United States Office of Research and EPA/600/R-98/076
Environmental Protection Development June 1998
Agency Washington DC 20460
vvEPA Development of a
Hazardous Waste
Incinerator Target Analyte
List of Products of
Incomplete Combustion
Prepared for:
Office of Solid Waste
Prepared by:
National Risk Management Research Laboratory
Research Triangle Park, NC 27711
-------
FOREWORD
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's
land, air, and water resources. Under a mandate of national environmental laws, the Agency
strives to formulate and implement actions leading to a compatible balance between human
activities and the ability of natural systems to support and nurture life. To meet this mandate,
EPA's research program is providing data and technical support for solving environmental
problems today and building a science knowledge base necessary to manage our ecological
resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for investigation of
technological and management approaches for reducing risks from threats to human health and the
environment. The focus of the Laboratory's research program is on methods for the prevention and
control of pollution to air, land, water, and subsurface resources; protection of water quality in
public water systems; remediation of contaminated sites and groundwater; and prevention and
control of indoor air pollution. The goal of this research effort is to catalyze development and
implementation of innovative, cost-effective environmental technologies; develop scientific and
engineering information needed by EPA to support regulatory and policy decisions; and provide
technical support and information transfer to ensure effective implementation of environmental
regulations and strategies.
This publication has been produced as part of the Laboratory's strategic long-term research plan. It
is published and made available by EPA's Office of Research and Development to assist the user
community and to link researchers with their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental Protection
Agency, and approved for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use. This document is available to the public
through the National Technical Information Service, Springfield, Virginia 22161.
-------
DEVELOPMENT OF A HAZARDOUS WASTE INCINERATOR TARGET
ANALYTE LIST OF PRODUCTS OF INCOMPLETE COMBUSTION
FINAL REPORT
Prepared by:
Paul M. Lemieux and Jeffrey V. Ryan
U.S. Environmental Protection Agency
National Risk Management Research Laboratory
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
Prepared for:
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
Pilot-scale incineration experiments were performed to develop a comprehensive list of products of
incomplete combustion (PICs) from hazardous waste combustion (HWC) systems. The goals of
this project were: 1) to develop an expanded list of HWC target analytes for EPA's Office of Solid
Waste (OSW) to use as a basis for a PIC-based regulatory approach; 2) to identify the total mass of
organic compounds sufficiently to estimate the toxicity of the complex mixture; and 3) to enable
OSW to assess the relative importance of poly chlorinated dibenzo-p-dioxins and polychlorinated
dibenzofurans (PCDDs/PCDFs) to other PICs.
These tests were performed under varied combustion conditions feeding a mixed surrogate waste,
resulting in the generation of numerous PICs. While many of these PICs were identified as target
analytes using standardized sampling and analytical methods, the majority of PICs present in the
incineration emissions were not target analytes. Although a substantial number of PICs have been
tentatively identified, a considerably larger number have not been identified at this time. It can be
concluded from these experiments that the current sampling and analytical schemes for
characterizing HWC emissions provide an incomplete picture of the emission profile.
Innovative analytical techniques, such as multi-dimensional gas chromatography (MDGC) appear
to show great promise for identifying the unknown compounds present in the stack gases. In
many cases, "clean" chromatographic peaks were not able to be identified via mass spectral search
algorithms because what appeared to be a single peak was really many compounds co-eluting off
the column. When these types of peaks were analyzed using the MDGC system, the co-eluting
compounds were resolved and identified.
As a result of these experiments, an expanded list of PIC target analytes has been developed. This
list is by no means complete or comprehensive. This list should be viewed in context with this
particular set of experiments; i.e., waste mix. The PICs generated from the incineration of other
mixed waste streams have not been evaluated.
The PICs identified fall into several chemical classes. A wide variety of chloro, bromo, and mixed
bromochloro alkanes, alkenes, alkynes, aromatics, and polyaromatics were detected. In addition,
nonhalogenated hydrocarbon homologues along with oxygenated, nitrogenated, and sulfonated
organics were detected. Analytical methods specifically suited to identify these chemical classes
are needed to enhance PIC characterizations. Of the non-target semivolatile organic compounds
that were detected but not identified, the vast majority were large alkanes (with more than 10
carbons), esters of high molecular weight carboxylic acids, and phthalates. The authors believe
that improved analytical methodologies emphasizing validation and quantification of these
compounds would provide the greatest opportunity to reduce uncertainty in risk assessment
calculations.
