hydrocarbons, as measured in both the main and bypass stacks, not to
exceed 20 parts-per-million volume (ppmv) at the monitoring location used for BIF compliance. Although
the operating conditions necessary for achieving high total hydrocarbon emissions may vary by facility,
maximum total hydrocarbon levels are likely to be achieved by some combination of high production rate,
high gas temperatures at the raw material feed end of the kiln, and low oxygen at the raw material feed end
of the kiln. Dellinger and others (1993) observed an inverse relationship between total hydrocarbons and
stack oxygen concentrations. The organic content of the raw material can also significantly influence
hydrocarbon levels, but the raw materials are not easily controlled for the purpose of testing. If total
hydrocarbon levels increase substantially due to changes in raw materials, then re-testing may be
necessary. Organic emissions from LWAKs are generally expected to be less than those from cement
kilns. This is because the feed material is usually shale or slate with low organic carbon content.
However, the objectives for maximizing total hydrocarbons still apply, consistent with those provided for
cement kilns.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
48
-------
In the context of D/F and other organic testing, total hydrocarbons are used as an operating parameter
indicating levels of organics within raw materials that may be chlorinated from the hazardous waste fuel.
In this case, total hydrocarbons are not being used as an indicator of good combustion or combustion
efficiency. The SSRA quantifies risks from organic emissions from the HWC facility, regardless of
source. Therefore, facilities that only monitor carbon monoxide under the BIF regulations (some
LWAKs), or cement kilns that only monitor carbon monoxide or total hydrocarbons in a bypass stack, may
need to install temporary total hydrocarbon monitors on the main stack prior to and during the D/F and
other organic tests to ensure that total hydrocarbon emissions are being maximized. The need for
permanent total hydrocarbon monitoring is assessed by the permit writer after the SSRA is completed and
potential risks are compared to target risk levels. Carbon monoxide may not always be a good indicator of
organic emissions from cement kilns. Carbon monoxide is generated during the calcining of calcium
carbonate, and may also be formed at the kiln exit where some of the total hydrocarbons from the raw
materials are oxidized.
Normal levels of chlorine in wastes should be maintained during D/F and other organic emissions testing.
It has been proposed that the highly alkaline environment in a cement kiln scavenges available chlorine,
making it unavailable for chlorination of organics. However, equilibrium calculations show lower chlorine
capture at high temperatures and conversion of HC1 to C12. Thus, even a highly basic chemical species
such as calcium hydroxide would not be expected to effectively control chlorinated hydrocarbon formation
(including D/Fs) at temperatures above 400 ฐF (Dellinger and others 1993).
Naturally occurring D/F inhibitors, such as sulfur, are expected to be present in the coal used for co-firing a
cement kiln. During the D/F testing, coal should not be fed at higher-than-normal rates, and low sulfur
coal is preferred if a facility uses several coal suppliers. Other potential D/F inhibitors, such as calcium,
are already present in the raw materials.
Metal catalysts in the waste are not expected to be relevant to D/F testing at cement kilns. Spiking wastes
with copper were not observed to affect D/F emission rates during full-scale testing of a cement kiln
(Lanier and others 1996). Also, other metals that have been studied as D/F catalysts (iron and aluminum)
are major ingredients in cement kiln raw materials.
D/F precursors at cement kilns are expected to be dominated by precursors in the raw material, and not by
precursors in the waste. However, if a facility burns wastes with significant quantities of D/F precursors,
these would be preferred over wastes without the precursors.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 49
-------
4.6
D/F EMISSIONS FROM LWAKS
The operation of LWAKs is similar to cement kilns in that (1) the operating temperature range is dictated
by ASTM consideration of the final product and (2) temperature at the hot end varies from 2,050 to 2,300
ฐF (EPA 1996a). Combustion gas exit temperatures vary from 300 to 1,200 ฐF depending on the feed and
system design. LWAKs typically burn only high-Btu, liquid fuel, and do not burn wastes at locations other
than the hot end. According to EPA (1996a), all LWAKs using hazardous waste as a fuel use fabric filters
for PM control.
Table 4-2 summarizes operating parameters associated with D/F and other organic emissions. As with
cement kilns, dry APCD inlet temperature is the primary operating parameter related to D/F formation.
The need to demonstrate combustion parameters should be evaluated on a case-by-case basis. As
appropriate, permit writers may also wish to consider combustion parameters as permit conditions.
LWAKs do not operate at combustion temperatures as high as those in cement kilns. However, the
potential for combustion transients may be minimized because LWAKs typically only burn high-Btu,
liquid wastes in the flame zone.
An additional concern for some LWAKs is the use of long runs of duct work (between the kiln, fabric
filter, and stack) that can lead to particle entrainment and high D/F emissions. This particulate holdup area
should be evaluated as a primary issue related to D/F formation. D/F emission data collection is most
appropriate during the upper end of the temperature operating envelope (SRE) due to the importance of the
inlet temperature to the dry APCD and duct work.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
50
-------
5.0 ORGANIC EMISSIONS OTHER THAN DIOXEVS AND FURANS
Completion of human health and ecological SSRAs includes the evaluation of organic emissions other than
D/Fs (EPA in press a, in press b). EPA has conducted research to identify the types and quantities of
organics emitted from HWC facilities (EPA 1997rc; Ryan and others 1996, 1997; and Midwest Research
Institute and A.T. Kearney 1997). These research efforts indicate frequent detection of volatile and
semivolatile organics including chloro-, bromo-, and mixed bromochloro-alkanes, alkenes, alkynes,
aromatics, polyaromatics, and D/Fs; nitrogenated and sulfonated organics; and short-chain alkanes (such as
methane and propane). Additional semivolatile compounds detected include oxygenated compounds
(PAHs and heterocycles), phthalates, phenols, halogenated mono- and polyaromatics, and nitrogenated and
sulfonated compounds. EPA (1997c) reports frequent detection of chlorinated and brominated alkanes and
alkenes (such as chlorinated ethenes), and suggests that chlorinated ethenes can serve as potential
indicators of D/F formation. EPA supports ongoing research to improve the identification and
quantification of organic emissions. However, uncertainty remains regarding the full suite of organic
emissions from HWC facilities and associated potential risks. Therefore, it is anticipated that all HWC
facilities will need to test for organic emissions to assess potential risks on a site-specific basis.
Testing for organic emissions should accomplish the following two objectives:
Identification and quantification of specific toxic organic compounds (such as PAHs and
PCBs) to assess their contribution to the total potential risk posed by the facility
Completion of a mass balance for total organics (including nontoxic organics) to reduce
and evaluate the uncertainty associated with the risk assessment process
Target analyte lists for identifying specific organics during laboratory analysis are provided in Appendix B
of this document. Standard EPA methods (1996g) can be used to identify and quantify many organics that
are potentially toxic, persistent, and bioaccumulative, such as PCBs and PAHs. For volatile and
semivolatile organic compounds, SW-846 Methods 8260 and 8270 gas chromatography/mass spectrometry
(GC/MS) procedures are preferred (EPA 1996g). Finally, a MS library search to tentatively identify
nontarget compounds is recommended. Tentatively identified compounds that could significantly
contribute to risk should be confirmed and quantified through the use of known standards.
The EPA methods, as written, are intended as guidance and should be used as starting points for the
development of standard operating procedures that will actually be used. For collection of SSRA
emissions data, method flexibility is allowed and guidance on appropriate method modifications is
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
51
-------
provided in SW-846. (Refer to Chapter 2 of EPA [1996g] and the applicable methods.) However, the
facility must be able to demonstrate and document that the methods used to generate the data meet the data
quality objectives (DQO) for the particular application. Other methods not found in SW-846 may also be
used provided that the user can demonstrate and document the methods used to generate data that meet the
appropriate DQOs.
Unfortunately, only a limited number of organic compounds can be accurately identified and quantified
using standard stack gas sampling and analysis methods. The mass of organic emissions that cannot be
readily identified and quantified using standard EPA (1996g) methods is estimated using the TO procedure
which sums organic fractions determined for three boiling point ranges (EPA 1996f). The TO procedure is
discussed further in Appendix B. TO values can be good indicators of uncertainty in the SSRAs because
they serve as a measure of the fraction of TO mass that has not been identified and quantified (TO mass
minus the total mass of speciated compounds). Thus, a TO determination should be performed to
supplement the organic emissions testing at each HWC facility.
An additional consideration in developing planned sampling and analytical procedures is that many
organics lack EPA toxicity values, or they may not be toxic. These compounds are typically not included
on standard target analyte lists. However, an important advantage of detecting, identifying, and
quantifying as many organics as possible is the reduction of uncertainties associated with the risk
assessment process. Determinations for methane, propane, and other short-chain aliphatics add little cost
to the emissions testing, and can potentially alleviate concerns about the percentage of the total organic
mass that might be toxic. This is discussed in more detail in Appendix B.
Operating parameters and conditions affecting organic emissions are expected to vary between different
types of HWC units. The following sections provide some guidelines regarding differences in operating
parameters for HWIs, boilers, cement kilns, and LWAKs.
5.1
ORGANIC EMISSIONS FROM HWIS AND BOILERS
The generation of organic PICs from HWIs and boilers is generally associated with poor combustion
(conditions related to time, temperature, and turbulence). Berger and others (1996) documented the effects
of inefficient burner operation and oxygen control in the PCC and SCC of HWIs. These conditions led to
incomplete combustion and subsequent increases in flyash and carbon monoxide and total hydrocarbon
concentrations. In general, adequate design and operation of the combustion zone operating parameters
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
52
-------
ensure the availability of excess oxygen and destruction of combustion gases in the SCC. Historically,
levels of oxygen, carbon monoxide, and total hydrocarbons have been used as surrogates of good
combustion to minimize PIC emissions.
Mixing (turbulence) with oxygen also is critical in order to minimize production of fly ash and PM, which
can be correlated to PIC production (Berger and others 1996). Different organic wastes will vary in their
tendency to form flyash and PM and PICs. Other examples of operating conditions that could lead to
increased PIC formation are as follows:
Excess flue gas velocity or flow leading to shorter residence times in the combustion unit
and incomplete combustion
Batch-fed containers with high oxygen demand that can lead to transient conditions of low
oxygen and incomplete combustion
Feeding of low-Btu solids or aqueous liquids that can reduce combustion temperatures
Insufficient temperatures in either the PCC or SCC leading to incomplete combustion
These types of operating issues typically are addressed during performance testing under DRE conditions.
Historically, DRE tests have included operating conditions that result in minimum combustion
temperatures in the PCC and SCC, maximum carbon monoxide emissions, maximum flue gas velocity or
flow, and maximum feed rate of each waste type (EPA 1989, 1997f). Therefore, a successful DRE test
should result in the maximum organic PIC emissions for HWIs and boilers.
The operating parameters associated with organic emission testing for HWIs and boilers are the same as
those identified in Table 4-1. These parameters include:
Combustion temperature
Combustion gas velocity
Waste feed rates
Waste composition
Carbon monoxide and total hydrocarbon levels
These parameters are usually demonstrated in conjunction with the DRE test condition; however, there
may be situations where collection of non-D/F organic emissions data would occur during a separate risk
burn test at normal operating conditions. For example, the use of surrogate waste for the DRE test might
conflict with the objective to use actual wastes during the non-D/F organic testing. Facilities may request
that the permitting authority allow testing at normal conditions provided that the facility can establish a
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 53
-------
monitoring plan and is willing to accept permit conditions to ensure that the test conditions are
representative of long-term operations. In all cases, the permit writer has a responsibility to ensure that
permit condition based on the non-D/F organic testing will be protective. Therefore, a permit writer's
decision to approve testing during normal (or average) conditions will depend on the extent to which those
conditions represent potential emissions and risks over the permitted operating range. If organic emissions
data are collected under normal conditions, caution should be exercised with respect to operating
parameters where the average value is significantly different than operating extremes.
Waste selection for PIC testing at HWIs and boilers can be very important, because the types of organic
PICs may relate directly to the chemical composition of the waste. Actual wastes, and not surrogate wastes
synthesized from pure compounds, should be used whenever possible. However, if a facility burns a large
number of small waste streams, or if large quantities of uniform wastes are not available, then the
inventory of actual wastes might be insufficient for completing the non-D/F organics emissions testing.
For example, a commercial facility burning containerized wastes may only have a few drums from each
generator. A combination of real waste and simulated wastes should be used in this situation. In the case
of containerized wastes, it will generally be more important to focus on maximizing the amount of volatile
material within each container, as discussed in Section 4.1.3.2, than to use a real waste with less oxygen
demand.
When choosing actual wastes, the emphasis should be on high quantity, routine, and recurring waste
streams. Examples of historical information that can assist in a review of actual wastes include the
following:
A listing of the top percentage (e.g., top 75 percent) waste receipts by quantity for a given
period with descriptions of:
source or process generating the waste
annual weight treated
hazardous waste code(s)
physical description
chemical description
feed location into the HWC unit
A ranking by weight of the top percentage of chemical constituents treated over a given
period
A ranking which identifies wastes containing constituents with the highest toxicity,
persistence, and bioaccumulation potential
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
54
-------
A summary of minimum, maximum, and average values for each feed location into the
HWC unit over a given period for:
heating value
percentage total chlorine
percent ash
percent water
percent solids
viscosity (as appropriate)
percent sulfur
concentrations for each metal of concern
Candidate wastes for organic emissions testing should be recommended and justified considering the
ability of the waste to achieve the desired combustion conditions (i.e., tax the combustion unit); the ability
of the waste to represent toxic, persistent, and bioaccumulative compounds; and predicted availability of
sufficient quantities of the waste streams to complete the relevant portions of the trial burn. Efforts should
be made to schedule the trial bum for a time period when the target wastes will be available. However,
waste selection may need to be evaluated a few weeks prior to the trial burn based on actual waste
stockpiles or scheduled waste receipts.
The facility should also provide protocols for monitoring and tracking waste stream information on an
ongoing basis. The protocols should include methods and procedures for comparing future waste
characteristics to the wastes burned during the organic emissions testing. This information may be used to
determine the potential need for retesting.
5.2
ORGANIC EMISSIONS FROM CEMENT KILNS AND LWAKS
Sections 4.5 and 4.6 describe how conditions related to good combustion are less relevant for cement kilns
and LWAKs than for HWIs and boilers. This also applies to emissions of organics other than D/Fs. The
design basis of a cement kiln rarely leads to poor combustion. However, naturally occurring organics are
driven out of the raw materials at the cold end of the kiln.
The operating parameters associated with organic emissions from cement kilns and LWAKs are the same
as those identified in Table 4-2. Conditions relative to good combustion are applicable only to cement
kilns and LWAKs that feed wastes at locations other then the hot end of the kiln. Conditions relative to
raw material organics are most likely to be achieved during a high temperature or SRE test.
Selection of wastes burned at the hot end of a cement kiln or a LWAK is expected to be less complex than
for HWIs and boilers, because wastes are blended to meet fuel specifications. However, the waste
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
55
-------
characterization information discussed in Section 5.1 should be provided for the as-blended fuel to the
extent possible, as well as for wastes fed at locations other than the hot end of the kiln. For blended fuel, it
may not be possible to perform a characterization by top percentage of waste receipts. In this case, an
alternate ranking scheme would need to be developed to reflect the most prevalent waste codes and toxic,
persistent, and bioaccumulative constituents in the blended fuel. The information on minimum, maximum
and average physical and chemical characteristics would be provided for the as-blended fuel.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
56
-------
6.0 METAL EMISSIONS
Metal emissions from HWC facilities are primarily dependent on metal volatility and APCD type, rather
than type of HWC facility. Therefore, this subsection focuses primarily on volatility groupings and
APCDs and associated PM controls. Metals are currently regulated under EPA's BIF rule (40 CFR 266
Subpart H) and under omnibus authority of Section 3005(cj(3) of RCRA and 40 CFR 270.32(b)(2).
Additional information on approaches for establishing permit limits for metals is found in EPA (1992b,
1997f, 1997g, 1997h, 1997i, and 1997J). In general, HRA metal feed rate limits are established in RCRA
Part B permits based on either a BIF Tier I approach, or high temperature (SRE) test results. The BIF Tier
I approach assumes no removal of metals. The SRE test involves quantifying metal contents of waste
feeds, spiking with specific metals during SRE testing, stack gas sampling and analysis, and determining
risk-based limits based on potential human health risks through the direct exposure (inhalation) pathway.
The 10 BIF metals originally listed in 40 CFR 266 Subpart H include antimony, arsenic, barium, beryllium,
cadmium, chromium, lead, mercury, silver, and thallium. Nickel and selenium were later regulated through
omnibus authority of RCRA (EPA 1992a). While antimony, arsenic, and selenium typically exhibit only
weak metallic properties and may commonly be referred to as metalloids, they are defined here as metals
for simplicity.
The traditional approach of establishing HRA feed rate limits based on either a BIF Tier I approach or
from an SRE test is required for BIFs and is recommended for most HWIs. However, the emission rates
associated with maximum SRE or Tier I metal feed rates could exceed target risk levels when evaluated in
multi-pathway SSRAs. Because the SSRAs include direct and indirect exposure routes for all
contaminants, it is expected that metal emission limits (in the final RCRA permits) will be lower than those
based on BIF air modeling approaches and the direct inhalation pathway. Lower limits are especially
likely for mercury, because of its propensity to bioaccumulate, and for nickel, because it will now be
evaluated as a potential carcinogen through the direct inhalation exposure route (EPA 1998a, in press a).
A facility has two options for achieving the target risk levels using feed rate controls if the emission rates
corresponding to a BIF Tier I approach or from an SRE test are likely to exceed target risk levels. First,
the facility should determine how far the emission and feed rates should be lowered to meet target risk
levels. In some cases, the SRE test may be expanded to include more metals, or spiking of wastes-with
specific metals may no longer be appropriate (in general, spiking with mercury is discouraged). Next, the
facility should evaluate whether compliance with the reduced feed rate limits can be achieved on an HRA
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
57
-------
basis. If so, the net result of the SSRA would simply be to lower the HRA metal feed rate permit limits
from those established using BIF procedures.
If lower HRA feed rate limits will not allow sufficient operating flexibility, the facility may opt to conduct
a separate risk burn at normal metal feed rates to establish long-term feed rate limits, in addition to the
traditional HRA limits. Facilities that wish to use metals emissions data collected during normal operating
conditions in the SSRA will be subject to a dual testing and permitting scheme. Permit conditions will be
based on the following tests:
A SRE test, to establish HRA limits and to demonstrate that maximum emissions meet BIF
emission limits (or risk-based emission limits considering only inhalation for HWIs)
A normal test, to establish long-term average operating limits and to generate emissions
data for evaluation in the multi-pathway SSRAs
The short-term HRA limits are required for BIFs, and are necessary for HWIs in order to alleviate concerns
regarding short-term, high emission spikes that would not be controlled by long-term limits. Limiting the
potential for short-term spikes based on an inhalation pathway evaluation is a conservative and reasonable
approach (in the absence of acute health benchmarks) since indirect effects and bioaccumulation are not
factors for the inhalation pathway. Additional information on acute inhalation risks is found in EPA (in
press a). The long-term limits are used to assure that operating conditions evaluated in the SSRA remain
representative.
In addition to the original 10 BIF metals, metals that will be considered as COPCs for the SSRA will
include aluminum, cobalt, copper, manganese, nickel, selenium, vanadium, and zinc (EPA in press a, in
press b). Summaries of specific human health and ecological concerns related to these metals are found in
Section 2.3 of this document. With the possible exception of nickel (which is considered as a potential
carcinogen) it is generally anticipated that feed rates of these non-BIF metals will not need to be limited on
an HRA basis. In addition, testing for these metals will not typically involve spiking. However, tracking
could be required to monitor potentially significant changes in feed rates that would warrant re-evaluation
in the SSRA, or long-term average feed rate limits may be established.
Specific operating parameters related to control of metal emission rates are expected to be consistent with
operating parameters previously identified by EPA (1989, 1992b, 1994a, 1997f, 1997g, 1997h, 1997i, and
1997J). Most of these operating parameters are summarized in Section 6.3 of this document. Technical
support documents related to EPA's proposed MACT standards (EPA 1996a) provide additional
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
58
-------
information on control technologies and effectiveness of APCDs (EPA 1996b, 1996c, 1996d, 1996e).
Thus, a significant body of knowledge exists on metals and PM control, operation of APCDs, and specific
APCD operating parameters. This document relies on this accumulated information.
Overall, metal behavior (and potential risks) in HWC facilities is influenced by issues related to metal
speciation, metal volatility, and system operating conditions (EPA 1992b, 1996a). These issues include (1)
higher volatility of metal chlorides than metal oxides, (2) thermodynamic considerations related to
temperature profiles, (3) metal valence and phase, (4) removal efficiencies for specific types of APCDs,
and (5) system-specific operating parameters such as cement kiln dust recycling rate (EPA 1996a).
Additional information on metal speciation, volatility, system operating conditions, and control measures
are discussed below.
6.1
METAL SPECIATION
Speciation is important for the volatile metal mercury and the low-volatile metals chromium and nickel.
Currently, EPA's human health and ecological risk assessment protocols (EPA in press a, in press b) do not
rely on analytical methods for quantifying metals speciation. The model associated with the protocols
includes default assumptions to partition the various forms of the elements. Based on the conservative
nature of these assumptions, a facility may want to perform speciation sampling or present other
information to replace the default assumptions with site-specific data.
Mercury can be present in the divalent (oxidized) form or the elemental (reduced) form. The divalent form
is water soluble and more likely to adsorb to particles. Stack emissions can thus be removed from the
atmosphere by precipitation which can lead to partitioning into methyl mercury in nearby surface waters.
This will tend to drive indirect ecological risks and food chain risks to humans. Divalent mercury emission
can be reduced through the use of a wet scrubber or other APCD. Elemental mercury emissions are more
difficult to control, and are more likely to become part of the global mercury cycle. Mercury speciation is
discussed in greater detail in EPA (in press a). EPA (in press a) assumes that mercury is partitioned as 80
percent vapor phase, and 20 percent particle-bound. The 80 percent vapor phase is segregated into 20
percent elemental and 60 percent divalent. Of the 20 percent particle-bound, all of it is assumed to be
divalent. Potential sampling methods for mercury speciation are discussed in Appendix B of this
document.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
59
-------
Chromium exists in multiple valance states. The trivalent form has relatively low toxicity based on
indirect exposure pathways, while the hexavalent form is carcinogenic to humans via the direct inhalation
route. The hexavalent form typically occurs as either chromate or dichromate. Other chromium valence
states are not considered in the SSRA. EPA (in press a) initially assumes all chromium to be emitted in the
hexavalent (carcinogenic) form. However, the risk model can be adjusted to show indirect effects and
ecological uptake through the trivalent form. Laboratory work by Linak and others (1996) and Linak and
Wendt (in press) indicates that the fraction of hexavalent chromium emitted may be very small, typically
ranging from less than one percent to approximately two percent of the total chromium measured. The
hexavalent fraction in the exhaust was found to be enhanced somewhat by high levels of chlorine, but was
reduced to below analytical detection limits by the addition of small quantities of sulfur. EPA (1996g)
provides an in-stack emissions method (Method 0061) for differentiating between trivalent and hexavalent
chromium.