Other secondary goals of this project were also realized. It was observed that increases in feed
bromine concentration could dramatically impact emissions of many chlorinated organics,
including PCDDs/PCDFs. It was also observed that concentrations of chlorinated alkenes dropped
as residence time in the secondary combustion chamber increased, while ring growth reactions
were observed in-flight in moderate temperature regions prior to gas quenching. Finally, evidence
has been found to support the use of certain easily measured volatile organic PICs as surrogates for
PCDD/PCDF emissions.
-------
Some goals of this project were not attained. A mass balance between identified PICs and total
hydrocarbon (THC) measurements was not established. THC concentrations were in the very low
ppm range, within the analytical accuracy of the instruments. Attempts to measure non-chlorinated
alkanes, alkenes, and alkynes via bag sampling did not detect measurable levels of those
compounds.
ACKNOWLEDGMENTS
This work was performed by Acurex Environmental Corp., under EPA Contract No. 68-D4-0005,
under the direction of EPA Work Assignment Manager Paul Lemieux. The authors would like to
thank Kevin Bruce, J. Frank Day, Tony Lombardo, Mark Calvi, Ray Thomas, and Eric Squier,
Acurex Corp. for their dedication in performing these tests. The authors would also like to thank
Senior Environmental Employees (SEE) John Dawkins, Robert King, and Bob Frazier for
operating the continuous emission monitoring systems. The authors would also like to thank
Wayne Rubey of the University of Dayton Research Institute for his invaluable help in performing
the multi-dimensional gas chromatography analyses.
-------
TABLE OF CONTENTS
AB S TRACT ii
ACKNOWLEDGMENT S iii
LIST OF FIGURES vi
LIST OF TABLES vii
1.0 INTRODUCTION 1-1
1.1 - Focus 1-1
1.2 - Regulatory Basis 1-1
1.3 - Surrogate Indicators 1-2
1.4 - Emission Characterization 1-2
1.5 - Limitations 1-3
2.0 EXPERIMENTAL APPROACH 2-1
2.1 - Focus 2-1
2.2 - Experimental Equipment 2-1
2.2.1 - Rotary Kiln Incinerator Simulator 2-1
2.2.2 - Flue Gas Cleaning System 2-2
2.3 - Waste Feed 2-3
2.4 - Sampling Approach 2-4
2.4.1 - General Sampling Information 2-4
2.4.2 - Continuous Emissions Monitors 2-7
2.4.3 - On-Line GC 2-7
2.4.4 - Volatile Organics 2-8
2.4.5 - Semivolatile and Non-Volatile Organics 2-9
2.4.6 - PCDDs/PCDFs 2-10
2.5 - Analytical Approach 2-10
2.5.1 - General Analytical Information 2-10
2.5.2 - Volatile Organics 2-10
2.5.3 - Semivolatile and Non-Volatile Organics 2-11
2.5.4 - PCDDs/PCDFs 2-12
3.0 RESULTS AND DISCUSSION 3-1
3.1 - Results from Continuous Measurements 3-1
-------
3.2 - Volatile Organic Results 3-3
3.2.1 - On-Line GC Results 3-3
3.2.2 - VOST and Tedlar Bag Results 3-5
3.3 - Semivolatile and Non-Volatile Organics 3-16
3.3.1 - Conventional GC/MS Analytical Results 3-16
3.3.2 - Multi-Dimensional GC/MS Analytical Results 3-22
3.4 - PCDDs/PCDFs and PBDDs/PBDFs 3-23
3.5 - Surrogate Performance Indicators 3-25
4.0 CONCLUSIONS 4-1
4.1 - Target Analyte List 4-1
4.2 - Effect of Presence of Bromine 4-2
4.3 - Surrogate Performance Indicators 4-2
4.4 - Implications of These Results 4-3
4.5 - Recommendations 4-3
5.0 REFERENCES 5-1
APPENDIX A: QUALITY CONTROL EVALUATION REPORT A-l
A.I - Continuous Measurement Results A-l
A.2 - Volatile Organic Compound Analyses A-2
A.2.1 - VOST Samples A-2
A.2.2 - Tedlar Bag Samples A-3
A.3 - Semivolatile Organic Compound Analyses A-3
A.4 - PCDD/PCDF and PBDD/PBDF Analyses A-4
A.5 - Online GC Samples A-5
-------
LIST OF FIGURES
Figure 2-1. Rotary kiln incinerator simulator 2-2
Figure 2-2. Flue gas cleaning system 2-3
Figure 2-3. Metal solution injection system 2-6
Figure 2-4. On-line GC system 2-8
Figure 2-5. MDGC-MS Setup (Courtesy of UDRI) 2-12
Figure 3-1. OLGC results of tetrachloroethylene concentrations at choke and SCC exit 3-3
Figure 3-2. OLGC results of 1,2 dichlorobenzene concentrations at choke and SCC exit 3-5
Figure 3-3. Average concentrations of analogous Ci and C2 halogenated compounds 3-15
Figure 3-4. MDGC/MS Analysis of Methylene Chloride Extract from Run 10. The upper trace is
for the single-column, "one-dimensional" analysis. "Two dimensional" resolution of a singlet and
a doublet are shown. (Courtesy of UDRI) 3-22
Figure 3-5. Trichloroethylene vs. Total PCDDs; R2=0.6476 3-25
Figure 3-6. Trichloroethylene vs. Total PCDFs; R2=0.6956 3-26
Figure 3-7. Trichloroethylene vs. Total PCDDs+PCDFs; R2=0.6915 3-26
VI
-------
LIST OF TABLES
Table 2-1. Waste Feed Composition 2-5
Table 2-2. Test Conditions 2-6
Table 2-3. Samples Taken During Each Test for Which Analytical Results Are Available 2-7
Table 3-1. Temperature Results (°C) 3-1
Table 3-2. CEM Emissions Results 3-2
Table 3-3. On-line Gas Chromatograph Results for Volatile Organic PICs (|lg/m3) 3-4
Table 3-4. Tedlar Bag Results: Target Compounds (|ig/m3) 3-6
Table 3-5. Tedlar Bag Results: Tentatively Identified Compounds (|ig/m3) 3-7
Table 3-6. VOST Results (|ig/m3) 3-8
Table 3-7. Target Volatile Organic Compounds Detected 3-10
Table 3-8. Tentatively Identified VOST Compounds 3-11
Table 3-9. Combinations of Detected Ci and C2 Compounds 3-12
Table 3-10. Ci & C2 Halogenated Hydrocarbons (jig/m3) 3-14
Table 3-11. Halogenated Aromatic VOC Results (Jig/m3) 3-16
Table 3-12. Semivolatile Organic Target Results (|ig/m3) 3-17
Table 3-13. Semivolatile Organic Tentatively Identified Compounds (|lg/m3) 3-20
Table 3-14. Compounds identified via MDGC/MS 3-23
Table 3-15. Polychlorinated and Polybrominated Dioxins and Furans 3-24
Table A-l. Data Quality Indicator Summary for Critical Measurements A-2
Vll
-------
1.0 INTRODUCTION
1.1 - Focus
Assessing the risk posed by combustor emissions requires sampling and analysis of what is
leaving the stack. The chemical analysis must be compound specific in order to consider the
toxicity of each compound. Efficient and cost effective sampling and analysis for routine
regulatory control requires a target analyte list to focus the effort. A list of Products of Incomplete
Combustion (PICs) suitable for focusing this effort is not well developed. The primary goal of
this project is to develop such a list. This list will help serve as a basis for EPA's Office of Solid
Waste (OSW) to pursue a PIC-based regulatory approach.
In the past, the Appendix VIII1 list of hazardous compounds has become the de facto list for
hazardous waste combustor (HWC) investigations. The Appendix VIII list was generated by
appending lists of chemicals that were previously regulated by other government agencies (U.S.