Nickel can achieve several oxidation states (up to four), with the most common being +2. The most
prevalent forms of nickel are sulfides, oxides, chlorides, and silicates. Some forms of nickel (including
nickel carbonyl, nickel subsulfide, and nickel refinery dust) are considered as potential human carcinogens.
Currently, EPA (in press a, 1998a) recognizes nickel emissions from HWC facilities as a potential human
carcinogen via the direct inhalation exposure pathway based on data collected on nickel refinery dust.
Overall, this assumption is fairly conservative because emissions of carcinogenic forms of nickel by HWC
facilities is unproven. Based on this, a facility may present data indicating the absence of carcinogenic
nickel refinery dust components or the presence of noncarcinogenic species such as soluble salts. Standard
EPA sampling methods for nickel speciation do not exist at this time. For exposure pathways other than
direct inhalation, nickel is assumed to be in the noncarcinogenic soluble form.
6.2
METAL VOLATILITY GROUPINGS
Metals can be grouped as volatile, semivolatile, or low volatile. Volatile metals are defined by EPA
(1992b, 1996a, 1996b, 1996c, 1996d, 1996e) and Energy and Environmental Research Corporation (EER)
(1996a, 1996b) as metals that have high vapor pressures (and are typically in vapor form) in the
combustion chamber as well as in the cooler downstream APCD of HWC facilities. Because of variable
removal efficiencies in the APCD, emissions of volatile metals are highly dependent on feed rate.
Semivolatile metals typically have higher vapor pressures at combustion temperatures and lower vapor
pressures at APCD temperature (EER 1996a, 1996b). This leads to vaporization in the combustion
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
60
-------
chamber followed by condensation onto participates (often of less than a 1-micron diameter) before
entering the APCD (EPA 1996a, 1996c, 1996d). Emissions of semivolatile metals are a function of both
feed rate and APCD removal efficiency (EPA 1996a).
Low-volatile metals vaporize at a lesser extent at combustion temperatures and partition to a greater extent
to bottom ash, other residue, cement kiln clinker, or entrained flue gas PM (EPA 1996a, 1996b, 1996c).
Low-volatile metal emissions are more strongly related to the operation of the APCD than to feed rate.
Evaluations conducted by EER (1996a, 1996b) indicate that SRE of metals increases strongly with
increasing feed rate for similar types of facilities, and that this relationship is stronger for semivolatile as
compared to low-volatile metals.
Figure 6-1, reproduced from Clark and Sloss (1992), provides a representation of metals volatility groups.
Group 1 represents the low-volatile metals, Group 2 represents the semivolatile metals, and Group 3
represents the volatile metals. The overlapping circles indicate that partitioning behavior can vary for
different combustion systems with different operating conditions.
Aluminum and silver were not categorized. However, based on vapor pressures and information found in
Dellinger (1993) aluminum would be expected to be low volatile as an oxide. As a chloride, aluminum
could be semivolatile to volatile. Based on the species expected to be formed within combustion
environments, silver would likely be classified as low volatile.
It is important to note that definitions of volatile, semivolatile, and low-volatile metals will vary among
studies, test conditions, and type of HWC facility (EPA 1996a, 1996b, 1996c 1996d, 1996e; EER 1991,
1996a, 1996b; Cesmebasi and others 1991; Linak and Wendt in press). Moreover, metal volatility is
affected by design differences in the HWC unit (such as differences among kilns, liquid injection, and
controlled air designs), APCD PM removal efficiency, raw materials (for cement kilns and LWAKs), the
presence of chlorine (which increases metal volatility), and fuel (natural gas, fuel oil, or coal). The figure
reproduced from Clark and Sloss (1992) may not fully represent the behavior of certain metals in cement
kilns. For example, EER (1996a, 1996b) found that the metals antimony and arsenic, exhibited low
volatility in cement kilns (probably due to their affinity for the cement matrix).
Specific issues related to volatile, semivolatile, and low-volatile metals and their expected behavior in
HWIs, cement kilns, and LWAKs are described in the following sections. These discussions are provided
in order to establish a general framework regarding metals behavior. However, this guidance does not
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
61
-------
recommend that permit limits be established based upon volatility groups (rather than individual metals).
The divisions between volatility groups are not absolute, and metals must be modeled individually in the
SSRA. In addition, the following discussions should not be construed as superseding current BE7
regulatory requirements for metals. Deviations from the traditional BIF approach may only be considered
in the context of permitting conducted pursuant to EPA's omnibus authority (e.g., SRE tests for HWIs, or
metals tests conducted at normal conditions for SSRAs).
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 62
-------
FIGURE 6-1
METAL VOLATILITY GROUPS
Group 3
Group 2
Group 1
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
63
-------
6.2.1
Volatile Metals (Mercury and Selenium)
EER (I996a, 1996b) and Clarke and Sloss (1992) identify mercury and selenium as volatile metals. These
metals tend to vaporize completely at combustion temperatures. Mercury may remain volatile in the
APCD, while selenium is more likely to condense out. Further considerations related to mercury are
provided below.
EPA (1996a, 1997e) and the U.S. Congress consider mercury to be a high priority hazardous air pollutant
with the potential to cause significant human health and environmental effects. Mercury is the most
volatile regulated metal, is the most difficult metal to control, and has been the subject of many detailed
studies, including the "Mercury Study Report to Congress" (EPA 1997e). At temperatures found in
RCRA-permitted HWC facilities, nearly all mercury volatilizes to form gaseous mercury that includes both
elemental (reduced) and divalent (oxidized) forms (EPA in press a). The divalent form is more likely to
adsorb to PM and to be removed in the facility's APCD. EPA (in press a) has found that partitioning
between elemental and divalent mercury during combustion is dependent on fuel and physical
characteristics of the source. For combustion sources containing relatively high concentrations of mercury
and chlorine, mercuric chloride is expected to be the dominant form of mercury (EPA 1996a).
Operating parameters related to mercury emissions from HWC facilities include the following (EPA
1996d):
Mercury emissions increase with increasing mercury feedrate. Permit limits should be
established for total maximum mercury feedrate (including hazardous waste, raw
materials, and fossil fuels). No separate limit is recommended for pumpable hazardous
wastes, since mercury is highly volatile in any form.
Chlorine feedrate may affect mercury emissions when wet scrubbers are used as APCDs.
In the case of mercury, the presence of chlorine is expected to decrease emissions through
conversion to the soluble divalent species. However, because only de-minimis levels of
chlorine are necessary for mercury conversion to the soluble salt, minimum chlorine limits
are not recommended. A maximum chlorine limit is also not necessary, since mercury is
already volatile.
Combustion chamber temperature is not believed to be a critical operating parameter. At
typical mercury feed rates and combustion temperatures, all mercury vaporizes in the
combustion chamber. However, experimental results indicate that conversion of elemental
mercury to mercuric chloride increases with temperature up to approximately 1,600 ฐF
(EPA 1996d).
Wet scrubbers have been demonstrated to be effective at controlling water soluble forms
of mercury. Operating parameters associated with wet scrubbers (such as pressure drops,
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
64
-------
liquid feed pressure, pH, and ratio of liquid to flue gas) should be established as permit
conditions. Caspar and others (1997) observed that rapid quenching of hot flue gases from
municipal waste incinerators reduced mercury removal in wet scrubbers. They believed
the lower temperatures affect the equilibrium reactions and shift equilibrium towards
elemental mercury. Similar observations have been made by EPA (1997d) at several HWC
facilities.
Mercury emissions from HWIs are currently controlled by limiting waste feed rates and using wet
scrubbers designed for acid gas removal. As described above, wet scrubbers are effective in removing
water soluble forms of mercury including mercuric chloride. EPA (1996a) also states that additional
alternative controls including carbon injection and carbon bed technologies would increase mercury
removal from HWIs up to 90 to 99 percent (EPA 1996d).
Cement kilns are currently regulated by BIF requirements, and they control mercury emissions by limiting
mercury feedrate. Raw materials and coal (as an auxiliary fuel) can add to the mercury emissions. EPA
(1996a) states that additional removal may be achieved by quenching the flue gases, followed by treatment
with carbon injection or carbon beds. Mercury can exit the kiln as volatile emissions, or partition to the
clinker product or cement kiln dust. According to EPA (1996a), all existing cement kilns rely on dry
APCDs that are either fabric filters or ESPs. Mercury is volatile at the operating temperature of these
units, and removal efficiencies are highly variable, ranging from zero to 90 percent (EPA 1996a).
Similar to cement kilns, LWAKs are currently regulated by BIF requirements, and mercury emissions are
controlled by limiting mercury feedrate. According to EPA (1996a), all LWAKs rely on fabric filters to
control PM emissions, although one facility uses a spray dryer, venturi scrubber, and wet scrubber in
addition to the fabric filter. The range of mercury emission rates is similar to that observed for cement
kilns. The raw material used by LWAKs seems to contribute to the total emissions rate. EPA (1996a)
states that additional removal may be achieved by quenching the flue gases, followed by treatment with
carbon injection or carbon beds.
6.2.2
Semivolatile Metals
Several operating parameters affect semivolatile metal emissions from HWC facilities. A summary of
operating parameters for the semivolatile metals based on EPA (1992b, 1996a, 1996d) follows:
Semivolatile metal emission rates increase as the metal feedrate increases. Permit limits
should be established for total maximum feedrates (including hazardous waste, raw
materials, and fossil fuels). A separate limit for pumpable feedstreams is not necessary
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
65
-------
because partitioning between the combustion flue gas and bottom ash or product does not
appear to be affected by physical state.
Chlorine increases metal volatility. An operating limit on maximum chlorine fed to the
HWC unit is recommended. The limit is based on total chlorine from all sources,
including organic and inorganic chlorine sources.
Maximum combustion chamber temperature traditionally is established under the BIF rule
to control metal volatility. However, maximum combustion chamber temperature is less
important than other operating parameters when considering metal emissions from most
HWC facilities. For semivolatile metals, typical combustion temperatures are generally
high enough to volatilize all of the metals in the combustion chamber.
APCD type and operating parameters are recognized as critical to the control of
semivolatile metals, and limits on APCD parameters should be established as permit
conditions. In addition, the operating temperature of the APCD may also be important. At
higher temperatures, a larger portion of some metals will be in the vapor phase. The
establishment of a maximum limit on the inlet temperature to the dry APCD is
recommended to control emissions from semivolatile metals.
Cement kilns and LWAKs that recycle collected PM require special consideration to
ensure that emissions reach steady-state prior to testing, as described in EPA (1992b).
Semivolatile metals are volatile at the high temperatures found within the combustion chamber, but
typically condense onto the fine particulates within a dry APCD such as a baghouse or ESP. Control of
semivolatile metals is most directly associated with PM control. However, these metals tend to condense
onto fine PM (less than 1 micron) (EPA 1996b), which is controlled less effectively than larger PM.
Therefore, removal efficiencies of particulates associated with semivolatile metals are lower than removal
efficiencies for total PM. EPA (1996a) also observed that venturi and wet scrubbers are effective in
controlling emissions of semivolatile metals. However, venturi scrubbers are generally considered most
effective for larger particulates (greater than 1 micron diameter). Wet scrubbers are designed for acid gas
removal and are also only effective in removing large particulates. Thus, fabric filters and ESPs would be
considered the most effective in removing smaller particulates (Clarke and Sloss 1992; EER 1991).
6.2.3
Low-Volatile Metals
Operating parameters associated with the low-volatile metals are consistent with those described for the
semivolatile metals in Section 6.2.2 (EPA 1996d), with the following exceptions:
A separate limit for low-volatile metals in pumpable feed streams is recommended, since
these may partition at a higher rate to the combustion flue gas.
Low-volatile metals are less apt to vaporize completely at typical combustion
temperatures. Thus, a limit on maximum combustion temperature may be more important
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
66
-------
for low-volatile metals than for semivolatile metals. However, maximum combustion
chamber temperature is most likely less important than APCD operating parameters. The
amount of additional vaporization at slightly higher temperatures could be negligible
compared to the amount of metals contained in entrained flue gas PM, especially for kilns
and pulverized coal boilers.
Similar to the semivolatile metals, EPA (1996a) observed that operation of the APCD is critical to removal
of low-volatile metals. Also, some low-volatile metals can behave as semivolatile metals in the presence of
chlorine or under reducing conditions. Venturi and wet scrubbers, ESPs, and fabric filters were all found
to be effective in controlling low-volatile metals. However, fabric filters and ESPs would be considered
the most effective in removing smaller particulates.
6.3
OPERATING CONDITIONS AND PARAMETERS FOR METALS
In general, metals emissions data are collected during high temperature (SRE) conditions as explained in
the BIF rule (40 CFR 266 Subpart H) and EPA (1992b, 1997f, 1997g, 1997h, 1997i, 1997J). As
appropriate, metal emissions data for SSRAs may also be collected during normal operating conditions, as
explained in Section 3.0 of this document. Operating parameters related to collection of metals emission
data are summarized in the documents cited above, as well as in Appendix A of this document. RCRA
permit limits for metal feed rates will generally be based on HRAs for BIF metals and, as appropriate,
long-term averages for any additional metals.
Operating parameters related to high temperature testing of BIF metals have traditionally included:
Maximum combustion temperatures
Maximum flue gas flow rate or velocity, or production rate for cement kilns and LWAKs
Maximum metal feed rates (total feed rates, total hazardous waste, and total pumpable
hazardous waste)
Maximum total chlorine feed rates
As explained previously, metal emissions are dependent on operation of APCDs and associated PM
controls. Specific operating parameters related to APCDs with wet scrubbers include:
Minimum pressure drops across the scrubber
Minimum liquid feed pressure
,ป Minimum liquid pH
Minimum liquid/flue gas ratio
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
67
-------
Maximum scrubber blowdown or maximum suspended solids
Specific operating parameters related to APCDs with dry systems include:
Maximum inlet temperature for fabric filters and ESPs
Minimum pressure drop across fabric filters
Minimum power input to ESPs
Minimum caustic feed rate to dry scrubbers
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 68
-------
7.0 PARTICLE-SIZE DISTRIBUTION AND HYDROGEN CHLORIDE
AND CHLORINE EMISSIONS
This section addresses data collection requirements and operating parameters for particle-size distribution,
HC1, and C12. Particle-size distribution is used in the air dispersion modeling completed as part of the
SSRAs, as described by EPA (in press a, in press b). HC1 and C12 emissions are relevant to the direct
inhalation pathway of the SSRA.
7.1
RELATIONSHIP BETWEEN PARTICLE SIZE AND POTENTIAL EMISSIONS
Quantification of indirect human health and ecological risks in the SSRAs is dependent upon annual
deposition rates at actual and reasonable future exposure scenario locations (EPA in press a). In general,
most semivolatile and low-volatile metals and some semivolatile and low-volatile organics occur only in
the particle phase (mercury is partitioned between vapor and particulate phases by the model associated
with the human health and ecological SSRAs). Particle size is the primary variable associated with the
transport of emitted particles (this includes wet and dry removal processes and subsequent deposition).
Particle terminal velocity is based on particle diameter and density.
Particle-size distribution information is necessary in order to perform air dispersion and deposition
modeling for the SSRA. Required inputs to the model include (1) particle density, (2) mass distribution by
particle-size category, and (3) surface area distribution by particle-size category. In an air model
sensitivity analysis (The Air Group 1997), the particle-size distribution element was found to be a
"moderately" sensitive parameter on a scale of "none" to "severe" sensitivity. Moderate sensitivity meant
that site-specific data could result in variations in the model outputs of up to 50 percent from default
assumptions.
A particle-size distribution that is more heavily weighted towards larger particles will result in higher
deposition near the source, and reduced concentration and deposition further away from the source. A
particle-size distribution which is more heavily weighted towards smaller particles will decrease deposition
near the facility and increase concentrations and deposition away from the source. In general (but not
necessarily in every case) higher deposition near the source is expected to result in higher potential risks.
Receptor locations relative to maximum deposition are a critical factor when quantifying potential risks.
In most cases, the effectiveness (and associated removal efficiency) of the APCD is the primary
determinant of particle size and total particle mass emitted from HWC facilities. Advances in air pollution
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
69
-------
control technology have led to particulate removal and removal efficiency improvements. However,
removal efficiencies are typically expected to decrease as particle size decreases. For example, according
to EPA (1996a), cyclone separators have typical removal efficiencies of less than 20 percent for particles
less than 1 micron in diameter, and about 5 percent for particles less than 0.5 microns in diameter. Fabric
filters are extremely efficient and have reported removal efficiencies of 99 to 99.99 percent for particles as
low as 0.1 micron diameter. ESPs are less efficient at capturing particles in the 0.1 to 1.0 micron range and
have reported removal efficiencies of 90 to 95 percent (EPA 1996b). Based on these observations, a well-
operated APCD greatly decreases the potential mass and mean particle size emitted from HWC facilities.
7.2
MEASURING PARTICLE-SIZE DISTRIBUTION
EPA (in press a) provides a nine-category particle-size default assumption which represents the particle-
size categories emitted from sources equipped with ESPs or fabric filters. The default particle-size
diameters range from less than 0.7 to greater than 15.0 microns. However, recent measurements of particle
sizes by HWC facilities for the purpose of conducting SSRAs have shown that the default distributions
may be overly biased towards larger particles. Much of the actual monitoring data are showing a majority
of the particle mass at less than 1 or 2 microns. In addition, the default assumptions would not be
appropriate for facilities equipped with wet APCDs, or facilities with no APCD. Thus, site-specific
particle-size measurements are recommended for each HWC facility.
Particle-size distribution may be determined using devices such as cascade impactors or by analysis of PM
deposited on filters using scanning electron microscopy (EPA 1998b). Preliminary data collected and
analyzed using these methods (EPA 1998b) indicate a large amount of uncertainty, especially for particle
sizes less than 1 micron. EPA (1997b) recognizes the inherent problems associated with measuring very
low masses of particulates and recommends modified Method 5 (Method 51) to improve accuracy,
precision, and representativeness by significantly reducing the variability and potential errors. Method 51
includes improved sample collection, elimination of possible contamination, and improved sample analysis
(EPA 1997b). Although Method 51 was not written specifically for particle-size determinations, it is
possible that use of 51 procedures in conjunction with the cascade impactors for particle-size
determinations could help reduce analytical uncertainty. A facility equipped with fabric filters or ESPs
may also be able to conduct preliminary modeling to demonstrate a reduced need for site-specific particle-
size information.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
70
-------
EPA recognizes the inherent uncertainties associated with collecting particle-size distribution data, given
(1) the removal efficiencies of APCDs, (2) the small particle sizes emitted, and (3) the practical difficulties
associated with measuring these particles using currently available techniques. Permit writers should work
with facilities to obtain the best data available under the circumstances. Additional guidance regarding
particle-size determinations is provided in Appendix B.
7.3
OPERATING CONDITIONS AND PARAMETERS FOR COLLECTION OF
PARTICLE-SIZE DISTRD3UTION DATA
EPA recommends that particle-size data be collected under normal or other operating conditions (as
defined by the facility) where ash spiking is not performed. Ash spiking will bias the particle-size
distribution and results would not be representative of actual combustion conditions.
7.4
HYDROGEN CHLORIDE AND CHLORINE EMISSIONS
HC1 limits for HWIs are found in 40 CFR Part 264.343. Appendix IV of 40 CFR Part 266 provides direct
inhalation RACs for total C12 and HC1 emissions from BIFs (including cement kilns, LWAKs, and boilers).
Since potential risks from HC1 and C12 are limited to the inhalation pathway, completion of multi-pathway
SSRAs is not expected to affect the current regulatory approach for HC1 and C12 emissions from BIFs.
However, it is now recommended that HWIs test their stack emissions for both HC1 and C12. Risk-based
limits may be established in the final permit for HWIs in lieu of the technology-based emissions limits
found in 40 CFR 264.343 if the risk-based limits are lower.
Continuous emissions monitoring systems (CEMS) for HC1 are available; however, HC1 and C12 are
controlled typically by limiting the feed rate of total chlorine in all feed streams and by establishing
appropriate APCD operating parameters for wet and dry scrubbers as listed previously in Section 6.3.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
71
-------
8.0 DATA ANALYSIS AND PERMIT CONDITIONS
As explained in Section 1.0 of this document, additional RCRA permit conditions may be established
based on the collection of emissions data (including D/Fs, organics other than D/Fs, metals, and HC1 and
C12) from HWC facilities and subsequent evaluation of these data in SSRAs. These data will be collected
during the appropriate test condition (as described in Section 3.0), and by relying on the appropriate
operating parameters (as discussed in Sections 4.0 through 6.0) for HWIs, boilers, cement kilns, and
LWAKs. The data are evaluated in the SSRAs to provide reasonable maximum exposure (RME) estimates
of potential risks to human and ecological receptors (EPA in press a, in press b). Final RCRA permit
conditions will ultimately depend on the operating practices, emissions levels, and SSRA results specific to
each facility. This section identifies information that should be included in the trial burn plan to clearly
communicate the relationship between the planned test protocol and potential permit conditions. In
addition, this section describes how trial burn data should be consolidated for evaluation in the SSRA.
Further examples regarding the relationship between test protocols and permit conditions are provided in
Appendix A of this document.
Consistent with existing EPA trial burn guidance (EPA 1989,1992b, 1997f, 1997g, 1997h, 1997i, and
1997J) it is assumed that three operating runs are conducted at each test condition, and that trial burn
testing frequently includes three test conditions (DRE, SRE, and normal test conditions). In certain cases,
additional test conditions may be required for a facility to develop the appropriate operating envelopes for
a variety of operating scenarios. This could be the case for facilities that burn wastes from different
production runs in discrete campaigns, facilities that cannot simultaneously maximize all of their feeds
during one test condition, or facilities that need to perform multiple tests to resolve conflicting operating
parameters.
Conflicting parameters are key operating parameters that cannot be demonstrated simultaneously (e.g.,
maximum combustion gas velocity and minimum fabric filter pressure differential). A facility may
encounter conflicts when attempting to integrate the key operating parameters for risk testing (identified in
Sections 4.0 through 6.0) with the key operating parameters for traditional DRE and SRE testing. To
overcome a conflict, a facility may need to perform duplicate tests as follows:
A first set of operating conditions to set limits for all operating parameters, excluding the
ones in conflict.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
72
-------
8.1
Additional operating condition(s) to set limits on the conflicting operating parameters. To
the maximum extent practicable, only the conflicting parameters should be varied from the
first set of operating conditions. All nonconflicting parameters should be maintained as
constant as possible during all operating conditions.
PERMIT LIMITS FOR KEY OPERATING PARAMETERS
The key operating parameters that could be limited in a facility's RCRA permit based on collection of
SSRA data are identified in Sections 4.0 through 6.0. Many of these parameters are assigned primary,
secondary, and tertiary hierarchial designations to indicate their relative significance. These designations,
together with consideration of site-specific operating practices and final SSRA results, should all be
considered by the permit writer in determining which operating parameters require final permit limits.
Also, depending on the significance of a specific operating parameter for the facility, a final permit limit
might be quantitative (such as an HRA limit), or it could involve periodic monitoring and reporting (for
permit limits based on normal operating conditions).
Compliance averaging periods for key operating parameters should also be considered. Frequently, permit
limits are established using HRA averaging periods. However, EPA (1996a) states that shorter averaging
periods (such as instantaneous or 10-minute rolling averages) may be appropriate in situations where short-
term perturbations outside of a certain operating range could result in high emission rates that cannot be
offset by lower emission rates during periods of more normal operations. The short-term average limit,
used to control perturbations, would be used in conjunction with an HRA limit to control average
emissions. For the operating parameters in Tables 4-1 and 4-2, EPA (1996a) suggests that short-term limits
are appropriate for minimum combustion temperatures and maximum post-combustion temperatures.
Alternatively, EPA (1996a) acknowledges that averaging periods longer than an hour may be appropriate
for situations where control of longer-term average emissions is the only concern. Examples include
metals and chlorine feed rate limits. Long-term averages may be considered for any operating parameter
where short-term perturbations are not a concern, and where an averaging period is not already prescribed
by regulation. Control charts, similar to those used for industrial quality control, provide an alternate
means of assessing ongoing operations to ensure that the SSRA remains representative over the long term.
For example, the control charts could be used to ensure that long-term variation does not exceed one or two
standard deviations from the mean. EPA (1996a, 1997J) provide further information on averaging periods
and control charts.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
73
-------
Early communication and coordination between the permit writer and facility is essential, because an
understanding of how the final permit limits will be developed is integral to the design of the test protocol.
The facility owner/operator is ultimately responsible for assuring that the trial burn provides adequate data
to support permit conditions that are acceptable to the facility. The following recommendations identify
specific types of information that should be provided by the facility to facilitate this process:
A summary matrix (for each test condition) should be provided listing the planned
emissions data to be collected and proposed operating limit for each relevant operating
parameter. Operating parameters include those listed in Sections 4.0 through 6.0, as well
as the operating parameters that must be demonstrated during the traditional DRE, SRE,
PM, and HC1 and C12 demonstrations.
A list of conflicting parameters (key operating parameters that cannot be maximized or
minimized simultaneously) and a detailed explanation of the reasons for the conflict
should be provided. This should include a summary to explain the two or more test
conditions that will be performed to resolve the conflict, and to identify changes in other
operating parameters that may be necessary to resolve the conflict.
A complete list of operating and waste feed parameters which are expected to be limited in
the permit should be provided. This list should be consistent with the proposed test
conditions and the following information should be included for each waste feed and
operating parameter:
The anticipated permit limit, assuming that all test conditions are executed
as planned.
An indication of whether the permit limit will be based on the trial burn or
other information (such as BIF Tier I for metals).
Specific information on how each permit limit will be established based
on trial burn results. This includes specifying the test condition that will
be used to establish the numerical limit and how the numerical limit will
be calculated (such as the average over all test runs).
Specific information on how each operating parameter will be monitored
and recorded to demonstrate compliance with the permit limit (i.e.,
whether the operating parameter will be continuously monitored, the type
of record to be maintained and the recording frequency, and what
averaging period will be used for the compliance determination).
Statements as to whether the parameter will be interlocked with the
AWFCS, and the corresponding proposed setpoint.
Appendix A includes a detailed list of final permit limits for an example HWI. The example facility is a
liquid injection incinerator with a heat recovery boiler and an APCD system consisting of a fabric filter
and venturi scrubber.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
74
-------
8.2
APPLICATION OF TRIAL BURN DATA IN THE SSRA
In addition to information on key operating parameters, the trial burn plan should indicate how the
emissions data from the various test conditions will be analyzed and consolidated for evaluation in the
SSRA. Data analysis for each test condition involves several steps, including evaluation of non-detected
constituents, final COPC selection, and determination of COPC emission rates as described in Section 2.3
of EPA (in press a). Consolidation of data from multiple test conditions can also encompass a range of
options, depending on how the test conditions are structured and how the data will be presented for
comparison to target risk levels. For example, separate risk values can be quantified based on different test
conditions to provide a risk range. Alternatively, a single risk estimate might be calculated by combining
the highest emission rates from multiple tests. These options should be considered carefully prior to
establishing test conditions and subsequent data collection.
Data analysis starts with decisions regarding treatment of non-detected constituents. Section 2.4.2 of EPA
(in press a) provides a recommended procedure. Several terms for non-detected constituents apply. These
include reliable detection limits (RDL) for constituents analyzed with non-isotope dilution methods,
estimated detection limits (EDL) for constituents analyzed with most isotope dilution methods, and
estimated maximum possible concentrations (EMPC) associated with SW-846 Method 8290 (EPA 1996g)
or other appropriate method.
Constituents that are not detected during any of the test runs should be re-evaluated for inclusion in the
SSRA, consistent with the final COPC selection process (and associated flowchart) found in EPA (in press
a). As part of the final COPC selection process, some constituents that are not detected during any of the
test runs will be eliminated from further consideration hi the SSRA. However, care should be taken not to
eliminate non-detected constituents that will be limited in the final RCRA permit based on the SSRA
results. This is because compliance with an emission rate limit of zero cannot be determined. Pollutants
that may be limited in the final permit based on SSRA results include D/Fs, specific metals, HC1 and C12,
and significant waste constituents (e.g., chemical agents). If these pollutants are not detected, then the full
RDLs, EDLs, or EMPCs (as appropriate) should be used in the emissions calculations. As appropriate,
permit limits may be considered for other contaminants that are found to be risk drivers during the risk
assessment process. However, emission limits are not expected to be routinely established for individual
organics, other than D/Fs.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
75
-------
For COPCs detected in at least one run of the test (as well as for COPCs that may be limited in the final
permit) the RME emission rate is calculated as the 95th percentile of the arithmetic test mean, or the
maximum value from the three test runs, whichever is lower (EPA in press a). RDLs, EDLs, or EMPCs (as
appropriate) should be used in the calculations when a COPC is not detected during a test run.
Next, consolidation of results from multiple test conditions must be considered. Possible options include
(1) calculation of risk ranges corresponding to specific test conditions (e.g., DRE and SRE), (2)
consolidation of the highest emissions data from the various test conditions for calculation of a single high-
end risk value, (3) combining test data with estimates of emission rates of COPCs to calculate high-end
risks, or (4) some combination of these. Unless all emissions determinations (D/Fs, other organics, and
metals) are conducted during each operating test condition, then consolidation of data from multiple
conditions must be considered. Further explanation regarding these possibilities are provided below and in
Appendix A.
Calculation of Risk Ranges Corresponding to Specific Test Conditions
A facility may choose to collect emissions data for use in the SSRA in conjunction with the SRE and DRE
tests, without performing a risk burn under normal operating conditions. In this case, the metals emissions
are expected to be higher in the SRE test, and the organic emissions may be higher in the DRE test. If all
emissions data are collected (D/Fs, other organics, and metals) during both test conditions, then the facility
could evaluate the emissions data sets from each separate test condition to calculate risk ranges.
Consolidation of Multiple Tests for Calculation of a Single High-End Risk Value
If the facility described above did not collect data for all COPCs during both test conditions (e.g., sampling
for metals emissions was not performed during the DRE test), then the metals emissions from the SRE test
and the higher of the D/F and organic emissions from either the SRE or DRE tests should be combined for
calculating potential RME risk estimates.
Combining Test Data with Emissions Estimates
For either of the above approaches, the facility may not have spiked mercury during the SRE test. Instead,
the facility estimated maximum mercury emissions by assuming zero SRE at a maximum mercury feed
rate. This maximum mercury emissions estimate could be combined with measured emissions data for
calculation of either a risk range or a single risk estimate.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
76
-------
Combined Approach
It is likely that metals emissions would not be determined during the DRE test, because of stack port
constraints or to limit metals spiking. In addition, it is likely that the facility would not spike mercury
during either test. In this case, a single RME risk estimate could be calculated by consolidating the
maximum estimated mercury emissions rate (and estimated emission rates for any other Tier I metals) with
the emissions data for the remaining metals collected during the SRE test, and the higher of the D/F and
organic emissions from either the SRE or DRE tests. The approach for calculation of a risk range would
be similar, except that separate risk calculations would be performed using the D/F and organic emissions
data from the SRE test versus those from the DRE test.
Obviously, if a facility chooses to collect emissions data under additional operating test conditions, then
the data management and analysis become more complex. Again, early communication and coordination
between the permit writer and facility is essential.
The primary objective of SSRA data analysis and consolidation is to ensure that the trial burn (and/or risk
burn, as appropriate) provide emissions data suitable for quantifying potential human and ecological risks
based on RME conditions. In addition, operating conditions associated with these risks should be clearly
limited by the permit terms and conditions.
Recent EPA direction in characterizing and communicating risk (EPA 1995a) to the public should also be
considered during the risk assessment process. EPA (1995a) states that risk assessments should include
risks based on data reflecting both RME and central tendency (average) conditions. As appropriate,
SSRAs could be completed based on both average and RME data. This could assist in the risk
communication process by allowing the facility and EPA to discuss a broader risk range with the public
and to consider effects of variations in operating conditions on potential risks. However, RME conditions
are generally expected to form the basis of permit conditions
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
77
-------
9.0 REPORTING CHANGES OF WASTE FEED AND OPERATING PARAMETERS
Previous sections of this document identified appropriate operating conditions and parameters for
collecting risk-based emissions data for HWC facilities including HWIs, boilers, cement kilns, and
LWAKs. Performance testing and Group A, B, and C operating parameters are used to define the HWC
facility operational envelope and to demonstrate performance standards. Risk-based emissions data are
used to establish baseline conditions for either reasonably conservative operating conditions (based on
DRE and SRE testing) or normal (or average) operating conditions. These data define a facility's baseline
with respect to long-term impacts, including potential effects on human health and the environment. This
is especially relevant for HWC facilities that conduct emissions testing under normal operating conditions.
The trial burn and corresponding RCRA permit conditions are based on assumptions that include:
Waste composition, physical form (liquid, solid, aqueous, or organic), and feed rates
Operating conditions and parameters including those related to combustion conditions
(temperature, turbulence, and residence time) and APCDs (pressure changes, flow rates,
liquid to gas ratios, inlet temperatures, and power)
Monitoring and reporting requirements should be incorporated into the final permit to ensure that any
significant changes to the facility's baseline that could affect emissions (including types of wastes treated
or major changes to operating parameters) will be reported to EPA or the appropriate state agency. This
information should be used to determine whether any additional risk-based data collection or risk analyses
will be required.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
78
-------
REFERENCES
Acharya, P., DeCicco, S.G., Novak, R.G. 1991. "Factors that Can Influence and Control the Emission of
Dioxins and Furans from Hazardous Waste Incinerators." Journal Air and Waste Management
Association. 41: 1605-1615. December.
The Air Group. 1997. "Model Parameter Sensitivity Analysis." May 23.
Agency for Toxic Substances and Disease Registry (ATSDR). 1990. "Toxicological Profile for Copper."
December.
ATSDR. 1992. "Toxicological Profile for Aluminum." July.
Altwicker, E.R., Schonberg, J.S., Ravi, K., Konduri, N.V., and Milligan, M.S. 1990. "Polychlorinated
Dioxin/Furan Formation in Incinerators." Hazardous Waste and Hazardous Materials. 7(1): 73.
Bruce, K.R. 1993. "Effect of Sulfur on Formation of PCDD/PCDF During Incineration." Presented at the
1993 International Conference on Incineration and Thermal Treatment Technologies. Knoxville,
Tennessee. May.
Berger, R., Baumbach, G., Vatter, J., and Hagenmaier. 1996. "Emissions of Polychlorinated
Dibenzodioxins and Polychlorinated Dibenzofurans of an Industrial Incineration Process at a
Precious Metal Recovery Facility in Germany and their Reduction." Presented at the 1996
International Conference on Incineration and Thermal Treatment Technologies. Savannah,
Georgia. May.
Cesmebasi, E., Dempsey, C.R., Eicher, A.R., Lee, K.C., Mukerjee, D., and Seeker, W.R. 1991. "Metals
Behavior in Waste Combustion Systems." Summary Report from ASME/EPA Workshop on the
Control of Metal Emissions from Waste Combustion Devices. Cincinnati, Ohio. November.
Clarke, L.B. and Sloss, L. L. 1992. "Trace Elements - Emissions from Coal Combustion and
Gasification." IEACR/49. IEA Coal Research, London. July.
Cundy, V.C., Sterling, A.M., Lester, T.W., Jakway, A.L., Leger, C.B., Lu, C., Montestruc, A.N., and
Cunway, R.B. 1991. "Incineration of Xylene/Sorbent Packs: A Study of Conditions at the Exit of
a Full Scale Industrial Incinerator." Environmental Science and Technology. 25: 223.
Dellinger, H.B., Pershing, D.W., and Sarofim, A.F. 1993. "Evaluation of the Origin, Emissions and
Control of Organic and Metal Compounds from Cement Kilns Co-Fired with Hazardous Wastes."
A Report of the Scientific Advisory Board on Cement Kiln Recycling. June 8.
Doull, J., Klaasen, C.D., and Amdur, M.O. 1991. Toxicology the Basic Science of Poisons. Fourth
Edition. Macmillan Publishing Co., Inc. New York.
Energy and Environmental Research Corporation (EER). 1991. "Metals Behavior in Waste Combustion
Systems." Draft Report Submitted to the International Joint Power Generation Conference.
Atlanta, Georgia. November.
EER. 1996a. "Updated Guidance on Metals Interpolation and Extrapolation for Hazardous Waste
Combustors." Draft Report. Prepared for EPA Office of Solid Waste. September 24.
EER. 1996b. "Updated Guidance on Metals Surrogates for Hazardous Waste Combustors." Draft Report.
Prepared for EPA Office of Solid Waste. September 30.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
79
-------
Ganrot, P. 1986. "Metabolism and Possible Health Effects of Aluminum." Environmental Health
Perspectives. 65: 363-441.
Caspar, J.A., Widmer, N.C., Cole, J.A., and Seeker, W.R. 1997. "Study of Mercury Speciation in a
Simulated Municipal Waste Incinerator Flue Gas." Presented at the 1997 International Conference
on Incineration and Thermal Treatment Technologies. Oakland, California. May.
Griffin, R.D. 1986. "A New Theory of Dioxin Formation in Municipal Solid Waste Combustion."
Chemosphere. 15: 1987 - 1990.
Gullett, Brian K. and Lemieux, P.M. 1994. "Role of Combustion and Sorbent Parameters in Prevention of
Polychlorinated Dibenzo-p-dioxin and Polychlorinated Dibenzofuran Formation during Waste
Combustion." Environmental Science and Technology. 28: 107-118.
Gullett, Brian K. and Raguhunathan, K. 1997. "Observations on the Effect of Process Parameters on
Dioxin/Furan Yield in Municipal Waste and Coal Systems." Chemosphere. 34: 1027-1032.
Harris, R.E., Lanier, W.S., and Springsteen, B.R. 1994. "PCDD and PCDF Emission Characteristics from
Hazardous Waste Burning Cement Kilns." Presented at the 1994 International Conference on
Incineration and Thermal Treatment Technologies. Houston, Texas. May.
Lanier, W.S., Stevens, F.M., Springsteen, B.R., and Seeker, W.R. 1996. "Dioxin Compliance Strategies
for the HWC MACT Standards." International Conference on Incinerator and Thermal Treatment
Technologies. Savannah, Georgia. May.
Lemieux, P.M., Linak, W.P., McSorley, J.A., Wendt, J.O., and Dunn, J.E. 1990. "Minimization of
Transient Emissions from Rotary Kiln Incinerators." Combustion Science and Technology. 74:
311-325.
Lemieux, P.M. and Ryan, J.V. 1998. "Enhanced Formation of Dioxins and Furans from Combustion
Devices by Addition of Trace Quantities of Bromine." Presented at the 1998 International
Conference on Incineration and Thermal Treatment Technologies. Salt Lake City, Utah. May.
Lemieux, P.M. and Ryan, J.V. In press. "Enhanced Formation of Chlorinated PICs by the Addition of
Bromine." Accepted for publication in Combustion Science and Technology.
Linak, W.P., Ryan, J.V., and Wendt, J.O.L. 1996. "Formation and Destruction of Hexavalent Chromium
in a Laboratory Swirl Flame Incinerator." Combustion Science and Technology. 116-117:479.
Linak, W.P. and Wendt, J.O.L. In press. "Partitioning of the Refractory Metals, Nickel and Chromium, in
Combustion Systems." Accepted for Publication in Combustion Science and Technology.
Midwest Research Institute and A.T. Kearney, Inc. 1997. "Products of Incomplete Combustion Emission
Test." Draft Report. Prepared for EPA Office of Solid Waste. April.
Paasivirta, J. 1991. Chemical Ecotoxicology. Lewis Publishers, Inc. Chelsea, Michigan.
Raghunathan, K. and Gullett, B.K. 1994. "Effect of Sulfur in Reducing PCDD/PCDF Formation."
Presented at the 1994 International Conference on Incineration and Thermal Treatment
Technologies. Houston, Texas. May.
Raghunathan, K., Gullet, B.K., Chun, W.L., and Kilgroe, J.D. 1997. "Reducing Dioxin Formation
Through Coal Co-Firing." Presented at the 1997 International Conference on Incineration and
Thermal Treatment Technologies. Oakland, California. May.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
80
-------
Research Triangle Institute (RTI). 1996. "North Carolina Protocol for Performing Indirect Exposure Risk
Assessments for Hazardous Waste Combustion Units." Draft. Prepared by RTI for the State of
North Carolina. 92D-6489-000. March.
Rigo, H.G., Chandler, A.J., and Lanier, W.S. 1995. "The Relationship Between Chlorine in Waste
Streams and Dioxin Emissions from Combustors." Draft. Prepared for the American Society of
Mechanical Engineers. January 6.
Ryan, V.R., Lemieux, P.M., Lutes, C., and Tabor, D. 1996. "Development of PIC Target Analyte List for
Hazardous Waste Incineration Processes." Presented at the International Conference on
Incineration and Thermal Treatment Technologies. Savannah, Georgia. May.
Ryan, J.V., Lemieux, P.M., and Groff, P.W. 1997. "Evaluation of the Behavior of Flame lonization
Detection Total Hydrocarbon Continuous Emission Monitors at Low Concentrations." Presented
at the International Conference on Incineration and Thermal Treatment Technologies. Oakland,
California. May.
Santoleri, J.J. 1995. "Dioxin Emissions - Effect of Chlorine/Time/Temperature Relationship at 300 ฐC."
Presented at the 1995 International Conference on Incineration and Thermal Treatment
Technologies. Bellevue, Washington. May.
Schofield, B., Eicher, A.R., and Crouch, H.C. 1997. "Conducting the Maximum Waste Feed Rate,
Minimum Combustion Temperature Test Condition for Boilers which Might Burn High Btu Waste
- A Case Study." Presented at the 1997 International Conference on Incineration and Thermal
Treatment Technologies. Oakland, California. May.
Shaub, W.M. and Tsang, W. 1983. "Dioxin Formation in Incinerators." Environmental Science and
Technology. 17:721. December.
Sidhu, L., Maqsud, L., Dellinger, B., and Mascolo, G. 1994. "The Homogeneous, Gas-Phase Formation of
Chlorinated and Brominated Dibenzo-p-dioxins from 2,4,6-Trichloro and 2,4,6-Tribromophenols."
Presented at the 25th Combustion Symposium, the Colloquium on Incineration and Wastes.
Townsend, D.I., Wilson, J.D., and Park, C.N. 1995. "Mechanisms for Formation and Options for Control
of Emissions of PCDD's/PCDF's from Incineration." Presented at the 1995 International
Incineration Conference. Bellevue, Washington. May.
Ullrich, R., Davidson, B., and Grater, L. 1996. "Practical Experience with Dioxin Synthesis and Control
in a Variety of Full Scale Gas Cleaning Trains." Presented at the 1996 International Conference
on Incineration and Thermal Treatment Technologies. Savannah, Georgia. May.
U.S. Environmental Protection Agency (EPA). 1983. "Guidance Manual for Hazardous Waste Incinerator
Permits." Final. Prepared by the Mitre Corporation for the U.S. EPA Office of Solid Waste. SW-
966. July.
EPA. 1989. "Guidance on Setting Permit Conditions and Reporting Trial Burn Results, Volume II of the
Hazardous Waste Incineration Guidance Series." Office of Research and Development.
EPA/625/6-89/019. January.
EPA. 1992a. "Implementation of Boiler and Industrial Furnace (BIF) Regulations - New Toxicological
Data." Memorandum from Shiva Garg to EPA Regions 1 through 10. Office of Solid Waste and
Emergency Response. February.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities 81
-------
EPA. 1992b. "Technical Implementation Document for EPA's Boiler and industrial Furnace Regulations."
Office of Solid Waste and Emergency Response. EPA-530-R-92-001. March.
EPA. 1993. "EPA Draft Strategy for Combustion of Hazardous Waste in Incinerators and Boilers." May
18.
EPA. 1994a. "Combustion Emissions Technical Resource Document." EPA530-R-94-014. May.
EPA. 1994b. "Exposure Assessment Guidance for RCRA Hazardous Waste Combustion Facilities."
Draft. Office of Solid Waste and Emergency Response. EPA530-R-94-021. April.
EPA, 1994c. "Strategy for Hazardous Waste Minimization and Combustion." EPA530-R-94-044.
November.
EPA. 1995a. "Guidance for Risk Characterization." Science Policy Council. February.
EPA. 1995b. "Dioxin Reassessment Review." Science Advisory Board Report. May.
EPA. 1996a. "Revised Standards for Hazardous Waste Combustors." Proposed Rule. Title 40 of the Code
of Federal Regulations Parts 60, 63, 260, 261, 264, 265,266, 270, and 271. Federal Register 61:
17358. April 19.
EPA. 1996b. "Technical Support Document for HWC MACT Standards." Draft. In Volume I,
"Description of Source Categories." February.