Department of Transportation (DOT) shipping labels, etc.). And, as such, it is not a list of
compounds well focused to HWC stack emissions. Moreover, this list focuses on compounds
possessing hazardous characteristics that are most often the Primary Organic Hazardous
Constituents (POHCs). As a result, existing required analytical methodologies focus on measuring
the POHC. Very few PICs that are formed are targeted by current analytical methodologies.
Analytical methodologies capable of identifying and quantifying PICs are required. This effort
avoids the focus provided by Appendix VTII by approaching the task with an open mind in order to
establish a list of compounds of importance to HWC emissions.
As a starting point, this study used existing trial burn data, laboratory-scale research literature, and,
where relevant, target analyte lists based on Appendix VIII and the hazardous air pollutant (HAP)
list from the 1990 Clean Air Act Amendments^ . It must be stressed, though, that this was only a
starting point. The vast majority of the effort for this study was consumed in identification and
quantification of unknown compounds.
1.2 - Regulatory Basis
HWCs have been regulated by the Resource Conservation and Recovery Act (RCRA),3 based on
the destruction and removal efficiency (DRE) of POHCs as defined in a trial burn. This approach
used the initial decomposition of the POHC, the first step in converting the organic POHC
molecule to carbon dioxide (CO2) and water (H2O), as a surrogate for the extent of complete
conversion to CO2 and H2O. The goal of reducing the toxicity of the hazardous constituents
requires many reactions (chlorobenzene has 12 bonds to break and 18 new bonds to make) to
completely react to CO2 and H2O. If the reaction sequence goes to completion, the toxicity is
reduced completely (i.e., CO2 and H2O are not toxic). However, partial destruction can mute the
reduction in toxicity, and reformation reactions can occur that cause molecular size growth; these
can also mute the reduction in toxicity or, in some cases, increase the toxicity from that of the
original organic molecule being incinerated^. Additionally, chlorine from the hazardous waste,
released in the form of hydrochloric acid (HC1) or diatomic chlorine (Cl2), can react with naturally
occurring hydrocarbons in the cool end of some incineration facilities (e.g., cement kilns) and
generate potentially toxic hazardous organic compounds^. A new PIC-based approach can
potentially avoid these problems associated with the POHC DRE approach.
-------
Current regulatory approaches use carbon monoxide (CO) as a surrogate for PICs. This approach
is based on the assumption that the oxidation of CO to CO2 is the final step in the long chain of
complex combustion reactions. Minimization of CO thus is assumed to minimize PICs.
Unfortunately, this assumption does not hold up well when polychlorinated dibenzo-p-dioxins and
polychlorinated dibenzofurans (PCDDs/PCDFs), which are generally formed in the cooler regions
of the incinerator, are taken into account. In the case where PCDDs and PCDFs constitute a
significant component of the organics-based toxicity of the mixture, the "CO-as-a-PIC-surrogate"
approach breaks down. CO appears to be a viable surrogate to distinguish between "poor"
combustion and "good" combustion, but as emissions limits get lower and lower, CO is not a
reliable surrogate to distinguish between "good" combustion and "great" combustion. At that
point, other parameters have a much more significant influence on the emissions of
PCDDs/PCDFs, such as the temperature at which the particulate control device operates.^ In other
words, minimization of CO is a necessary, but not sufficient condition for PIC minimization.
1.3 - Surrogate Indicators
A surrogate incinerator performance indicator is an easily measured parameter, compound, or
group of compounds whose variance can account for the variance in the measurements of a more
difficult-to-measure compound, such as PCDDs/PCDFs. Although this work will not be used
directly to develop surrogate indicators of performance, it will lay ground work for that purpose.
The task of choosing a surrogate indicator of performance implies that a significant PIC of concern
(one that can significantly influence the results of a risk assessment) is known. PCDDs/PCDFs
have gained notoriety as being potentially significant PICs in many cases, although some critics
have suggested that PCDDs/PCDFs are the most important PICs simply because they are the class
of PICs most frequently investigated. The problem that exists is that PCDDs/PCDFs are present at
the low parts-per-trillion (ppt) levels in the stacks of a well-operated combustion facility. Sampling
and analytical procedures to measure PCDDs/PCDFs are expensive and time consuming. If an
easily measured surrogate were available that gave a strong correlation with PCDDs/PCDFs,
routine compliance tests could potentially be replaced by continuous or semi-continuous
monitoring of that surrogate. In addition, the process could be optimized based on continuous
measurements of that surrogate.