EPA. 1996c. "Technical Support Document for HWC MACT Standards." Draft. In Volume III,
"Selection of MACT Standards and Technologies." February.
EPA. 1996d. "Technical Support Document for HWC MACT Standards." Draft. In Volume IV,
"Compliance with the Proposed MACT Standards." February.
EPA. 1996e. "Technical Support Document for HWC MACT Standards." Draft. In Volume VII,
"Miscellaneous Technical Issues." February.
EPA. 1996f. "Guidance for Total Organics, Final Report." EPA/600/R-96/036. March.
EPA. 1996g. "SW-846, Test Methods for Evaluating Solid Waste." Fourth Revision. December.
EPA. 1997a. "Revised Technical Standards for Hazardous Waste Combustion Facilities." Proposed Rule.
Title 40 of the Code of Federal Regulations Parts 60, 63, 260, 264, 265, 266, 270, and 271.
Federal Register 62:24211. May 2.
EPA. 1997b. "Notice of Data Availability and Request for Comments. Total Mercury and Particulate
Continuous Emissions Monitoring Systems." Proposed Rule. Title 40 of the Code of Federal
Regulations, Parts 60 and 63. Federal Register 62:67788. December 30.
EPA. 1997c. "Development of a Hazardous Waste Incinerator Target Analyte List of Products of
Incomplete Combustion." Final Report. Prepared by EPA National Risk Management Research
Laboratory, Research Triangle Park, North Carolina, for the Office of Solid Waste. July.
EPA. 1997d. Unpublished Data. Office of Solid Waste. August.
EPA. 1997e. "Mercury Study Report to Congress." Volumes I through Vffl. Final. Office of Air Quality
Planning and Standards and Office of Research and Development. December.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
82
-------
EPA. 1997f. "Hazardous Waste Combustion Unit Permitting Manual. Component 1 - How to Review a
Trial Burn Plan." Center for Combustion Science and Engineering, Multi Media Planning
Division, EPA Region 6. December.
EPA. 1997g. "Hazardous Waste Combustion Unit Permitting Manual. Component 2 - How to Review a
Quality Assurance Project Plan." Center for Combustion Science and Engineering, Multi Media
Planning Division, EPA Region 6. December.
EPA. 1997h. "Hazardous Waste Combustion Unit Permitting Manual. Component 3 - How to Review a
Part B Permit Application." Center for Combustion Science and Engineering, Multi Media
Planning Division, EPA Region 6. December.
EPA. 1997L "Hazardous Waste Combustion Unit Permitting Manual. Component 4 - How to Review a
Trial Burn Report." Center for Combustion Science and Engineering, Multi Media Planning
Division, EPA Region 6. December.
EPA. 1997J. "Hazardous Waste Combustion Unit Permitting Manual. Component 5 - How to Prepare
Permit Conditions." Center for Combustion Science and Engineering, Multi Media Planning
Division, EPA Region 6. December.
EPA. 1997k. "Notice of Draft Source Category Listing for Section 112(d)(2) Rulemaking Pursuant to
Section 112(c)(6) Requirements." Federal Register 62: 33625. June 20.
EPA. 19971. "Hazardous Waste Combustors; Revised Standards; Proposed Rule-Notice of Data
Availability and Request for Comments." Notice of Data Availability and Request for Comments
40 CFR Parts 60, 63.260, 261, 264, 265, 266,270, and 271. Federal Register: 960. January 7.
EPA. 1998a. Integrated Risk Information System (IRIS). On-line Database (http://www.epa.gov/iris).
EPA. 1998b. Unpublished Data. EPA Region 6. Dallas, Texas.
EPA. 1998c. Unpublished data. EPA Region 10. Seattle, Washington.
EPA. In press a. "Protocol for Human Health Risk Assessment at Hazardous Waste Combustion
Facilities." EPA-R6-098-002. Center for Combustion Science and Engineering, Multimedia
Planning Division, EPA, Region 6.
EPA. In press b. "Protocol for Screening Level Ecological Risk Assessment at Hazardous Waste
Combustion Facilities." EPA-R6-098-003. Center for Combustion Science and Engineering,
Multimedia Planning Division, EPA, Region 6.
Waterland, L.R. and Ghorishi, S.B. 1997. "Rapid High-Temperature Dioxin Formation: Pilot-Scale Test
Results from the U.S. EPA Incineration Research Facility." Presented at the 1997 International
Conference on Incineration and Thermal Treatment Technologies. Oakland, California. May.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities ง3
-------
APPENDIX A
TRIAL BURN CONDITIONS AND PERMIT LIMITS
FOR AN EXAMPLE HWC FACILITY
(10 pages)
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
A-l
-------
APPENDIX A
TRIAL BURN CONDITIONS AND PERMIT LIMITS
FOR AN EXAMPLE HWC FACILITY
This Appendix includes a detailed list of Group A, B, and C control parameters (and associated monitoring
requirements) for an example hazardous waste incinerator (HWI) (Facility Z). Facility Z includes a liquid
injection combustion chamber burning organic liquid and aqueous wastes, a heat recovery boiler, and an
air pollution control device (APCD) system consisting of a fabric filter and venturi scrubber. Facility Z
typically operates within a 1,750 + 50 ฐF combustion temperature window, with exit gas temperatures
from the heat recovery boiler in the range of 350 to 550 ฐF. Gas exit temperatures from the heat recovery
boiler generally increase and decrease with combustion temperature. Facility Z plans to perform three test
conditions (summarized in Table A-l) that consist of the following:
Destruction and Removal Efficiency Test Condition
The destruction and removal efficiency (DRE) demonstration is proposed at a minimum combustion
temperature of 1,600 ฐF to promote maximum operating flexibility. This test also involves the key
operating parameters for organic products of incomplete combustion (PIC); therefore, the facility plans to
measure PICs and total organics (TO) in conjunction with the DRE performance demonstration. A
temporary total hydrocarbon continuous emissions monitor (CEM) will be operational during the DRE/PIC
testing. Carbon monoxide monitoring will be performed during the DRE conditions.
Dioxin/furan (D/F) testing will also be performed during the DRE test. This is because of concerns that
the high combustion temperatures demonstrated during the system removal efficiency (SRE) test might not
adequately represent the presence of D/F precursors formed during poor combustion conditions. In
general, the formation of D/F precursors during conditions of poor combustion are not a significant
concern for this system (because the historical data review indicated very constant operating temperatures,
very low carbon monoxide, and very few waste feed cutoffs). In addition, the 350ฐF fabric filter inlet
temperature planned for the DRE test is outside of the critical D/F temperature range (400 - 750 ฐF).
However, the facility agreed that D/F sampling during DRE conditions would better represent the
operating envelope, and the stack sampling ports could accommodate all of the necessary sampling trains
for multiple determinations.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
A-2
-------
System Removal Efficiency Test Condition
Facility Z plans to demonstrate SRE for metals at maximum feed rates by spiking with arsenic, beryllium,
cadmium, chromium, lead, and nickel. The SRE demonstration for nickel is necessary based upon a
preliminary risk evaluation that quantifies potential carcinogenic risks from inhaled nickel emissions. The
SRE demonstration will be performed at a maximum combustion temperature of 1,850 ฐF and a maximum
inlet fabric filter temperature of 550 ฐF. Since all of the feed mechanisms at Facility Z are for liquid
feeds, there is no distinction between total metal feed rates and pumpable metal feed rates.
The primary operating parameters related to D/F formation for Facility Z are boiler exit temperature and
fabric filter inlet temperature (both operating parameters are represented by the same measurement location
for the Facility Z system configuration). D/Fs are expected to be maximized at the maximum fabric filter
inlet temperature of 550 ฐF, and will be measured in conjunction with the SRE test. Particulate matter
(PM) and hydrogen chloride (HC1) and chlorine (C12) will also be measured, because the SRE test
maximizes chlorine emissions, ash production, and flue gas velocity. Carbon monoxide monitoring will be
performed during the SRE conditions.
Normal Test Condition
Facility Z is capable of defining and maintaining a normal operating condition. A normal test is proposed
for metals, because of concerns that spiking with metals will result in risks that exceed target values. The
normal test will involve normal metal feed rates, a combustion temperature of approximately 1,750 ฐF, and
a fabric filter inlet temperature of 450 ฐF. Emissions testing will be performed for all 18 metals (the 12
boiler and industrial furnace [BIF] regulation metals and the six additional compounds of potential concern
[COPC] metals from Table A-l of Appendix A of EPA's "Protocol for Human Health Risk Assessment at
Hazardous Waste Combustion Facilities," EPA-R6-098-002, Center for Combustion Science and
Engineering, Multimedia Planning and Permitting Division, EPA Region 6, [in press]).
PM and HC1 and C12 emissions testing were also added to the normal test to establish minimum pressure
differential limits for the fabric filter and venturi scrubber. Minimum pressure differentials conflict with
the maximum combustion gas velocities demonstrated during DRE and SRE test conditions. A particle-
size determination was added to the normal test, since this test does not include ash spiking. Carbon
monoxide monitoring will be performed during the normal conditions.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities A-3
-------
Site-Specific Risk Assessment
Facility Z performed the trial burn according the approved test plan. The D/F emissions, as well as the
HC1 and C12 emissions, were highest during the SRE test condition. Emissions were consolidated from the
three test conditions for evaluation in the multi-pathway human health and ecological site-specific risk
assessments (SSRA) as follows:
D/F emissions from the SRE test condition
Metals emissions (18 metals) from the normal test condition
Organic PIC emissions from the DRE test condition
HC1 and C12 emissions from the SRE test condition
Total risks from these consolidated emissions were below target levels.
Next, a second SSRA evaluation was performed considering only the inhalation pathway, consistent with
the BIF regulations. This evaluation used emissions from the SRE test condition for the spiked metals
(arsenic, beryllium, cadmium, chromium, lead, and nickel), as well as emissions estimates for the
remaining metals based on maximum anticipated feed rates and an assumption of zero SRE. Emission rate
data for all other pollutants were unchanged. Total inhalation risks from these consolidated emissions
were also below target levels.
Final Permit Limits
Table A-2 provides the final permit limits for Facility Z. For both minimum combustion temperature and
maximum fabric filter inlet temperature, limits were established based on the following dual averaging
periods:
Hourly rolling averages (HRA)
Short-term limits (10-minute rolling averages [RA]) to control temperature spikes which
could result in excess D/F or organic PIC emissions
HRA feed rate limits for the six spiked metals (arsenic, beryllium, cadmium, chromium, lead, and nickel)
were established based on feed rates measured during the SRE test. For two of the unspiked metals
(barium and mercury), HRA feed rate limits were established using the inhalation risk-based limit (and an
assumption of no SRE). For the remaining 10 unspiked metals (aluminum, antimony, cobalt, copper,
manganese, selenium, silver, thallium, vanadium, and zinc), HRA feed rate limits were considered
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities A-4
-------
unnecessary because inhalation risks were either negligible or not applicable (considering the range of
inputs for these metals at Facility Z). Quarterly average metals feed rate limits were established for all
eighteen metals based on feed rates demonstrated during the normal test condition.
Total hydrocarbon levels were negligible during the ORE test condition. Therefore, the permit writer
determined that there was no need to require continued total hydrocarbon monitoring.
Final Permitted Emission Rates
Maximum emission rate limits were established in the permit for D/Fs and C12 based on multi-pathway
SSRA target risk levels. The limits were established for the purpose of periodic verification testing to
ensure that emissions remain below those evaluated in the SSRA. If emissions increases occur above
target levels, then the SSRA would need to be redone. Both "normal" and "maximum" emission rate limits
for metals were established corresponding to the multi-pathway SSRA and inhalation-only target risk
levels, respectively.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities A-5
-------
TABLE A-l
FACILITY Z TEST CONDITIONS
/ i! ^
"^
J f *"
A f U
* J<&
1 x
Combustion temperature
Fabric filter inlet
temperature
Organic liquid feed rate
Aqueous liquid feed rate
Combustion gas velocity
Ash feed rate
Chlorine feed rate
Spiked metals
(Arsenic, beryllium,
cadmium, chromium, lead,
and nickel)
Other metals
Fabric filter differential
pressure
Venturi differential
pressure
Venturi liquid-to-gas ratio
Venturi scrubber liquid
exit pH
Scrubber blowdown rate
. ' * ' lEST CONDITIONS; - -. t
2 . AND EMISSIONS DETERMINATIONS
"^ * s
DBE i ',,"--
fOHCs, PICs;IปFs, TO,
Total Hydrocarbons, Carbon 1
Monoxide ,- , '
v ' ' ;,
1,600 ฐF, HRA
(1,575 ฐF, 10-min. minimum
RA)
350 ฐF
Maximum
Maximum
Maximum *
Above normal
Maximum
N/A
N/A
Maximum *
Maximum *
Minimum
Minimum
Minimum
SRE * " \ '" ^
MeJais,D/Fs;- -, --
PM,HCI/C12, ',
Carbon Morioiide
1,850 ฐF HRA
550 ฐF
(565 ฐF, 10-min.
maximum RA)
Maximum
Minimum
Maximum *
Maximum
Maximum
Maximum
Normal
Maximum *
Maximum *
Minimum
Minimum
Minimum
Notes:
* = r.nnflir.tina naramfttftr
-------
TABLE A-2
FINAL PERMIT LIMITS FOR FACILITY Z
Operating Parameters
Value
Basis
Test
Established As:
Group A
Minimum combustion
temperature
Maximum combustion
temperature
Maximum combustion
gas velocity
Maximum organic
liquid feed rate
Maximum aqueous feed
rate
Maximum fabric filter
inlet temperature
Minimum fabric filter
pressure differential
Minimum venturi
scrubber differential
pressure
Minimum venturi
liquid-to-gas ratio
Minimum venturi
scrubber liquid exit pH
Minimum scrubber
blowdown flow rate
Maximum stack Carbon
Monoxide concentration
1,600 ฐF, HRA
1,575 ฐF, 10-min. RA
1,850 ฐF, HRA
cfm,HRA
lbs/hr,HRA
Ibs/hr, HRA
550 ฐF, HRA
565 ฐF, 10-min. RA
inches water column,
HRA
inches water column,
HRA
gal/cfm,HRA
pH,HRA
gpm,HRA
100ppmat7%
oxygen, dry basis,
HRA
DRE
DRE
SRE
SRE/
DRE
DRE
DRE
SRE
SRE
Normal
Normal
SRE
SRE
SRE
N/A
Avg. of the three lowest HRA
temperatures during the DRE test
Avg. of the three lowest 10-min. RAs
during the DRE test
Avg. of the three highest HRA
temperatures during the SRE test
Avg. of the six highest HRA velocities
during DRE and SRE
Avg." of the three highest HRA feed
rates during the DRE test
Avg. of the three highest HRA feed
rates during the DRE test
Avg. of the three highest HRA
temperatures during the SRE test
Avg. of the three highest 10-min. RAs
during the SRE test
Avg. of the three lowest HRA pressure
differentials during the normal test
Avg. of the three lowest HRA pressure
differentials during the normal test
Avg. of the three lowest HRA ratios
during the SRE test
Avg. of the three lowest HRA pHs
during the SRE test
Avg. of the three lowest HRA flow
rates during the SRE test
Limit based on established guidance
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
A-7
-------
TABLE A-2
FESTAL PERMIT LIMITS FOR FACILITY Z (Continued)
/ ^
Operating Parameters
Maximum combustion
chamber pressure
Group B
POHC incinerability
limits
Maximum chlorine feed
rate
Maximum ash feed rate
Maximum feed rates
(spiked metals):
- arsenic, beryllium,
cadmium, chromium,
lead, and nickel
Maximum feed rates:
- barium and mercury
Maximum feed rates:
- aluminum, antimony,
arsenic, barium,
beryllium., cadmium,
chromium, cobalt,
copper, lead,
manganese, mercury,
nickel, selenium,
silver, thallium,
vanadium, and zinc
* ~* *
'r.
* ""
Value
inches water column,
vacuum,
instantaneous limit
(interlocked with
AWFCS)
Allowable Appendix
VIE constituents
Ibs/hr, HRA
Ibs/hr, HRA
Ibs/hr, HRA
Ibs/hr, HRA
Ibs/hr, quarterly
average
Test
N/A
DRE
SRE
SRE
SRE
SRE
Normal
^ Basis '
'ia fs <
' Established As:
As necessary to maintain negative
pressure
Based on POHCs which achieved
99.99% DRE
Avg. of the three highest HRA feed
rates during the SRE test
Avg. of the three highest HRA feed
rates during the SRE test
Avg. of the three highest HRA feed
rates during the SRE test
Zero SRE; the inhalation risk-based
emission limit is used as the feed rate
limit
Avg. over all three runs of feed rates
during the normal test
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
A-8
-------
TABLE A-2
FINAL PERMIT LIMITS FOR FACILITY Z (Continued)
Operating Parameters
Value
' ' ''" Basis '
Test
Established As:
Group C
Maximum heat input
Burner/atomizer:
- Maximum viscosity
- Maximum turndown
- Maximum solids
- Minimum atomizing
pressure differential
Minimum venturi
scrubber nozzle
pressure
million Btu/hr
- centipoise
- gpm range
- percent solids
- psig (interlocked
withAWFCS)
psig (interlocked with
AWFCS)
N/A
N/A
N/A
Design basis
Manufacturer's recommendations
Manufacturer's recommendations
Summary of Performance Standards and Emission Limits
DRE
Maximum PM
emissions
Maximum HC1
emissions
Maximum C12 emissions
Maximum D/F
emissions
Maximum emissions,
each metal
Normal emissions, each
metal
99.99% for POHCs
0.08 gr/dscf
Lower of 4 Ibs/hr or
inhalation risk-based
limit
Inhalation
risk-based limit
SSRA risk-based limit
Inhalation SSRA risk-
based limits
Multi-pathway SSRA
risk-based limits
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Statutory requirement
Regulatory limit
Regulatory limit/SSRA target level
SSRA target level
SSRA target level
SSRA target level, established for the
purpose of periodic SRE verification
tests at conducted at maximum metals
feed rates and/or confirmation of Tier
I estimates
SSRA target level, established for the
purpose of defining the baseline for
the SSRA. Increases in normal metal
feed rates or emission rates could
require retesting and/or another SSRA
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
A-9
-------
Notes:
AWFCS
Avg.
Btu/hr
cfm
gal/cfm
gpm
Ibs/hr
min.
N/A
ppm
psig
POHC
gr/dscf
TABLE A-2
FINAL PERMIT LIMITS FOR FACILITY Z (Continued)
automatic waste feed cutoff system
average
British thermal units per hour
cubic feet per minute
gallons per cubic feet per minute
gallons per minute
pounds per hour
minutes
not applicable
parts per million
pounds per square inch, gauge
principle organic hazardous constituent
grains per dry standard cubic foot
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
A-10
-------
APPENDIX B
SAMPLING AND ANALYSIS
(51 pages)
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities
-------
CONTENTS
Section
Page
ACRONYM LIST B-iv
B.I OVERVIEW OF SAMPLING AND ANALYSIS PROCEDURES B-l
B.I.I EMISSIONS TESTING OBJECTIVES B-l
B.1.2 DETECTION LIMITS B-3
B.1.3 METHOD SUMMARY B-4
B.2 DIOXINS AND FURANS B-10
B.3 VOLATILE ORGANIC COMPOUNDS B-12
B.4 SEMIVOLATILE ORGANIC COMPOUNDS B-15
B.5 TENTATIVELY IDENTIFIED COMPOUNDS B-19
B.6 CHLOROBENZENES/CHLOROPHENOLS B-21
B.7 POLYCYCLIC AROMATIC HYDROCARBONS B-23
B.8 POLYCHLORINATED BIPHENYLS B-25
B.9 ANALYSIS OF MM5 SAMPLES FOR MULTIPLE POLLUTANT CLASSES B-28
B.10 ALDEHYDES/KETONES B-33
B.ll FACILITY-SPECIFIC COMPOUNDS B-34
B.12 TOTAL ORGANICS B-36
B. 13 TOTAL HYDROCARBON AND CARBON MONOXIDE CONTINUOUS EMISSIONS
MONITORS (CEM) B-42
B.14 METALS B-43
B.15 PARTICLE-SIZE DISTRIBUTION B-44
B.16 HYDROGEN CHLORIDE AND CHLORINE B-46
B.17 PROCESS SAMPLES B-48
REFERENCES B-49
Attachment
Method 0040 Clarifications
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-ii
-------
TABLES
Table Page
Bl-1 RISK-BASED STACK EMISSION DETERMINATIONS B-6
B3-1 VOLATILE TARGET COMPOUND LIST - METHOD 8260B ANALYTES B-13
B4-1 SEMIVOLATILE TARGET COMPOUND LIST - METHOD 8270C ANALYTES B-16
B6-1 CHLOROBENZENES AND CHLOROPHENOLS B-22
B7-1 POLYCYCLIC AROMATIC HYDROCARBONS B-24
B8-1 POLYCHLORINATED BIPHENYLS B-26
Bll-1 ORGANOCHLORINE PESTICIDES - METHOD 8081A ANALYTES B-35
B12-1 TOTAL ORGANICS FGC ANALYSIS B-39
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-iii
-------
jig
ug/mL
AED
A/K
AMS
APCD
BIF
BP
CARB
CAS
CB
CEM
CO
CP
D/F
DQO
ECD
EDL
EMPC
EPA
ESP
FGC
FID
GC
gm/cm3
GRAY
HEPA
HHRA
HPLC
HRGC
HRMS
ACRONYM LIST
Microgram
Microgram per cubic meter
Microgram per milliliter
Atomic emission detection
Aldehyde/ketone
Alkaline mercury speciation
Air pollution control device
Boiler and industrial furnace
Boiling point
California Air Resources Board
Chemical Abstract Services
Chlorobenzene
Continuous emissions monitor
Carbon monoxide
Chlorophenol
Dioxin and furan
Data quality objective
Electron capture detector
Estimated detection limit
Estimated maximum possible concentration
U.S. Environmental Protection Agency
Electrostatic precipitator
Field gas chromatography
Flame ionization detector
Gas chromatography
Grams per cubic centimeter
Gravimetric
High efficiency particulate air
Human health risk assessment
High performance liquid chromatography
High resolution gas chromatography
High resolution mass spectrometry
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-iv
-------
ACRONYM LIST (Continued)
HWC Hazardous waste combustion
HWI Hazardous waste incinerator
IUPAC International Union of Pure and Applied Chemists
LC Liquid chromatography
LRMS Low resolution mass spectrometry
MDGC Multi-dimensional gas chromatography
mg Milligram
MM5 Modified Method 5
MS Mass spectrometry
ng Nanogram
PAH Polycyclic aromatic hydrocarbon
PCB Polychlorinated biphenyl
ppb Parts per billion
ppm Parts per million
QA/QC Quality assurance/quality control
RCRA Resource Conservation and Recovery Act
RDL Reliable detection limit
RRF Relative response factor
SEM Scanning electron microscopy
SOP Standard operating procedure
SSRA Site-specific risk assessment
SVOC Semivolatile organic compound
TCO Total chromatographable organic
TEQ Toxic equivalent
THC Total hydrocarbon
TIC Tentatively identified compound
TO Total organic
VOC Volatile organic compound
VOST Volatile organic sampling train
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-v
-------
B.1 OVERVIEW OF SAMPLING AND ANALYSIS PROCEDURES
This appendix discusses recommended procedures for collection of stack emissions data for site-specific
risk assessments (SSRA). As explained previously in Section 2.0 of this document, emissions data needs
for SSRAs include the following:
Dioxins and furans (D/F)
Non-dioxin organics and total organic (TO) mass
Metals
Particle-size distribution
Hydrogen chloride and chlorine
Although the majority of this appendix focuses on the sampling and analysis procedures for obtaining this
stack emissions data, recommended characterizations for process samples are also discussed briefly in
Section B.I 7. Characterization of waste feeds, auxiliary fuels, and raw materials during the risk testing is
necessary in order to establish the basis for the SSRA.
This appendix is intended to be a tool for assisting permit writers and facility managers in making
informed decisions about the measurements necessary for their SSRA data needs. It identifies appropriate
methods and highlights sampling, sample recovery, and analytical considerations which can influence
detection limits, and therefore, the results of the SSRA. To a limited extent, it also provides method
clarifications and "lessons learned" that have not been widely published elsewhere.
B.1.1
EMISSIONS TESTING OBJECTIVES
Sampling and analyses for SSRA emissions data should accomplish two primary objectives as follows:
Achieve the most comprehensive emissions characterization possible by (1) identifying
and quantifying specific toxic constituents, in order to assess their contribution to the total
risk posed by the facility; and (2) identifying and quantifying as many other constituents as
possible, regardless of toxicity
Estimate the completeness of the organic emissions characterization, in order to evaluate
the uncertainty associated with the SSRA process
The first objective is achieved via analysis for specific target analytes. Target analyte lists, by design,
address known toxic compounds. Analyses for full, commercially available target analyte lists should
generally be completed by all facilities. A priori deletion of individual compounds (e.g., because they are
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-l
-------
not found in the waste, or because they are not expected to be risk drivers) is not appropriate. Organic
stack emissions cannot be predicted with certainty based upon waste characteristics. In addition, analyses
for complete target analyte lists, when conducted using standard U.S. Environmental Protection Agency
(EPA) methods, do not subject facilities to significant additional costs or burdens during the trial burn
process. Therefore, a priori deletion of individual compounds is not likely to significantly reduce costs
and could jeopardize the chances of identifying the greatest possible percentage of organics.
Measurements for non-toxic volatile organics are also highly recommended in order to improve the overall
emissions characterization.
The second, equally important objective, can be achieved by performing an organics mass balance for the
organic stack emissions. The mass balance endeavors to (1) quantify the TO mass being emitted from the
stack, and (2) estimate the quantity of unidentified organics, based on the difference between the TO mass
and the total quantity of identified organics. Measurements for TO mass support this second objective.
The human health risk assessment (HHRA) guidance (EPA in press) contains a detailed discussion
regarding the use of TO for performing the organics mass balance and estimating uncertainty.
This appendix emphasizes methods and standard target analyte lists that are commercially available.
Research by Ryan and others (1997a, 1997b), Lemieux and Ryan (1997b, 1998), and Midwest Research
Institute (MRI) and AT Kearney, Inc. (1997) provides valuable insight into the limitations of standard
methods for identifying and quantifying organics. However, this appendix incorporates virtually the entire
repertoire of commercially available methods for stack samples, and alternative options are limited at this
time. Some of the more innovative techniques suggested by Lemieux and Ryan (e.g., multi-dimensional
gas chromatography/mass spectrometry (MDGC/MS), gas chromatography/atomic emission detection
(GC/AED), and liquid chromatography/mass spectrometry (LC/MS) appear promising for further research,
but additional method development is needed to support the use of these methods for stack samples on a
commercial scale. Thus, it is unreasonable to presume that these innovative methods could be used for
typical trial burn applications.
Until further method development can be pursued, organics determinations should rely on the best
currently available methods for specific target analytes and tentatively identified compounds (TIC),
combined with a TO determination, to clearly communicate the portion of the organic mass that cannot be
identified. Research by Ryan, Lemiuex, and MRI highlights the significance of currently available options
for optimizing organic emissions characterization. All of their studies relied on comprehensive evaluations
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-2
-------
for TICs to expand identifications beyond standard target analyte lists (Lerhieux and Ryan 1997b, 1998;
MRI1997). TIC identification also played a key role in full-scale research (Energy and Environmental
Research Corporation [EER] 1997). Several of the studies (MRI 1997; EER 1997) emphasize the
importance of determining as many volatile compounds as possible, regardless of toxicity. In these studies,
screening for Cl through C4 straight-chain alkanes, alkenes, and alkynes was performed using either an
on-line gas chromatograph/flame ionization detector (GC/FID), or Cl-C4-targeted analyses of Method
0040 bag samples. Methane can comprise a significant percentage of the total stack organics (MRI 1997;
Johnson 1996a). Determinations for methane and other aliphatics add little cost to the emissions testing
and can potentially alleviate concerns about the percentage of the TO mass that might be toxic. Finally,
several of the studies indicate that total hydrocarbon (THC) measurements may not be adequate or
appropriate for supporting an organics mass balance (Ryan and others 1997a; MRI 1997). An obvious
drawback to THC monitors is that they only measure gas-phase organics. Particulate material, including an
indeterminate but sometimes significant fraction of the organic material, is removed by filtration and
discarded. This is clearly unacceptable when attempting to measure TOs (Johnson 1996a). In lieu of
THC, the TO mass determination has been repeatedly cited as the preferred starting point for the mass
balance (Lemieux and Ryan 1997b, 1998; Johnson 1996a; EPA 1996d).
B.1.2
DETECTION LIMITS
In planning the specific sampling and analysis procedures, each facility should carefully consider detection
limits that may be required in order to demonstrate risks below target levels. The HHRA guidance (EPA in
press) contains a detailed discussion regarding required treatment of reliable detection limits (RDL),
estimated detection limits (EDL), and estimated maximum possible concentrations (EMPC) in risk
assessments. If detection limits are not low enough to achieve target risk levels, then modifications to the
sampling and analytical procedures (e.g., longer sampling times, high resolution analytical techniques, etc.)
may need to be considered.
It is highly recommended that a facility evaluate anticipated detection limits in a preliminary risk
evaluation to determine whether modifications to the sampling and analytical procedures will be needed.
This guidance, as well as the cited references, provide limited information regarding detection limits that
can generally be expected from the various sampling train/analytical method combinations. However,
detection limits are best determined based upon discussions with the qualified analytical chemist that will
oversee the analysis. Detection limits for D/Fs have been found to be critical for the indirect risk pathway.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-3
-------
Other organics that may require particular attention because they could be important contributors to
indirect risk include the following:
Polycyclic Aromatic Hydrocarbons (PAH)
Benzo(a)pyrene
Benz(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Chrysene
Dibenz(a,h)anthracene
Indeno(l ,2,3-cd)pyrene
Polvchlorinated Biphenvls (PCB1
Total PCBs
13 dioxin-like coplanar PCBs, listed in Table B8-1
Phthalates
Bis(2-ethylhexyl) phthalate
Di(n)octyl phthalate
Nitroaromatics
1,3-Dinitrobenzene
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Nitrobenzene
Pentachloronitrobenzene
Other Chlorinated Organics
Hexachlorobenzene
Pentachlorophenol
B.1.3
METHOD SUMMARY
Table Bl-1 provides a summary of applicable stack sampling and analysis requirements for the SSRA data
needs. Specific information on each determination is provided in subsequent sections.
The emissions determinations rely on SW-846 methods (EPA 1996a), where available. Method numbers
correlate with those in SW-846 unless otherwise specified. The method suffixes correspond to those in the
December 1996 update (EPA 1996a). However, SW-846 will continue to be updated, and the latest
revision of the recommended methods should be used as guidance. Where SW-846 does not provide a
method for a particular determination, alternate sources are referenced. All of the procedures discussed in
this appendix have been intentionally limited to methods that are commercially available.
SW-846 methods, as written, are intended as guidance and should be used as starting points for standard
operating procedures (SOP) of the methods that will actually be used. SW-846 also provides guidance on
individual method modifications. Other methods, not in SW-846, may be used for generating SSRA
emissions data, provided that the user can demonstrate and document that the methods meet the data
quality objectives (DQO) for the particular application. Even when using SW-846 methods, either as
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-4
-------
written or modified, the user will need to demonstrate that the methods generate data that meet the DQOs.
Therefore, detailed information regarding the actual methods should be described in detail in the trial burn
plan, and the trial burn plan should define DQOs for the particular application. In some cases, additional
modifications to standard EPA methods may be appropriate in order to address site-specific circumstances
(e.g., sampling port limitations or a need for reduced detection limits). These modifications to standard
EPA methods should be described in detail in the trial bum plan and are subject to approval by the
regulatory agency.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-5
-------
TABLE Bl-1
RISK-BASED STACK EMISSION DETERMINATIONS
Pollutant
Category
Dioxins/Furans
Non-D/F Organics:
Volatile Organics
Volatile Organics
(continued)
Sampling
Method '
M0023A
TedlarBag
M0040
for
Organics
with
BP < 30 ฐC
VOST
M0030/0031
for
Organics
withBP
30-100 ฐC
Analysis
Method l
8290
HRGC/HRMS
GC/FID
GC/ECD
8260B
GC/MS
5041A/8260B
GC/MS
Constituents To Be
Determined
2,3,7,8-TCDD 2,3,7,8-TCDF
Total TCDDs Total TCDFs
2,3,7,8-PeCDD 2,3,7,8-PeCDFs
Total PeCDDs Total PeCDFs
2,3,7,8-HxCDD 2,3,7,8-HxCDFs
Total HxCDDs Total HxCDFs
2,3,7,8-HpCDD 2,3,7,8-HpCDFs
Total HpCDDs Total HpCDFs
OCDD OCDF
Alkanes, alkenes, alkynes
Volatile target analyte list, per
Table B3-1
30 TICs, per Section B.5
Volatile target analyte list, per
Table B3-1
30 TICs, per Section B.5
Applicability
All facilities
Optional, but
highly
recommended
All facilities
All facilities
Comments
Method 0023A may be modified to allow
simultaneous sampling and analysis of
PCBs, PAHs, or SVOCs. However,
specific approval is required for this
modification, and a detailed description of
the proposed methodology must be
provided.
Recommended in order to determine
quantities of non-toxic organics to reduce
uncertainty.
Examples: Methane, ethane, ethene,
acetylene, propane, propene, and propyne
Analysis by GC/ECD may also be
considered in order to determine low
concentrations of chlorinated compounds.
Analyze condensate.
Analyze condensate.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-6
-------
TABLE Bl-1
RISK-BASED STACK EMISSION DETERMINATIONS (Continued)
Pollutant
Category . ,
Semivolatile
Organics
Chlorobenzenes/
Chlorophenols
PAHs
PCBs
Aldehydes/
Ketones
Facility-Specific
Compounds
Sampling
Method1
MM5
M0010
for
Organics
with BP
>100 ฐC
MM5
M0010
MM5
M0010
M0023A
M0011
Compound-
specific, see
SectionB. 11
Analysis
Method '
3542/8270C
GC/MS
GC/MS
HRGC/HRMS
CARB 428 2
HRGC/HRMS
Method 16683
HRGC/HRMS
83 ISA
HPLC
Compound-
specific, see
SectionB. 11
Constituents To Be
'Determined
Semivolatile target analyte list,
per Table B4-1
30 TICs, per Section B.5
CBs/CPs in Table B6-1
PAHs in Table B7-1
Total PCBs, based upon summation of
homologue groups listed in Table B8-1
Thirteen dioxin-like congeners listed in
Table B8-1
See M001 1 for list, additional compounds
can be determined
Significant facility-specific compounds
may need to be determined as appropriate:
Pesticides per Table Bll-1
Nitroaromatics
Cyanides
Applicability
All facilities
All facilities
All facilities
Site-specific
determination as
discussed in
Section B.8
Site-specific
determination as
discussed in
Section B. 10
Site-specific
Comments
Method 0010 maybe modified to allow
simultaneous sampling and analysis of
PCBs, PAHs, and D/Fs. However, specific
approval is required for this modification,
and a detailed description of the proposed
methodology must be provided.
Separate analysis of the M0010 extract for
CBs/CPs is recommended.
Separate cleanup of the MM5 extract and
analysis for PAHs by HRGC/HRMS is
recommended.
In order to determine PCBs from the
M0023A sampling train, the condensate
and impinger contents must be retained and
analyzed.
Although HRMS is not specified for total
PCBs in CARB 428, it is recommended in
order to provide lower detection limits.
These additional compounds may need to
be determined as necessary to represent the
wastes burned at a particular facility.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-7
-------
TABLE Bl-1
RISK-BASED STACK EMISSION DETERMINATIONS (Continued)
Pollutant
Category
Total Organic
Mass4
THC/CO
Metals
Particle-Size
Distribution
Hydrogen
Chloride and
Chlorine
Sampling
Method '
TedlarBag
M0040
MM5
M0010
Analysis
Method '
GC/FID
GC/FID
Gravimetric
CEMs - 40 CFR Part 266 Appendix IX
Section 2.0
M0060/
0061
6010A/6020
7000-series
See discussion in Section B. 15
M0050/0051
M9057/9056
Constituents To Be
Determined
Total organic mass (ug) for
organics boiling between (-)160to 100
ฐC
Total organic mass (ng) for
organics boiling between
100-300 ฐC
Total organic mass (ug) for organics
boiling at > 300 ฐC
THC/CO, corrected to 7% oxygen, dry
basis
Al, Co, Cu, Mn, Ni, Se, V, and Zn, in
addition to As, Be, Cd, Cr, Ag, Ba, Hg,
Pb, Sb, Tl
Site-specific particle-size distribution
Hydrogen chloride and chlorine
Applicability
All facilities
All facilities
All facilities
See discussion in
Section B.I 5
All facilities
Comments
A separate M0010 train is required for the
total organic mass determination.
Baseline may be needed for continuous
performance assurance.
Tier 1 metals may be excluded (i.e.,
assumption of no partitioning/ removal).
Notes:
Unless specified otherwise, all sampling and analysis methods are from "Test Methods for Evaluating Solid Waste," SW-846, 3rd ed., as
updated by Updates I, II, HA, IIB, and HI, through December 1996. If detection limits are not low enough to achieve target risk levels, then
modifications to the sampling and analytical procedures may need to be considered. Other methods may be considered provided that the
user can demonstrate the methods meet the data quality objectives for the particular application.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-8
-------
TABLE Bl-1
RISK-BASED STACK EMISSION DETERMINATIONS (Continued)
2 California Environmental Protection Agency, Air Resources Board, CARB Method 428, Sacramento, CA.
3 EPA "Draft Method 1668 Toxic Polychlorinated Biphenyls by Isotope Dilution High Resolution Gas Chromotography/High Resolution
Mass Spectrometry," Office of Science and Technology, Office of Water, March 1997.
4 "Guidance for Total Organics - Final Report," prepared for EPA by Radian Corporation, EPA/600/R-96/033, March 1996. See also
"Determination of Total Organic Emissions from Hazardous Waste Combustors," Larry D. Johnson, Analytical Chemistry, Vol. 68, No.l,
January 1,1996.
See the Appendix B Acronym List for acronym definitions.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-9
-------
B.2 DIOXBVS AND FURANS
All testing for collection of SSRA data is expected to include stack sampling and analysis for D/F
compounds. In order to calculate a toxic equivalent factor (TEQ) in accordance with 40 CFR Part 266,
Appendix IX, Section 4.0 and EPA (1989), the analysis should have sufficient resolution to quantify each
congener with 2, 3, 7, 8 substitutions, as well as total quantities of other remaining congeners for each
homologue. As indicated in Table Bl-1, Method 0023A is used for stack sampling and sample
preparation, followed by analysis in accordance with Method 8290 high resolution gas chromatography/
high resolution mass spectrometry (HRGC/HRMS). Method 0023A may be modified to allow for
simultaneous sampling and analysis of other pollutant classes (e.g., PCBs, PAHs, semivolatile organic
compounds [SVOC], and chlorobenzenes/chlorophenols [CB/CP]). However, specific approval is required
for this modification. Simultaneous determinations from single Modified Method 5 (MM5) samples are
discussed further in Section B.9.
Method 0023 A supersedes Method 23 for Resource Conservation and Recovery Act (RCRA) testing (EPA
1997a). Procedures for addition of isotopically labeled standards to both the XAD-2ฎ sorbent trap and
filter, as well as separate extraction and analysis of the sorbent and filter, have been added in order to
quantify recoveries from each fraction. The labeled surrogate standards are spiked onto the XAD-2ฎ
sorbent prior to sampling, and onto the filter prior to extraction.
As discussed in Section B.I.2, target detection limits for D/Fs should be considered very carefully. Section
6.2.3 of Method 0023 A provides guidance on determining the required sampling time based upon desired
D/F detection limits. Desired detection limits for a particular application may be determined by
performing a preliminary risk assessment prior to testing.
Sample recovery and preparation steps are reviewed briefly in this section because they are critical to the
determination of whether Method 0023A can be modified to provide simultaneous analytical results for
PAHs, PCBs, SVOCs, or CB/CPs. Combined determinations are discussed in more detail in Section B.9.
After sampling, Method 0023 A specifies sequential acetone, methylene chloride, and toluene rinses of the
front half and back half portions of the sampling train. Methods 0023 A allows all of the solvents to be
combined in one container for the front half rinse and another for the back half rinse. However, the
toluene has to be stored separately if SVOCs and CB/CPs are also being determined because some of the
SVOCs could be lost in subsequent preparation steps. Also, if the sampling train is to be analyzed
exclusively for D/Fs, the impinger liquid may be discarded after weight or volume is recorded. However,
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-10
-------
if any other determinations are being made, the condensate and impinger liquid need to be retained and
analyzed.
In the lab, surrogate standards are added to the filter, and internal standards are added to both the
filter/front half and XAD-2ฎ resin/back half fractions. The two fractions are then separately Soxhlet
extracted with toluene and concentrated. One half of each fraction is archived, and the other portion is
solvent-exchanged to hexane and then subjected to three column chromatographic cleanup steps as
described in Method 8290. Finally, after addition of additional standards, the two fractions are analyzed
separately by Method 8290 HRGC/HRMS. High resolution gas chromography/low resolution mass
spectrometry (HRGC/LRMS) as provided hi Method 8280 results in order-of-magnitude higher detection
limits, and is not appropriate for collection of SSRA data. The mass of each isomer from the front half
train fraction is added to that from the back half fraction to obtain a train total before further calculations.
Section 7.4 of Method 0023 A provides a suggested approach for this addition when a measured amount of
an isomer is found in one fraction, but is below the detection limit in the other. Section 7.9.4 of Method
8290 provides guidelines regarding treatment of detection limits in the summation to determine total
homologue concentrations.
Analysis of stack samples for fluorine-, bromine- and sulfur-substituted D/Fs is not anticipated at this time.
Few calibration standards exist to accomplish these analyses, and analytical methods are not yet well
defined. EPA has conducted preliminary studies of chlorinated, brominated, and mixed bromochloro D/Fs
in stack emissions (Lemieux and Ryan 1997a, 1998). However, further research is necessary to better
quantify these compounds and to further develop the appropriate sampling and analytical methodologies
(EPA 1996b, 1996c). The HHRA guidance (EPA in press) recommends that these compounds be
addressed in the uncertainty section of the SSRA.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-ll
-------
B.3 VOLATILE ORGANIC COMPOUNDS
Two sampling methods are needed to determine volatile organic compounds (VOC). These include the
volatile organic sampling train (VOST) (Method 0030/0031) and Tedlar bag (Method 0040) methods.
VOST can typically achieve lower detection limits than the gas bag (i.e., parts per billion [ppb] verses parts
per million [ppm] levels). However, it will not reliably capture very volatile compounds (i.e., those boiling
below 30 ฐC). Thus, the bag sample is needed to augment the VOST results for very volatile compounds.
Attachment 1 of this appendix provides a memorandum clarifying certain aspects of the Method 0040
Tedlar bag method. The VOST samples are thermally desorbed by Method 5041A. The Tedlar bag
samples are introduced into the gas chromatograph through use of a sample loop. The condensate samples
from both the VOST and Tedlar bag determinations can be introduced by direct aqueous injection, or by
Method 5030 purge and trap. All samples (VOST, Tedlar bag, and condensates) should be analyzed by
GC/MS for the Method 8260B volatile target analyte list, summarized in Table B3-1, and TICs as
described in Section B.5. If a compound is detected in both the bag and VOST samples, the higher of the
two results should be used for the SSRA and in the summation of total identified organics. If a compound
is not detected, the lower of the two detection limits should be used.
Individual laboratories are likely to have target analyte lists for volatiles that differ slightly from Method
8260B and Table B3-1. Based upon a review of the lists from several laboratories, the differences are
usually limited to a handful of problem compounds. If a laboratory does not include all of the Table B3-1
compounds on their standard volatile target analyte list, the laboratory should identify the excluded
compounds and explain the reason for exclusion. The permit writer would likely approve use of the
laboratory's standard VOC list in lieu of Table B3-1, after consideration of the laboratory's rationale and a
review to ensure that the excluded compounds have not historically been found in stack emissions, based
on Appendix A-l of the HHRA guidance (EPA in press), Lemieux and Ryan (1997b, 1998), MRI (1997),
and EER (1997).
Table Bl-1 also indicates that Method 0040 bag samples should be analyzed by GC/FID for C1-C4
straight-chain aliphatics (alkanes, alkenes, etc.). As discussed in Section B.I.I, these compounds are
generally non-toxic, but can make up a significant portion of the TO mass. The GC/FID evaluation adds
little to the cost and is highly recommended in order to reduce uncertainty in the SSRA. Analysis by gas
chromatograph/electron capture detector (GC/ECD) may also be considered in order to determine low
concentrations of chlorinated compounds.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-12
-------
TABLE B3-1
VOLATILE TARGET COMPOUND LIST - METHOD 8260B ANALYTES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
, , Volatiles
Acetone2
Acrylonitrile
Benzene
Bromodichloromethane
Bromoform3
Bromomethane
2-Butanone4
Carbon Bisulfide
Carbon Tetrachloride
Chlorobenzene3
Chlorodibromomethane
Chloroethane
Chloroform
Chloromethane
Dibromomethane
Dichlorodifluoroniethane4
1 , 1 -Dichloroethane
1 ,2-Dichloroethane
1 , 1 -Dichloroethene
cis-1 ,2-Dichloroethene
trans- 1 ,2-Dichloroethene
1 ,2-Dichloropropane
cis- 1 ,3-Dichloropropene
trans-l,3-Dichloropropene
Ethylbenzene 3
lodomethane
Methylene Chloride
Styrene 3
1,1,2,2- Tetrachloroethane 3
Tetrachloroethene
Toluene
1,1,1 -Trichloroethane
1 , 1 ,2-Trichloroethane
Trichloroethene
Trichlorofluoromethane
1,2,3-Trichloropropane 3
Vinyl Chloride
Xvlenes (total) 3
CAS Number
67-64-1
107-13-1
71-43-2
75-27-4
75-25-2
74-83-9
78-93-3
75-15-0
56-23-5
108-90-7
124-48-1
75-00-3
67-66-3
74-87-3
74-95-3
75-71-8
75-34-3
107-06-2
75-35-4
156-59-2
156-60-5
78-87-5
10061-01-5
10061-02-6
100-41-4
74-88-4
75-09-2
100-42-5
79-34-5
127-18-4
108-88-3
71-55-6
79-00-5
79-01-6
75-69-4
96-18-4
75-01-4
1330-02-7
Boiling Point^C1
56
78
80
87
149
4
80
46
77
132
119
12
62
-24
97
-30
57
83
32
48
48
95
108
107
136
43
40
145
146
121
110
74
114
87
24
157
-13
137
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-13
-------
TABLE B3-1
VOLATELE TARGET COMPOUND LIST - METHOD 8260B ANALYTES (Continued)
Notes:
Existing sampling methods for volatiles are boiling point specific. The appropriate sampling
methods should be considered to achieve the required volatile target analyte list. For example,
compounds with boiling points less than 30 ฐC cannot be reliably sampled with Method 0030, and
should be sampled using Methods 0031 and/or 0040.
Certain compounds, including acetone, acrylonitrile, and iodomethane, cannot be reliably
determined using the VOST methodology. Therefore, the results for these compounds should be
considered semi-quantitative, at best. However, they have been retained on the target analyte list
in the interest of providing the most complete emissions characterization possible using currently
available methods.
These constituents boil at greater than 121 ฐC. Although they are listed as Method 8260B
analytes, they cannot be reliably sampled with VOST and should be added to the SVOC target
analyte list (Table B4-1).
Two constituents, 2-butanone and dichlorodifluoromethane, have been retained from the former
Method 8240 list, because these compounds have been found routinely in stack emissions.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-14
-------
B.4 SEMIVOLATILE ORGANIC COMPOUNDS
The MM5 sampling train (Method 0010) is used to sample semivolatile organics. In order to determine
SVOCs from the Method 0010 train, the sampling train components are prepared in accordance with
Method 3542. Method 3542 involves separate processing and analysis of three fractions: the filter/front-
half rinse, the XAD-2ฎ resin/back half rinse, and the condensate/condensate rinse. Each fraction should
be analyzed by GC/MS for the Method 8270C semivolatile target analyte list, summarized in Table B4-1,
and TICs as described in Section B.5.
Sample recovery and preparation steps are reviewed briefly in this section because they are critical to the
determination of whether Method 0010 can be modified to provide simultaneous results for D/Fs, PAHs,
PCBs, SVOCs, and CB/CPs (discussed later in Section B.9). After sampling, Method 0010 specifies that a
methanol/methylene chloride solvent be used for the front half and back half rinses. In the lab,
isotopically-labeled standards are added to the filter and XAD-2ฎ resin/back half fractions as specified in
Method 3542, and these fractions are separately Soxhlet extracted with methylene chloride. The front half
rinse is liquid-liquid extracted by separately funnel to recover the methylene chloride layer, which is added
to the filter extract. The condensate/condensate rinse is spiked with standards and is also liquid-liquid
extracted to recover the methylene chloride layer. All three methylene chloride extracts, from the
filter/front-half rinse, the XAD-2ฎ resin/back half rinse, and the condensate/condensate rinse, are
individually concentrated and analyzed by Method 8270C GC/MS. Detection limits for the Table B4-1
compounds vary between analytes, but are generally expected to be in the range of 1 to 10 micrograms per
milliliter ((Xg/mL) of final extract volume. Quantitation limits are on the order of 10 ug/mL.
Individual laboratories are likely to have target analyte lists for semivolatiles that differ slightly from
Method 8270C and Table B4-1. Based upon a review of the lists from several laboratories, the differences
are more extensive than for the volatiles, and could encompass as many as thirty compounds. If a
laboratory does not include all of the Table B4-1 compounds on their standard semivolatile target analyte
list, the laboratory should identify the excluded compounds and explain the reason for exclusion. The
permit writer would likely approve use of the laboratory's standard list in lieu of Table B4-1, after
consideration of the laboratory's rationale and a review to ensure that the excluded compounds have hot
historically been found in stack emissions based on Appendix A-l of the HHRA guidance (EPA in press),
Lemieux and Ryan (1997b and 1998), MRI (1997), and EER (1997).
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-15
-------
TABLE B4-1
SEMIVOLATILE TARGET COMPOUND LIST - METHOD 8270C ANALYTES
Semivolatiles
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
Acenaphthene
Acenaphthylene
Acetophenone
4-Aminobiphenyl
Aniline
Anthracene
Benzidine
Benzoic Acid
Benzo(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(g,h,i)perylene
Benzo(a)pyrene
Benzyl alcohol
Bis(2-chloroethoxy)methane
Bis-(2-chloroethyl)ether
Bis(2-ethylhexyl)phthalate
4-Bromophenyl-phenyl ether
Butylbenzylphthalate
4-Chloroaniline
4-Chloro-3-methylphenol
1 -Chloronaphthalene
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl-phenyl ether
Chrysene
Dibenz (aj) acridine
Dibenzo(a,h)-anthracene
Dibenzofuran
Di-n-butylphthalate
1 ,2-Dichlorobenzene
1 ,3 -Dichlorobenzene
1 ,4-Dichlorobenzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
p-Dimethylaminoazobenzene
7, 1 2-Dimethylbenz(a)anthracene
a,a-Dimethylphenethylamine
2,4-Dimethylphenol
CAS Number
83-32-9
208-96-8
98-86-2
92-67-1
62-53-3
120-12-7
92-87-5
65-85-0
56-55-3
205-99-2
207-08-9
191-24-2
50-32-8
100-51-6
111-91-1
111-44-4
117-81-7
101-55-3
85-68-7
106-47-8
59-50-7
90-13-1
91-58-7
95-57-8
7005-72-3
218-01-9
224-42-0
53-70-3
132-64-9
84-74-2
95-50-1
541-73-1
106-46-7
91-94-1
120-83-2
87-65-0
84-66-2
60-11-7
57-97-6
122-09-8
105-67-9
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-16
-------
TABLE B4-1
SEMIVOLATILE TARGET COMPOUND LIST - METHOD 8270C ANALYTES (Continued)
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
Semivolatiles - "
Dimethylphthalate
4,6-Dinitro-2-methylphenol
2,4-Dinitrophenol
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Di-n-octyl phthalate
Diphenylamine
Ethyl methanesulfonate
Fluoranthene
Fluorene
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopentadiene
Hexachloroethane
Indeno(l,2,3-cd)-pyrene
Isophorone
3-Methylcholanthrene
Methyl methanesulfonate
2-Methylnaphthalene
2-Methylphenol
4-Methylphenol
Naphthalene
1 -Naphthylamine
2-Naphthylamine
2-Nitroaniline
3-Nitroaniline
4-Nitroaniline
Nitrobenzene
2-Nitrophenol
4-Nitrophenol
N-Nitroso-di-n-butylamine
N-Nitrosodimethylamine
N-Nitrosodiphenylamine
N-Nitroso-di-n-propylamine
N-Nitrosopiperidine
2,2'-oxybis (1-Chloropropane)1
Pentachlorobenzene
Pentachloronitrobenzene
Pentachlorophenol
Phenacetin
CAS Number
131-11-3
534-52-1
51-28-5
121-14-2
606-20-2
117-84-0
122-39-4
62-50-0
206-44-0
86-73-7
118-74-1
87-68-3
77-47-4
67-72-1
193-39-5
78-59-1
56-49-5
66-27-3
91-57-6
95-48-7
106-44-5
91-20-3
134-32-7
91-59-8
88-74-4
99-09-2
100-01-6
98-95-3
88-75-5
100-02-7
924-16-3
62-75-9
86-30-6
621-64-7
100-75-4
108-60-1
608-93-5
82-68-8
87-86-5
62-44-2
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-17
-------
TABLE B4-1
SEMIVOLATILE TARGET COMPOUND LIST - METHOD 8270C ANALYTES (Continued)
Semivolatiles
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
Phenanthrene
Phenol
2-Picoline
Proriamide
Pvrene
1 ,2,4,5-Tetrachlorobenzene
2,3,4,6-Tetrachlorophenol
1 ,2,4-Trichlorobenzene
2,4,5-Trichlorophenol
2,4,6-Trichlorophenol
CAS Number
85-01-8
108-95-2
109-06-8
23950-58-5
129-00-0
95-94-3
58-90-2
120-82-1
95-95-4
88-06-2
Additional Constituents from the Method 8260B Target List:
92.
93.
94.
95.
96.
97.
98.
99
Bromoform
Chlorobenzene
Ethylbenzene
Styrene
1 , 1 ,2,2-Tetrachloroethane
Toluene2
1,2,3-Trichloropropane
Xvlenes ("total")
75-25-2
108-90-7
100-41-4
100-42-5
79-34-5
108-88-3
96-18-4
1330-02-7
Notes:
1 Also known by the name bis(2-Chloroisopropyl) ether
2 Toluene is recommended in order to confirm sampling of constituents down to 100 ฐC
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-18
-------
B.5 TENTATIVELY IDENTIFIED COMPOUNDS
For both the Method 8260B and Method 8270C determinations discussed in Sections B.3 and B.4 (i.e.,
VOCs by VOST and Tedlar bag, and SVOCs by MM5), analysis for specific target analytes should be
accompanied by an evaluation of TICs. Besides target analytes, there are generally a number of non-target
components observed in the chromatogram. Attempts to identify and quantify these unknown
chromatographic peaks can improve the percentage of identified organics.
The mass spectrum of the non-target compounds can be searched against a library of mass spectra. A
forward library search selects the largest mass spectral peaks from the unknown spectrum and looks for
library spectra containing those peaks. A reverse search looks for the peaks in the library spectrum that
occur in the unknown spectrum. Any components that are identified are referred to as TICs, since there is
no reference standard analyzed at the same time as the unknown.
Without a compound-specific calibration, TICs can only be quantified using the nearest internal standard
and an assumed relative response factor (RRF) of one (1.0). The resulting concentration is considered
"estimated," due to lack of a compound-specific response factor. The error introduced by an assumed RRF
of 1.0 is unknown and will vary from compound to compound.
The EPA Contract Laboratory Program provides guidelines regarding TIC determinations (EPA 1994).
These are reproduced here as follows:
A library search should be executed for non-target sample components for the purpose of tentative
identification. Up to 30 organic compounds of greatest apparent concentration not listed on the
relevant target compound list, and excluding internal standards and surrogates, should be tentatively
identified via a forward search of the NIST/EPA/NIH (May 1992 release or later) and/or Wiley (1991
release or later), or equivalent mass spectral library. Computer generated library search routines
should not use normalizations which would misrepresent the library or unknown spectra when
compared to each other. Only after visual comparison of sample spectra with the nearest library
searches will the mass spectral interpretation specialist assign a tentative identification. Guidelines
for making tentative identification are as follows:
Relative intensities of major ions in the reference spectrum (ions greater than 10% of
the most abundant ion) should be present in the sample spectrum.
The relative intensities of the major ions should agree within +/- 20%.
Molecular ions present in the reference spectrum should be present in the sample
spectrum.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-19
-------
Ions present in the sample spectrum but not in the reference spectrum should be
reviewed for possible background contamination or presence of co-eluting
compounds.
Ions present in the reference spectrum but not in the sample spectrum should be
reviewed for possible subtraction from the sample spectrum because of background
contamination or co-eluting compounds. Data system library reduction programs can
sometimes create these discrepancies.
If, in the judgement of the mass spectral interpretation specialist, no valid tentative identification can
be made, the compound should be reported as "unknown". The mass spectral specialist should give
additional classification of the unknown compound, if possible (i.e., unknown aromatic, unknown
hydrocarbon, unknown chlorinated compound). If probable molecular weights can be distinguished,
these should also be included.
An estimated concentration for tentatively identified non-target compounds should be determined by
the internal standard method. For quantitation, the nearest internal standard free of interferences
should be used, using an assumed relative response factor of one (1). The resulting concentration
should be qualified as "estimated", due to lack of a compound-specific response factor, and as
"presumptive evidence of presence", indicating the quantitative and qualitative uncertainties
associated with the non-target component. An estimated concentration should be calculated for all
tentatively identified compounds, as well as those identified as unknowns.
Where possible, additional standards containing TICs should be prepared and analyzed to confirm
identification. A multiconcentration calibration with the additional standards can be used to establish
RRFs specific to each compound, and the semivolatile extracts can be re-analyzed to enhance quantitative
accuracy. For volatiles, the RRFs can be used to re-quantify prior analyses to enhance quantitative
accuracy.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-20
-------
B.6 CHLOROBENZENES/CHLOROPHENOLS
Some labs offer an expanded GC/MS target analyte list for CB/CPs, similar to Table B6-1. Although there
is some overlap with the Table B4-1 SVOC constituents, the CB/CP list is more extensive and is
recommended. The CB/CP analysis uses the same extracts prepared for the SVOCs, as discussed in
Section B.9. Therefore, dual analysis for both SVOCs and CB/CPs should not adversely impact the
detection limits for either determination. If a CB/CP compound is detected in both the SVOC and CB/CP
analyses, the higher of the two results should be used for the SSRA and in the summation of total identified
organics. If a compound is not detected, the lower of the two detection limits should be used.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-21
-------
TABLE B6-1
CHLOROBENZENES AND CHLOROPHENOLS
Chlorobenzenes
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
1,3,5-Trichlorobenzene
1 ,2,4-Trichlorobenzene
1 ,2,3-TrichIorobenzene
1 ,2,3,5-Tetrachlorobenzene1
1 ,2,4,5-Tetrachlorobenzene1
1 ,2,3,4-Tetrachlorobenzene
Pentachlorobenzene
Hexachlorobenzene
CAS
Number
95-50-1
541-73-1
106-46-7
108-70-3
120-82-1
87-61-6
634-90-2
95-94-3
634-66-2
608-93-5
118-74-1
Chlorophenols
2-Chlorophenol
3-Chlorophenol2
4-Chlorophenol2
2,4-Dichlorophenol
2,5-Dichlorophenol
2,3-Dichlorophenol
2,6-Dichlorophenol
3,5-Dichlorophenol
3 ,4-Dichlorophenol
2,3 ,5-Trichlorophenol
2,4,6-Trichlorophenol
2,4,5-Trichlorophenol
2,3 ,4-Trichlorophenol
2,3 ,6-Trichlorophenol
2,3,5,6-
Tetrachlorophenol3
2,3,4,5-
Tetrachlorophenol3
Pentachlorophenol
CAS
Number
95-57-8
108-43-0
106-48-9
120-83-2
583-78-8
576-24-9
87-65-0
591-35-5
95-77-2
933-78-8
88-06-2
95-95-4
15950-66-0
933-75-5
935-95-5
490-151-3
87-86-5
Note:
Co-elute, reported as totals
Co-elute, reported as totals
Co-elute, reported as totals
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-22
-------
B.7 POLYCYCLIC AROMATIC HYDROCARBONS
Although the Method 8270C list (Table B4-1) includes most PAHs, detection limits for PAHs have been
found to be critical for the indirect risk pathway. Therefore, it is recommended that PAHs be determined
by separate HRGC and either HRMS or LRMS analysis after the application of chromatography cleanup
procedures designed to remove interfering organics. If a PAH compound is detected in both the SVOC and
PAH analyses, the higher of the two results should be used for the SSRA and in the summation of total
identified organics. If a compound is not detected, the lower of the two detection limits should be used.
California Environmental Protection Agency, Air Resources Board (CARB) Method CARB 429 (CARB
1997) provides guidelines for sampling and analysis of the PAHs listed in Table B7-1. Method CARB 429
specifies sample collection with a MM5 sampling train. After sampling, the train components are
sequentially rinsed with acetone, hexane, and methylene chloride. The filter, XAD-2ฎ resin, concentrated
rinses, and condensate are Soxhlet extracted with methylene chloride, as are the impinger contents/rinses.
The methylene chloride extracts are concentrated, solvent exchanged to hexane, and subjected to a silica
gel or alumina column clean up.
Following the cleanup, the extracts are analyzed by HRGC, and either isotope dilution HRMS or LRMS.
The decision regarding specific procedures will depend on the detection limits which are determined to be
necessary for PAHs based upon a preliminary risk evaluation. PAHs can be analyzed by LRMS with usual
detection limits ranging from 1 to 5 micrograms (ng) per fraction. PAHs can be analyzed by HRMS with
usual detection limits ranging from 10 to 50 nanogram (ng) per fraction. Actual detection limits are limited
by the background level of PAHs observed in the XAD-2ฎ resin.
The Method CARB 429 PAH procedures can likely be successfully merged with either the Method 0023 A
D/F procedures or the Method 0010/3542 SVOC procedures with careful planning. An extract that has
been subject to chromatography cleanup procedures for PAHs would not be suitable for any other
semivolatile determination. However, absent this final step, the sample preparation and extraction can be
designed to achieve multiple semivolatile determinations. This is discussed further in Section B.9.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-23
-------
TABLE B7-1
POLYCYCLIC AROMATIC HYDROCARBONS
PAH
Acenaphthene
Acenaphthylene
Anthracene
Benzo(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(g,h,i)perylene
Benzo(a)pyrene
1-Chloronaphthalene
2-ChloronaphthaIene
Chrysene
Dibenz(aJ)acridine
Dibenzo(a,h,)anthracene
Dibenzo(a,e)pyrene
7, 1 2-Dimethylbenz(a)anthracene
Fluoranthene
Fluorene
Indeno( 1 ,2,3 -cd)pyrene
3-Methylcholanthrene
2-Methylnaphthalene
Naphthalene
Phenanthrene
Pyrene
:.: .:;;;; .'0-CAS.Number.
83-32-9
208-96-8
120-12-7
56-55-3
205-99-2
207-08-9
191-24-2
50-32-8
90-13-1
91-58-7
218-01-9
224-42-0
53-70-3
192-65-4
57-97-6
206-44-0
86-73-7
193-39-5
56-49-5
91-57-6
91-20-3
85-01-8
129-00-0
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-24
-------
B.8 POLYCHLORINATED BBPHENYLS
The need for sampling and analysis of PCBs requires careful consideration and is discussed in detail in the
HHRA guidance (EPA in press). Determinations regarding the need for PCB analysis may become more
straight forward as additional data are collected and evaluated. However, in the interim, the following
general guidelines may be considered in conjunction with site-specific concerns:
PCBs should automatically be tested at combustion units that burn PCB-contaminated
wastes or waste oils
PCBs should be tested at units that burn highly variable waste streams, such as municipal
or commercial wastes, for which PCB contamination is reasonable
PCBs should be tested at facilities that burn highly chlorinated waste streams
In regards to PCB testing at other hazardous waste combustion facilities, the HHRA guidance (EPA in
press) states, "due to the toxicity and uncertainties associated with combustion chemistries the permitting
authority may choose to confirm the absence of these compounds from stack emissions via stack gas
testing for units burning hazardous wastes." For any facility where uncertainty regarding PCB formation
may be cause for concern, a PCB determination can be easily added to the D/F test procedures.
Method CARB 428 (CARB 1990a), which provides guidelines on sampling and recovery procedures for
PCB determinations, specifies sampling with a MM5 sampling train followed by Soxhlet extraction with
toluene. Since Method 0023A for D/Fs also requires toluene extraction, PCBs and D/Fs may be
determined from the same Method 0023 A sampling train. However, if both D/Fs and PCBs are to be
determined from the same sampling train, it is necessary to split the sample after extraction for different
preliminary fractionation and cleanup procedures. PCB cleanups typically involve silica gel or florisil. In
addition, the condensate and impinger contents must be retained and analyzed. This is discussed further in
Section B.9.
PCBs are often discussed in terms of commercial mixtures, or Arochlors. However, PCBs change in the
environment, and their compositions differ from the commercial mixtures. Characterization of combustion
emissions in terms of Arochlors, for example by Method 8082, would be both imprecise and inappropriate.
EPA's current toxicity approach (1996e) requires data on (1) the total PCB concentration and (2)
congener-specific analyses for the 13 toxic dioxin-like coplanar and mono-ortho-substituted PCBs listed in
Table B8-1. Dioxin TEQs are applied to the congener-specific concentrations to evaluate dioxin-like
toxicity. Risks from the dioxin-like congeners (evaluated using dioxin TEQs) are then added to risks from
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-25
-------
TABLE B8-1
POLYCHLORINATED BIPHENYLS
Dioxin-Like Coplanar PCBs
S^S'^'-Tetrachlorobiphenyl
2,3,4,3',4'-Pentachlorobiphenyl
2,3 ,4,5,4'-Pentachlorobiphenyl
2,4,5,3',4'-Pentachlorobiphenyl
3,4,5,2',4-Pentachlorobiphenyl
3,4,5,3',4'-Pentachlorobiphenyl
2,3,4,5,3',4I-Hexachlorobiphenyl
2,3,4,3',4'>5'-Hexachlorobiphenyl
2,4,5,3'.4l,5'-Hexachlorobiphenyl
SAS^'^'.S'-Hexachlorobiphenyl
2,3,4,5,2',3',4'-Heptachlorobiphenyl
2,3,4>5,2I,4',51-Heptachlorobiphenyl
2J3,4,5,3'J4',5'-Heptachlorobiphenyl
IUPAC77
IUPAC 105
IUPAC114
IUPAC 118
IUPAC 123
IUPAC 126
IUPAC 156
IUPAC 157
IUPAC 167
IUPAC 169
IUPAC 170
IUPAC 180
IUPAC 189
CAS Number
32598-13-3
32598-14-4
74472-37-0
31508-00-6
65510-44-3
57465-28-8
38380-98-4
68782-90-7
52663-72-6
32774-16-6
35065-30-6
35065-29-3
39635-31-9
Total Homologue Groups (Sum to Determine Total PCBs)
Monochlorobiphenyls
Dichlorobiphenyls
Trichlorobiphenyls
Tetrachlorobiphenyls
Pentachlorobiphenyls
Hexachlorobiph6nyls
Heptachlorobiphenyls
Octachlorobiphenyls
Nonachlorobiphenyls
Decachlorobiphenyls
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-26
-------
the rest of the mixture (evaluated using slope factors applied to total PCBs reduced by the amount of
dioxin-like congeners). The HHRA guidance (EPA in press) provides more information on this issue.
Method CARB 428 provides guidelines for analysis to determine the total homologue group concentrations
listed in Table B8-1, which can be summed to provide a total PCB concentration. Knowledge of the PCB
distribution by homologue group is necessary in order to determine which toxicity value to apply. Method
CARB 428 calls for analysis by HRGC/LRMS, giving target detection limits on the order of 0.1 to 1.0 ug
per sample per homologue group when both D/F and PCBs are determined from the same sample.
However, since PCBs have been found to be risk drivers, CARB 428 should be used with HRMS to
provide PCB homologue group concentrations at lower detection limits.
Method CARB 428, as written, will not provide information on the 13 co-planar PCBs. However,
guidelines on isotope dilution HRGC/HRMS analysis for the thirteen co-planars have been published as
Draft Method 1668 (EPA 1997b), and Method 1668 determinations are already being provided at some
commercial laboratories with high resolution GC/MS capabilities.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-27
-------
B.9 ANALYSIS OF MM5 SAMPLES FOR MULTIPLE POLLUTANT CLASSES
As discussed in the previous sections, the MM5 sampling train is used to sample several different
categories of semivolatile organics:
Section B.2 - Dioxins/furans
Section B.4 - Semivolatile organic compounds
Section B.6 - Chlorobenzenes/chlorophenols
Section B.7 - Polycyclic aromatic hydrocarbons
Section B.8 - Polychlorinated biphenyls
The sample recovery and preparation steps associated with the various SVOC, CB/CP, PAH, and PCB
determinations significantly influence whether the determinations can be conducted from a single Method
0010 sampling train, or in conjunction with the Method 0023 A D/F train. This topic is discussed in more
detail in this section.
In some cases (e.g., sampling platform constraints, sample port constraints, to increase sampling efficiency,
etc.) it may be necessary to analyze a single MM5 train sample for multiple pollutant classes. This could
encompass all of the semivolatile determinations discussed previously, D/Fs, SVOCs, CB/CPs, PAHs, and
PCBs. In this situation, all aspects of the sampling and analysis must be considered carefully to ensure that
the resulting data will not be invalid for one or more of the compound classes. A separate MM5 train must
always be run exclusively for TOs, as discussed later in Section B.12.
Procedures for multiple determinations are subject to approval by the regulatory agency. The facility
should provide a detailed description in the trial burn plan of how the sampling analysis for multiple
compounds will be performed. Documentation should include detailed information on sampling, recovery,
spiking, analysis, quality assurance and quality control (QA/QC) procedures, and anticipated impacts on
detection limits.
Potential liabilities associated with combined semivolatile determinations are discussed in detail in
Johnson (1995). The following briefly summarizes key information from that paper.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-28
-------
Detection Limits:
Increased detection limits can occur as a result of combining semivolatile determinations. It is often
necessary to split the sample for separate processing through two different preparatory procedures. If the
sample is split into two equal parts, each may have its detection limit doubled.
Sample Preparation:
The choice of solvents is critical. If methylene chloride is combined with toluene or other higher boiling
point solvent, it may become difficult to concentrate the resulting extract without potential loss of the more
volatile compounds.
Cleanup procedures are also critical. Many analytical methods include cleanup procedures, such as
column chromatography or gel-permeation chromatography, designed to remove interfering organics but to
have minimal effect on the target analytes. When combining determinations, it is important to ensure that
the target analytes will not be removed with the unwanted compounds. Determinations for D/Fs, PAHs,
and PCBs each involve specific cleanup procedures, which make the resulting extract unsuitable for any
other determinations.
Standards:
In combining methods, it is important to consider whether internal standards, surrogate standards, or
recovery standards could be a source of incompatibility between methods. Assistance should be obtained
from a well qualified analytical chemist who understands the methods and calculations involved.
Alternatively, standards can serve a beneficial function when multiple methods are combined because
standards can be used to ensure that losses are not incurred. For example, isotopically-labeled D/F,
SVOC, PAH, and PCB standards can be used at various stages as follows:
Standards can be added to the sampling module, to gauge potential losses during sampling
Standards can be added prior to sample recovery, to gauge losses during extraction,
concentration, and cleanup
Standards can be added immediately prior to analysis, primarily to compensate for
instrument response changes
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-29
-------
Although most of the methods already utilize these types of standards, additional standards should be
added as needed to address concerns associated with combined methods.
The importance of consulting a qualified analytical chemist on these issues cannot be overstated.
However, procedures which have been used by some commercial labs are summarized below in order to
highlight some of the basic considerations.
Determinations from One Sampling Train
All determinations (D/F, SVOCs, CBs/CPs, PAHs, and PCBs) can be made from a single Method 0023A
sampling train, if necessary. However, the toluene train rinses from the D/F train recovery have to be
stored separately, the condensates and impinger contents have to be retained and analyzed, and two Soxhlet
extractions are required.
Procedures used by one lab illustrate these points. This lab performs the first Soxhlet extraction with
methylene chloride. The resulting extract is split in half, one half for the D/F determination and the other
half for everything else. The toluene rinse is added to the remaining Soxhlet extractor contents, and a
second extraction occurs with toluene. Half of the resulting toluene extract is combined with the half
methylene chloride extract, and the combined extract is subjected to D/F cleanup and analysis (the other
half of the toluene extract is not used). From the remaining methylene chloride extract, one portion is
subjected to PAH cleanup and analysis, one portion is subjected to PCS cleanup and analysis, and one
portion can be analyzed directly for SVOCs and/or CB/CPs.
The D/F detection limit is not compromised by this methodology, because Method 0023A already calls for
a 1:1 split of the extract (half is archived). The combined methodology essentially uses part of the archive
for the other analyses. The SVOC and CB/CP detection limits are doubled by this methodology (a dilution
factor of 2x must be applied, because of the 1:1 methylene chloride extract split). The PAH and PCB
detection limits may or may not be affected, depending on the portions removed for clean up and the final
concentrated volumes prior to analysis.
Alternatives to this methodology exist. Since Method CARB 428 requires a toluene extraction for PCBs,
PCBs would preferably be determined with D/Fs from the combined toluene/methylene chloride extract
instead of with SVOCs from the straight methylene chloride extract. One portion would be used for D/F
cleanup and analysis, and another portion would be removed for PCB clean up and analysis.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-30
-------
Determination from Two Sampling Trains
If two sampling trains can be used, a reasonable combination would include D/Fs and PCBs from a Method
0023 A train, and all other determinations from a second Method 0010 train. The D/F and PCB
combination is natural, because both components are amenable to toluene extraction as described in
Method CARS 428.
In order to determine PCBs from Method 0023 A, it is recommended that at least four isotopically labeled
PCB surrogates be spiked onto the XAD-2ฎ resin together with the D/F surrogates prior to sampling. The
amount of surrogates and internal standards added to the samples should be adjusted to compensate for the
additional analysis (Ryan 1998). The Method 0023A train components would be subject to a single
Soxhlet extraction with toluene, and the extract would be split. One portion would be used for D/F clean
up and analysis, and another portion would be removed for PCB cleanup and analysis. The D/F detection
limit would not be compromised by this methodology, and the PCB detection limit would depend on the
portion removed for clean up and the final concentrated volume prior to analysis.
The second MM5 train, Method 0010, would be subject to a single Soxhlet extraction with methylene
chloride. One portion would be subjected to PAH cleanup and analysis, and another portion could be
analyzed directly for SVOCs and/or CB/CPs. The detection limits for SVOCs and CB/CPs would not be
compromised by this methodology. The detection limits for PAHs would depend on the portion removed
for cleanup, and the final concentrated volume prior to analysis.
PAHs
With either method, the PAHs could possibly be determined with the D/F extract, because toluene is an
acceptable alternative extractant for PAHs (Johnson 1995). Care would have to be taken not to take the
extract to dryness, because PAH vapor pressures could cause significant losses with complete evaporation.
Also, the sample extract would have to be split for separate D/F cleanup and analysis, PCB cleanup and
analysis, and PAH cleanup and analysis.
Clean Up Procedures
Determinations for D/Fs, PAHs, and PCBs each involve specific cleanup procedures. An extract which has
been subjected to any of these cleanups should not be analyzed for any other pollutants. However, absent
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-31
-------
the final cleanup steps, the extraction and preparation for all three determinations can probably be
combined.
Multiple Analysis of Single Extracts
SVOCs and CB/CPs, and perhaps other semivolatiles, can always be determined simultaneously with no
compromise in detection limits. This holds true for any determinations that simply involve multiple
injections of a single extract, with no splitting during preparation.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities B-32
-------
B.10 ALDEHYDES/KETONES
The final category of specific organic analytes is aldehydes/ketones (A/K). Some A/Ks are already
included on the Table B3-1 and B4-1 constituent lists. These have been retained on Tables B3-1 and B4-1
because the VOST, Tedlar bag, and MM5 methods often provide information on A/K concentrations.
However, in order to reliably determine A/Ks, specific sampling and analytical procedures are required.
Reliable A/K determinations require a Method 0011 sampling train. The stack gases are sampled
isokinetically and collected in aqueous acidic 2,4-dinitrophenylhydrazine solution. A dinitrophenyl-
hydrazone derivative is formed, which is then extracted, solvent-exchanged, concentrated, and analyzed by
high performance liquid chromatography (HPLC) according to Method 8315. This methodology can
determine concentrations of the following A/Ks, although many laboratories have extended the method to
other aldehydes and ketones:
Formaldehyde
Acetaldehyde
Acetophenone
Isophorone
Propionaldehyde
The need for Method 0011 A/K sampling should be considered carefully. Stack sampling port and
platform restrictions will likely limit the number of isokinetic sampling trains that can be run at one time.
Unless a facility burns large quantities of A/K wastes, the facility may wish to submit a justification for use
of A/K data from a similar facility, or for A/K data collected previously at the same facility. Also, as
stated previously, the VOST, Tedlar bag, and MM5 methods are likely to provide some A/K information,
although the data would not be of the same quality as that from Methods 0011/8315.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-33
-------
B.11 FACILITY-SPECIFIC COMPOUNDS
The previous sections identified target analyte lists that are generally applicable to all facilities. However,
it may be necessary for individual facilities to sample and analyze for additional compounds based upon
the wastes that they burn. Potential candidates for additional facility-specific determinations include any
highly toxic, persistent, or bioaccumulative constituents burned in large quantities. These could
encompass compounds such as pesticides, nitroaromatics, and cyanides. These constituents would be in
addition to, and not substitutes for, the target analyte lists described previously in this appendix. Table
Bl 1-1 lists organochlorine pesticides which can be determined by Method 8081A GC/ECD, together with
the waste codes which might contain these compounds.
Other facility-specific considerations might include DO 17 waste (2,4,5-TP [Silvex]), DO 16 and U240 waste
(2,4-D), K025 waste (1,3-Dinitrobenzene), and cyanide wastes (F007-F012, F019, K007, KOI 1, K013,
K027, K060, K088, and U246). Method 8151A provides analytical procedures for chlorinated herbicides.
Method 8270C can provide data on nitroaromatics. Method CARS 426 (CARB 1991) can be used for
determination of cyanide emissions.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities B-34
-------
TABLE Bll-1
ORGANOCHLORINE PESTICIDES - METHOD 8081A ANALYTES
Compound
Aldrin
a-BHC
P-BHC
Y-BHC (Lindane)
8-BHC
Chlorobenzilate
a-Chlordane
y-Chlordane
Chlordane - mixed isomers
1 ,2-dibromo-3-chloropropane
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate
Dieldrin
Endosulfan I
Endosulfan II
Endosulfan sulfate
Endrin
Endrin aldehyde
Endrin ketone
Heptachlor
Heptachlor epoxide
Hexachlorobenzene
Hexachlorocyclopentadiene
Isodrin
Methoxychlor
Toxaphene
CAS Number
309-00-2
319-84-6
319-85-7
58-89-9
319-86-8
510-15-6
5103-71-9
5103-74-2
57-74-9
96-12-8
72-54-8
72-55-9
50-29-3
2303-16-4
60-57-1
959-98-8
33213-65-9
1031-07-8
72-20-8
7421-93-4
53494-70-5
76-44-8
1024-57-3
118-74-1
77-47-4
465-73-6
72-43-5
8001-35-2
Waste Codes
P004
D013, U129
D013,U129
D013, U129
D013, U129
U038
D020, K097, U036
D020, K097, U036
D020, K097, U036
U066
U060,U061
U060, U061
U060, U061
U062
P037
P050
P050
P050
D012, P051
D012,P051
D012,P051
D031,K097,P059
D031,K097,P059
D032
U130
P060
D014,U247
D015,K041,K098,P123
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-35
-------
B.12 TOTAL ORGANICS
The importance of determining the TO mass being emitted from the stack has been discussed previously in
Section B.I.I. Procedures for determining TO mass are outlined in a final EPA (1996d) report dated
March 1996. Although other methods for measuring TOs are often proposed, a discussion regarding why
these other methods, including THC monitoring, are inappropriate for the TOs determination is found in
Johnson (1996a).
TOs, as determined by the published methodology, means the total amount of organic material which is
recoverable by means of solvent extraction or other preparatory steps used in the survey analysis. The
results are reported as "ng total organic per m3." This result can be compared directly to the summed mass
of all of the identified constituents. It is not necessary, or actually appropriate, to convert the TO result to
another basis, such as ppm carbon or ppm propane.
In order to determine the unidentified organic mass, the masses of specific quantified toxic compounds are
subtracted from the results of the TO determination. This would include compounds such as D/Fs, PAHs,
PCBs, SVOCs, VOCs, and A/Ks. The mass of organic material remaining after correction for identified
compounds is referred to as the "unspeciated (or unidentified) organic mass." The HHRA guidance (EPA
in press) provides more detail regarding the use of this information in the SSRA.
The TO procedure is based on the determination of organics in three boiling point ranges (1) the volatile,
field gas chromatography (FGC) fraction (boiling point less than 100 ฐC), (2) the semivolatile, total
chromatographable organics (TCO) fraction (boiling point between 100 and 300 ฐC), and (3) the
nonvolatile, gravimetric (GRAY) fraction (boiling point greater than 300 ฐC). The sum of the constituent
concentrations in the three fractions represents the estimated TOs.
The TO procedure involves two separate sampling procedures. The FGC fraction is collected using a
Tedlar bag as described in Method 0040. The TCO and GRAY fractions are collected during the same
sampling period as the FGC fraction by using a separate Method 0010 sampling train. The operation of the
Method 0010 train for TO will be the same as that used for the SVOCs. However, a separate Method 0010
train must be run exclusively for TO. The MM5 trains used for sampling specific semivolatile constituents
(D/F, SVOCs, etc.) cannot be used for the TOs determination, because the addition of standards to those
trains would register as TOs and would complicate interpretation of results.
Each fraction is discussed in more detail below.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities B-3 6
-------
Field Gas Chromatography Fraction (boiling points < 100 ฐC)
The FGC fraction is collected using a Tedlar bag, Method 0040. The analysis procedures are normally
performed in the field to minimize sample loss due to storage and shipping. The analysis is by GC/FID,
and results are reported in Table B12-1 as seven subcategories according to boiling point range.
Total FGC organics are determined by summing the peak areas, converted to concentration values in
micrograms per cubic meter (ug/m3), in each retention time window. It is strongly recommended that
methane be determined as an individual compound during the FGC analysis. Methane can be a major
component of the volatile organic emissions, and it is beneficial to have its mass in the assigned side of the
organics material balance rather than as part of the unidentified component (Johnson 1996a).
The condensate fraction from the Method 0040 sampling is normally transferred to a zero-headspace vial
and shipped to the laboratory for analysis. The condensate is analyzed by purge and trap GC/FID, based
on calibration with a C5-C7 mixture. Results can be reported as C5, C6, and C7 fractions.
Total Chromatographable Organics Fraction (boiling points 100-300 ฐC)
The TCO and GRAY fractions are both collected using a single MM5 sampling train, SW-846 Method
0010. Organics are recovered from the various components of the sampling train according to Method
3542. After recovery, the three resulting methylene chloride extracts are combined before analysis for
TCO and GRAY.
TCO analysis is a form of low-resolution GC/FID, where the area under the response curve is integrated
between boiling points 100-300 ฐC (e.g., by summing peak areas in the C7-C17 range). For purposes of
comparison, "hot" THC continuous emission monitor (CEM) samples are only required to be maintained
above 150 ฐC, and do not involve filter analysis.
Gravimetric Mass Fraction (boiling points > 300 ฐC)
The GRAY fraction is the only fraction for which the mass is determined directly. The GRAY procedure
is carried out by analysis of an aliquot of the same methylene chloride extract as was used for the TCO
determination. The aliquot is placed in a weighing pan, allowed to dry, and weighed. The mass (in ug) is
recorded as the GRAY value. When divided by the sample volume, the result in ug/m3 can be added to the
FGC and TCO values for the TOs determination.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
' August 1998 Peer Review Draft
B-37
-------
Because high field blank results for the gravimetric portion have been reported, trouble-shooting measures
have been identified to minimize potential sources of contamination (EPA 1997c). This section provides a
summary of these "lessons learned" for the gravimetric determination. In order to obtain the most accurate
results possible for the nonvolatile portion of the TO emissions, it is important to ensure that the XAD-2ฎ
resins used in the gravimetric analysis are clean. Field results have shown high blank results that were
attributed to old and dirty XAD-2ฎ resins. It is recommended that only XAD-2ฎ resins that have been
recently cleaned (preferably no longer than 14 days prior to sample analysis) be used to analyze trial burn
samples for the purpose of generating data for an SSRA. The following gravimetric laboratory procedures
are recommended for minimizing sources of contamination:
Assure all glassware and equipment have been cleaned thoroughly with high quality
reagents
Use high quality reagents for performing procedures (extractions, rinses, etc.);
"Ultrapure" reagents are recommended
Cover the pan that composite extracts are transferred to for drying, by building a tent with
aluminum foil, shiny side out
Run a control pan
Check calibration on balance prior to each weighing
Use a balance precise to 10 ug
Take extra care in handling XAD-2ฎ resin (make sure that the resin does not float out of
the extraction thimble)
Run three non-filtered and filtered extracts through gravimetric measurements (using
filters specified in Method 0010) and compare the results to determine if a carry through
problem may exist; as a rule of thumb, if the XAD-2ฎ resin and condensateftack-half
rinse extract exceeds the criteria for clean XAD-2ฎ in Method 0010, Appendix A, the
resin is the likely source of the problem
Confirm that the clean XAD-2ฎ resin used in the laboratory and in the field meets Method
0010, Appendix A requirements
Reconstitute the gravimetric sample from the extract composite and run it on the GC/MS
for rhombic sulfur; also, run TO extract prior to performing the procedure for rhombic
sulfur on the GC/MS (This should only be applied to actual samples, not blanks, where
there can be significant sources of sulfur in the fuels or wastes.)
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-38
-------
TABLE B12-1
TOTAL ORGANICS FGC ANALYSIS
Compound
Methane
Ethane
Propane
Butane
Pentane
Hexane
Heptane
Boiling Point
- , "C ' -
-161
-88
-42
0
36
69
98
Reporting Range
$' - .
-160 to -100
-100 to -50
-50 to 0
OtoSO
30 to 60
60 to 90
90 to 98
Report As
i
Cl
C2
C3
C4
C5
C6
C7
1 Comparisons
Cold THC CEM:
Boiling point <
0ฐC
VOST:
Boiling Point
30-100 ฐC
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-39
-------
Attempts to more specifically characterize the GRAY fraction have met with mixed results. The GRAY
fraction may include organic and/or inorganic mass not directly attributable to organic incinerator
emissions. These artifacts may be composed of inorganic salts, super-fine particulate, fractured XAD-2ฎ
resin, or some other unknown (Lemiuex and Ryan 1998).
The TO measurement is an estimate. For both the FGC and TCO fractions, the estimated mass is
calculated based on detector responses to specific compounds that are presumed to be representative of the
specific fraction. However, the TO measurements are strongly believed to be the best currently available
procedure for generating a TOs analysis for the purpose of indicating uncertainty due to the fraction of
organics that have not been quantified. TO measurements may be very conservative measurements of
uncertainty because the GRAY fraction may contain inorganic species.
It is recommended that the individual boiling point category and subcategory data from the various
components of a TO determination be reported, since this information may be very useful during later
interpretation. For example, the unidentified mass in the GRAY range cannot be due to vinyl chloride, just
as the unidentified material in the FGC analysis cannot be dioxin or PAH compounds (Johnson 1996a).
Estimates of compounds that are potentially associated with the three TO fractions are summarized from
MRI (1997). The FGC fraction would be expected to contain lighter hydrocarbons and halogenated
alkanes and alkenes. This could include compounds such as methane and halogenated ethanes, ethenes,
and propanes. The TCO fraction will contain a wide range of semivolatile compounds that could include
compounds such as D/Fs, phthalates, phenols, halogenated aromatics, and nitrogenated and sulfonated
compounds. The GRAY portion is extremely difficult to analyze. However, it would be expected to
contain high molecular weight organics of C17 or greater, D/Fs, many PAHs, and high molecular weight
organic acids and salts. Potential contaminants in the GRAY fraction include salts. One attempt to
characterize the gravimetric fraction of reaction products from D/Fs sorbed on a calcium-based sorbent
involved thin-layer chromatography separation followed by multiple analytical techniques (Gullett and
others 1997). This showed the gravimetric portion from that particular research to be higher molecular
weight, chlorinated compounds with both aromatic and aliphatic components. The author noted that these
higher molecular weight products would not likely be detected by conventional GC/MS analysis.
Lemieux and Ryan (1997b, 1998) have suggested a number of techniques which may be explored to
identify more of the TOs, with particular emphasis on the semivolatile and non-volatile (TCO and GRAY)
fractions. Each sample would be segregated or fractionated based on polar characteristics using HPLC.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-40
-------
Then, each fraction could be analyzed by GC/MS as well as MDGC/MS. These recommendations involve
innovative techniques and would first need to be demonstrated at lab-scale test facilities prior to field use.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-41
-------
B.13 TOTAL HYDROCARBON AND CARBON MONOXIDE CONTINUOUS EMISSIONS
MONITORS (CEM)
Even if THC monitoring is not required by current regulations at a facility, it should be performed in
conjunction with the organic emissions determinations during risk testing. This can be accomplished by
having a temporary monitor brought in during the testing, if necessary. Performance specifications for
THC monitoring are provided at in 40 CFR Part 266, Appendix IX, Section 2.2.
Although THC monitoring is not adequate for determining TO mass, as discussed in Section B. 1.1, it may
prove to be a useful tool for continuous performance assurance. Both THC and carbon monoxide (CO) can
be indicators of whether good combustion practice is being maintained, and whether organics emissions
may have changed from the baseline determined during the risk testing.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-42
-------
B.14 METALS
As indicated in Table B1-1, the additional metals which will be evaluated in SSRAs (i.e., aluminum,
cobalt, copper, manganese, nickel, selenium, vanadium, and zinc) can all be collected in the Method 0060
multiple metals train with the 10 traditional metals. Thus, there are no new sampling considerations
related to total metals.
Although total metals determinations are not substantially impacted by SSRAs, metals chemical speciation
data, especially for mercury, represents a new data need. The HHRA guidance (EPA in press) provides
default assumptions for ratios of divalent particulate mercury, divalent mercury vapor, and elemental
mercury vapor. However, these default assumptions could be replaced by site-specific data if mercury
speciation sampling were performed.
Based on a preliminary review of mercury partitioning data from the Method 0060 train components, it
would appear that the HHRA (EPA in press) default assumptions probably over-estimate particulate
mercury for most hazardous waste combustion sources, and underestimate elemental mercury vapor for
cement kilns. If this is the case, then speciated mercury data could potentially reduce the estimated risk for
cement kilns by reducing the local deposition of divalent mercury. The impact for other facilities would be
more difficult to predict, but it is likely that local deposition rates would decrease if speciated mercury data
showed less divalent particulate and more divalent vapor.
A modified mercury speciation sampling train has recently been developed under direction of the National
Exposure Research Laboratory (Giglio 1997). This train, the alkaline mercury speciation (AMS) method,
uses dilute sodium hydroxide impingers for capturing divalent mercury, followed by an empty impinger to
stop carryover and two acidified potassium permanganate impingers for capturing elemental mercury.
AMS has been shown in bench-scale testing to be highly effective at collecting both divalent and elemental
mercury, with fewer interferences from both sulfur dioxide and chlorine than observed in previous
speciation methods. However, further method development is needed. At the present time, this method
has not been subject to 40 CFR Part 63, Appendix A, Method 301 field validation.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-43
-------
B.15 PARTICLE-SIZE DISTRIBUTION
Particle size distribution may be determined using cascade impactors, or by analysis of particulates
deposited on filters using scanning electron microscopy (SEM). Unfortunately, preliminary data collected
and analyzed using these methods indicates a large amount of uncertainty, especially for particle sizes less
than 1 micron.
Method CARB 501 (CARB 1990b) specifies procedures for collection of particle-size data using cascade
impactors. Method CARB 501 recommends stage configurations representing a range of 0.25 to 10
microns, with five to eight cuts and a total sample size of 50 milligrams (mg). For hazardous waste
combustion (HWC) facilities with very low particulate emissions, a very long sampling period may be
needed to achieve resolution for the smaller sizes. In some cases, the weight gain on the filter may be
negligible or zero. In addition, HWC facilities with wet stacks may experience a problem with particle
agglomeration.
For SEM, filter samples are photographed at magnification, and diameter measurements of the individual
particles are made. The data can then be processed to produce particle-size distributions. Again, particle
agglomeration can be a problem. A disadvantage of SEM is that it only provides physical particle size,
whereas the impactors give aerodynamic size.
EPA (1997d) recognizes the inherent problems associated with measuring very low masses of particulates
and recommends modified Method 5 (Method 51) to improve accuracy, precision, and representativeness
by significantly reducing variability and potential errors. Although Method 51 was not written specifically
for particle-size determinations, it is possible that some of the procedures recommended in Method 51
could also be used in conjunction with cascade impactor or SEM measurements to improve performance.
Since there are no default particle-size distribution assumptions for HWC facilities with no air pollution
control devices (APCD), these facilities will generally need to attempt a particle-size distribution
measurement using currently available procedures. The permit writer should be cognizant of the inherent
problems associated with these procedures, and should work with the facility to obtain the best data
possible under the circumstances. HWC facilities equipped with wet APCD systems or APCD components
other than electrostatic precipitators (ESP) or fabric filters (such as dry scrubber systems or high efficiency
particulate air [HEPA] filters) will either need to attempt a particle-size distribution measurement, or may
be able to obtain particle-size data from the APCD vendor.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities B-44
-------
For facilities equipped with ESPs or fabric filters, the HHRA guidance provides a nine-category particle-
size default assumption. However, since the default assumption may be overly biased towards larger
particles, a particle-size distribution measurement is highly recommended. If the majority of the particle
mass is very small (less than 1 or 2 microns), it will generally not be necessary to achieve high resolution
in cut sizes for the small particles. A sensitivity analysis (The Air Group 1997) concluded that as few as
three size categories could be used in the model with little impact (less than 10 percent) on concentration,
dry and wet deposition. Also, the sensitivity analysis noted that small particles in the range of 1 micron
have very low terminal velocities and effectively are suspended in air. This generally translates to reduced
potential for risk, because indirect risks tend to be driven by deposition.
For facilities equipped with ESPs and fabric filters, it may also be possible to demonstrate via preliminary
air dispersion and deposition modeling that the HHRA guidance default assumptions are conservative (in
that they will result in higher potential for risk), and that site-specific measurements are therefore
unnecessary. The sensitivity analysis (The Air Group 1997) clearly showed that deposition rates will be
substantially higher for distributions which are skewed towards larger particles for most downwind
distances up to 9-15 kilometers. Based on recent measurements, the default assumptions appear to be
biased towards larger particles. If site-specific modeling (using the default assumptions, versus an
assumption of smaller particle sizes in the 1 to 2 micron range) confirmed that deposition would be higher
for key receptor locations using the default assumptions, then particle-size measurements could be avoided.
A final consideration is that particles from cement kilns may have a higher density (2 grams per cubic
centimeter [gm/cm3]) than the density default assumption (1 gm/cm3) in the HHRA guidance. This can
result in up to 50 percent greater dry deposition near the source.
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities B-45
-------
B.16 HYDROGEN CHLORIDE AND CHLORINE
Since potential risks from hydrogen chloride and chlorine are limited to the inhalation pathway, the multi-
pathway SSRA is not expected to impact the current regulatory approach with respect to hydrogen chloride
and chlorine for boiler and industrial furnace (BIF) facilities. However, hazardous waste incinerators
(HWI) will now be expected to characterize their stack emissions for both hydrogen chloride and chlorine,
and the 40 CFR Part 264.343 technology-based limits may be superseded by risk-based limits, if the risk
limits are more stringent. Sampling procedures are specified in Methods 0050/0051, and the collected
samples are analyzed for hydrogen chloride and chlorine using Method 9057. For analytical determination
of additional halides, Method 9056 may be used instead of Method 9057.
Johnson (1996b) provides a convenient summary of research evaluating the use of Methods 0050/0051 for
hydrogen chloride and chlorine determinations. This summary discusses areas of potential concern, and
highlights special considerations, which can help to ensure the quality of the final results.
One area of concern repeatedly noted by cement kiln representatives is that hydrogen chloride
determinations from Methods 0050/0051 could be biased high because volatile particulate chloride salts,
such as ammonium chloride, could penetrate the filter and be converted to hydrogen chloride within the
sampling train (Gossman 1997). Industry has proposed correcting the hydrogen chloride results based
upon analysis of the impinger solutions for cations including Na+, Ca+, K+, and NH4+.
EPA has considered this issue and does not believe that salts will significantly bias the results (Johnson,
1996b, EPA 1996f). Johnson (1996b) emphasizes that correction of the hydrogen chloride results is not
appropriate, because it is not possible to determine how and in what form the ionic material entered the
impingers. The prescribed filter will not pass significant quantities of solid halide salts such as sodium
chloride (NaCl), calcium chloride (CaCl2), or potassium chloride (KC1). Therefore, the presence of Na+,
Ca+- or K+ in the impingers could reflect contamination during handling, a broken filter, or operation with
a wet filter. These problems can be addressed by use of a cyclone and adequate heating in the sampling
train, and careful handling of the train components with special concern for minimization of
contamination.
There is some evidence that it may be possible for ammonium chloride to penetrate the filter as a vapor. If
ammonium chloride is believed to be causing a significant bias, then an infrared spectroscopy-based CEM
monitor for hydrogen chloride could be considered. However, potential inaccuracies relative to Methods
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities B-46
-------
0050/0051 which could bias hydrogen chloride high are not expected to significantly impact the SSRA
process, because excess risks due to hydrogen chloride have not historically been an issue.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-47
-------
B.17 PROCESS SAMPLES
Trial burn protocols for collection of SSRA data will also need to address analysis procedures for
completely characterizing the trial burn wastes, fuels, raw materials, and spike materials. Data equivalent
to the following should be generated:
Quantification of total metals feed rates for aluminum, antimony, arsenic, barium,
beryllium, cadmium, chromium, cobalt, copper, lead, manganese, mercury, nickel,
selenium, silver, thallium, vanadium, and zinc
Proximate analysis, or a comparable evaluation, to determine physical properties including
moisture, percent solids, heating value, ash, and viscosity or physical form, as well as to
determine approximate chemical properties including total organic carbon, total organic
halogens, and elemental composition
Survey analysis or a comparable evaluation for (1) TO content, (2) organic compound
class types, and (3) major organic components, including analysis for VOCs, SVOCs,
PCBs, and PAHs using standard analytical methods
These data define a facility's baseline with respect to long-term impacts and potential effects on human
health and the environment. Any significant changes to the facility's baseline can necessitate additional
risk-based data collection or risk analyses.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-48
-------
REFERENCES
The Air Group. 1997. "Model Parameter Sensitivity Analysis." May 23.
California Air Resources Board (CARB). 1990a. "Method 428 Determination of Polychlorinated
Dibenzo-p-Dioxin (PCDD), Polychlorinated Dibenzofuran (PCDF), and Polychlorinated Biphenyl
Emissions from Stationary Sources." September 12.
CARB. 1990b. "Method 501 Determination of Size Distribution of Particulate Matter Emissions from
Stationary Sources, Stationary Source Test Methods, Volume 1, Methods for Determining
Compliance with District Nonvehicular (Stationary Source) Emission Standards." Monitoring and
Laboratory Division, September 12.
CARB. 1991. "Method 426 Determination of Cyanide Emissions from Stationary Sources, Stationary
Source Test Methods, Volume 3, Methods for Determining Emissions of Toxic Air Contaminants
from Stationary Sources." Monitoring and Laboratory Division, December 13.
CARB. 1997. "Determination of Polycyclic Aromatic Hydrocarbon (PAH) Emissions from Stationary
Sources." July 28 Amendment.
Energy and Environmental Research Corporation (EER). 1997. "Products of Incomplete Combustion
(PICs) from a Hazardous Waste Co-fired Cement Kiln." Report prepared for EPA. January.
Giglio, Jeffrey J., et al. 1997. "The Development of a Method for the Speciation of Source Mercury
Emissions." Presented at the International Conference on Incineration and Thermal Treatment
Technologies. Salt Lake City. May.
Gossman Consulting, Inc. 1997. "Comments on the Trial Burn Guidance Document." September.
Gullett, Brian K., D. Natschke, and K. Bruce. 1997. "Thermal Treatment of 1,2,3,4-Tetrachlorodibenzo-
p-dioxin by Reaction with Ca-Based Sorbents at 23-300 ฐC". Environmental Science and
Technology, Volume 31, No. 7.
Johnson, Larry D. 1995. "Analysis of Modified method Five Train Samples for Multiple Pollutant
Classes." Presented at EPA/A&WMA International Symposium on Measurement of Toxic and
Related Air Pollutants. Research Triangle Park, NC. May.
Johnson, Larry D. 1996a. "Determination of Total Organic Emissions from Hazardous Waste
Combustors." Analytical Chemistry, Vol. 68, No. 1. January 1.
Johnson, Larry D. 1996b. "Stack Sampling Methods for Halogens and Halogen Acids." Proceedings of
the EPA/A&WMA International Symposium: Measurement of Toxic and Related Air Pollutants,
Research Triangle Park, NC. VIP-64, Air & Waste Management Association, Pittsburgh, PA,
1996, pp. 626-633. May.
Lemieux, P.M. and Ryan, J.V. 1997'a. "Interactions Between Bromine and Chlorine in a Pilot-Scale
Hazardous Waste Incinerator." Presented at the International Conference on Incineration and
Thermal Treatment Technologies. Oakland. May.
Lemieux, P.M. and Ryan, J.V. 1997b. "Development of a Hazardous Waste Incinerator Target Analyte
List of Products of Incomplete Combustion." Presented at the International Conference on
Incineration and Thermal Treatment Technologies. Oakland. May.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-49
-------
Lemieux, P.M. and Ryan, J.V. 1998. "Development of a Hazardous Waste Incinerator Target Analyte List
of Products of Incomplete Combustion." EPA-600/R-98-076. Final Report. July.
Midwest Research Institute (MRI) and A.T. Kearney, Inc. 1997. "Products of Incomplete Combustion
Emission Test." Draft Report. Prepared for EPA Office of Solid Waste. April.
Ryan, J.V., Lemieux, P.M., and Groff, P.W. 1997a. "Evaluation of the Behavior of Flame lonization
Detection Total Hydrocarbon Continuous Emission Monitors at Low Concentrations." Presented
at the International Conference on Incineration and Thermal Treatment Technologies. San
Francisco. May.
Ryan, J.V., Lemieux, P.M., Lutes, C. and Tabor, D. 1997b "Development of PIC Target Analyte List for
Hazardous Waste Incineration Processes." Presented at the International Conference on
Incineration and Thermal treatment Technologies. San Francisco. May.
Ryan, J.V. 1998. Personal communication with Catherine Massimino of EPA Region 10, Seattle,
Washington.
U.S. Environmental Protection Agency (EPA). 1989. "Interim Procedures for Estimating Risks
Associated with Exposures to Mixtures of Chlorinated Dibenzo-p-Dioxins and Dibenzofurans
(CDDs and CDFs)." March.
EPA. 1994. "Statement of Work for Organics Analysis, Multi-Media, Multi-Concentration." EPA
Contract Laboratory Program OLM03.0, Revision OLM03.1. August.
EPA. 1996a. "Test Methods for Evaluating Solid Waste, Physical/Chemical Methods." EPA Publication
SW-846 [Third Edition (November 1986), as amended by Updates I (July 1992), II (September
1994), HA (August 1993), HE, (January 1995), and HI (December 1996)].
EPA. 1996b. "Formation of Dioxin-like PICs During Incineration of Hazardous Wastes." Memorandum
To The Record from Dorothy Cantor, Ph.D., Science Advisor, Office of Solid Waste and
Emergency Response. June 21.
EPA. 1996c. "Johnston Atoll Chemical Agent Disposal System (JACADS) Risk Related Issues."
Memorandum from Timothy Fields, Jr., Deputy Assistant Administrator, Office of Solid Waste
and Emergency Response, to Julie Anderson, Director, Waste Management Division, EPA Region
9. October 2.
EPA. 1996d. "Guidance for Total Organics, Final Report." EPA/600/R-96/036. March.
EPA. 1996e. "PCBs: Cancer Dose-Response Assessment and Application to Environmental Mixtures."
EPA/600/P-96/001. National Center for Environmental Assessment, Office of Research and
Development. September.
EPA. 1996f. "Revised Standards for Hazardous Waste Combustors." Proposed Rule. Title 40 of the Code
of Federal Regulations Parts 60, 63,260, 261, 264, 265,266, 270, and 271. Federal Register
61:17358. April 1.
EPA. 1997a. "Hazardous Waste Management System; Testing and Monitoring Activities." Final Rule.
Federal Register 62: 32452. June 13.
EPA. 1997b. "Draft Method 1668 Toxic Polychlorinated Biphenyls by Isotope Dilution High Resolution
Gas Chromotography/High Resolution Mass Spectrometry." Office of Science and Technology,
Office of Water, March.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-50
-------
EPA. 1997c. "TOC Gravimetric Lab Procedures Troubleshooting Options to Minimize and Identify
Sources of Contamination." Memorandum from Catherine Massimino, EPA Region 10. May 1.
EPA. 1997d. "Notice of Data Availability and Request for Comments. Total Mercury and Patticulate
Continuous Emissions Monitoring Systems." Proposed Rule. Title 40 of the Code of Federal
Regulations, Parts 60 and 63. Federal Register 62:67788. December 30.
EPA. In press. "Protocol for Human Health Risk Assessment at Hazardous Waste Combustion Facilities,"
EPA-R6-098-002. Center for Combustion Science & Engineering, Multimedia Planning and
Permitting Division, EPA Region 6.
Guidance on Collection of Emissions Data to Support Site-Specific
Risk Assessments at Hazardous Waste Combustion Facilities
August 1998 Peer Review Draft
B-51
-------
ATTACHMENT 1
METHOD 0040 CLARIFICATIONS
(3 pages)
Guidance on Collection of Emissions Data to Support Site-Specific August 1998 Peer Review Draft
Risk Assessments at Hazardous Waste Combustion Facilities B-52
-------
\
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
NATIONAL EXPOSURE RESEARCH LABORATORY
Research Triangle Park, NC 27711
Office of
Research and Development
MEMORANDUM:
DATE: April 3, 1998
SUBJECT: Method 0040 Questions
FROM: Larry D. Johnson
Source Apportionment and Characterization Branch (MD-47)
TO: Catherine Massimino, U.S. EPA Region 10
This memo is to transmit clarifications to the passages of Method 0040 about which you asked
questions. I'm also sending it to Beth Antley so she can include the explanations in her new Trial
Burn Guidance. Other interested parties are also receiving copies. The appropriate section of
M0040 is reproduced below in Arial type. Your question, as understood by me, follows in italics.
The answer follows in regular GC Times type. If I misinterpreted or omitted any of the questions,
or if the explanations aren't clear, please let me know. Thank you very much for bringing these
confusing instructions to our attention.
7.4.3.4 Draw at least eight times the sample volume of flue gas, or purge for
at least 10 minutes, whichever is greater.
Isn't eight times the sample volume an unreasonably large amount of purge volume, it takes a long
time to carry this out in the field.
You certainly have a good point. The confusion here is that "sample volume" refers to the volume
contained in the part of the train being purged (the probe and lines, typically less than a liter) rather
than the bag volume of 30-40 liters. Purging 40 liters times 8 at 0.5 liters per minute would take
640 minutes. The 10 minute purge mentioned in the method would give an 8 fold volume exchange
for a 0.6 liter tram volume (at 0.5 liters per minute). If the tram volume is estimated to be larger
than 0.6 liters, then the purge tune needs to be increased accordingly.
-------
7.6.5.2 Rinse the condenser, the condensate trap and the sample line three
times with 10 mL of HPLC grade water and add the rinsings to the measuring cylinder containing
the condensate. Record the final volume of the condensate and rinse mixture on the field sampling
data form. High moisture sources (such as those with wet control devices) may require a 150-mL
or 200-mL measuring cylinder while low moisture sources (such as some rotary kilns and pyrolytic
incinerators) may require only a 100-mL size.
Does the method mean 3 rinses of 10 mL each, or three rinses totaling 10 mL?
Three rinses of 10 mL each was the intended instruction. The contractor who wrote and field
tested the method, feels that it is difficult to achieve an adequate rinse with much less than 10 mL.
7.6.5.3 Pour the contents of the measuring cylinder into a 20- or 40-mL amber
glass VOA vial with a Teflonฎ septum screw cap. Fill the vial until the liquid level rises above the
top of the vial and cap tightly. The vial should contain zero void volume (i.e., no air bubbles).
Discard any excess condensate into a separate container for storage and transport for proper
disposal.
If three rinses of 10 mL each are carried out, there -will always be a minimum of 30 mL of
condensate. Why is a 20 mL vial an option? Do you throw out the extra liquid?
The section above does instruct the sampler to discard any excess condensate, so the 20 mL vial
would be an option, but not necessarily the best one. If more than 10 mL of actual condensate is
collected, even the 40 mL vial would not hold the combined volume. If the analytical method is
to use purge-and-trap technology (it usually does), then it is best to discard as little of the liquid as
possible. The detection limit may be minimized by purging the target compound from as much of
the original volume as possible.
This question also brings up another important point which is inadequately addressed in the
method. In Section 7.6.5.2 above, the sampler is instructed to record the final volume of
condensate and rinse mixture on the field sampling data form (shown in Figure 6). This volume
is called "Total condensate volume" on the form in Fig.6, and becomes Vlc as defined (Total
volume of liquid collected in the condensate knockout trap) in Section 7.8.2. Vlc is used in equation
15 in Section 7.8.9.2 to calculate the amount of target compound collected in the condensate. This
whole sequence is all well and good, as long as there is excess condensate, or exactly enough to
fill a vial. However, in the case where the condensate and rinses do not completely fill the vial and
the vial must be "topped off' with rinse water as described in Section 7.6.5.3, then the vial volume
should be used as Vlc rather than using the volume measured with the measuring cylinder (and
recorded as "Total condensate volume" on the form in Fig. 6).
Example:
Given that the condensate plus rinses makes up 30 mL. If a 20 mL vial is used to ship the sample,
then Vlc should be 30 mL, since the sample concentration determined in the lab was contained hi
30 mL of liquid. If a 40 mL vial was used, then the sample was diluted with 10 mL of water, so
-------
Vlc must be 40. The lab determined concentration was contained in 40 mL rather than the original
30. In this example, the use of the wrong volume would introduce a negative bias of 25 % into the
condensate result. The magnitude hi actual cases would depend on the dilution ratio. Since 30 mL
is pretty much the minimum total volume, as discussed above, the example is likely to be a worst
case. Of course, most of the target compound will usually be collected in the bag rather than the
condensate, so even a 25% bias in the condensate value might have little effect on the overall
results. This just demonstrates, once again, that no matter how many highly competent people
write and review a method (and there were quite a few in this case), it takes years of use to
discover all the ambiguities and oversights.
cc
Beth Antley, Region 4
Bob Fuerst, ORD
Barry Lesnik, OSW
Jeff Ryan, ORD
Gene Riley, OAQPS
Joan Bursey, Eastern Research Group
-------
------- |