1.4 - Emission Characterization
An additional issue this work may help to address is that of "what fraction of the emissions are
toxic and what fraction are low or non-toxic?" By attempting to quantify as large a percentage of
the mass of organic emissions as possible (in a research level effort) it may be possible to get a
better handle on the question. The public has been quick to assume that the unidentified
compounds are hazardous; since they have not been identified it is not possible to assure the public
that they are of low toxicological significance. This research effort and the Omnibus regulatory
effort intend to identify and quantify both the toxic and low/non-toxic compounds to the extent
possible. It is expected that the bulk of the emissions will be low molecular weight low/non-toxic
compounds.
Although PCDDs/PCDFs, due to their high toxicity^, are likely to be the most toxic organic hazard
in the HWC stack, they are typically present in minute quantities. In addition, there may be entire
classes of PICs that are not even being measured, some of which could potentially influence the
risk assessment calculations. The conservative nature of risk assessment assumes that unknown
1-2
-------
compounds are toxic. Because of this, risk assessment uncertainties can be influenced not only by
not detecting PICs that are important from a toxicological point of view, but also by not detecting
harmless compounds that potentially comprise much of the mass of stack emissions. Sampling
and analytical methodologies may not be sufficiently developed to generate reliable emissions data.
Compounds that fall into this category are the brominated and bromochloro analogs to
PCDDs/PCDFs (the polybrominated dibenzo-p-dioxins and polybrominated dibenzofurans
[PBDDs/PBDFs] and mixed bromochloro dibenzo-p-dioxins and mixed bromochloro
dibenzofurans [PXDDs/PXDFs]), and polycyclic aromatic hydrocarbons (PAHs) substituted with
various species (oxygen, chlorine, sulfur)°. Another issue is the measurement of compounds such
as phthalates, which are frequently detected in HWC emissions, but may be artifacts of sampling
and analytical treatments.
1.5 - Limitations
The experiments were performed on EPA's rotary kiln incinerator simulator (RKIS) located in
Research Triangle Park, NC. Exact quantification of concentrations was not a primary goal for
this study. A more important goal was to derive a detailed list of target compounds that can be
found at levels above the detection limits. The existing database of PIC data from bench,
laboratory, pilot, and full-scale was used as a starting point for development of this list.
It is critical to understand that all quantified PICs generated in this study are based on the pilot-
scale RKIS, burning the chosen waste mix, at the given conditions, prior to any flue gas cleaning
equipment. The RKIS is a small pilot-scale kiln, and many of the fluid mechanical features of full-
scale kilns that can produce excess emissions are not present in the RKIS. As such, the system
sometimes needs to be operated slightly outside what would constitute normal incinerator operating
conditions in order to properly quantify important emission trends and measure subtle phenomena.
It is believed that this system generates qualitatively applicable data, although emissions results
from the RKIS should not be quantitatively compared to full-scale systems.
1-3
-------
2.0 EXPERIMENTAL APPROACH
2.1 - Focus
The emphasis of this effort was placed on analytical operations rather than sampling operations.
The sampling methods selected were appropriate for the quantitative capture of volatile,
semivolatile, and non-volatile organics. The issue was how to retrieve and analyze the organic
compounds captured by these methods. Both standard and non-routine approaches were used.
Methods development/validation was not within the scope of this project. It must be reiterated that
the emphasis of this project was to identify PICs that are not routinely identified by conventional
methodologies. Once these PICs have been identified and their relative toxicological importance
evaluated, emphasis can more appropriately be placed on method development and validation.
Certain samples, such as those collected using SW-846 Draft Method 0040^ (Tedlar bags) or
Method 003Q10 (VOST), must be analyzed soon after the samples have been taken. These
analyses were performed within 24 hours. Other samples, though, such as Method 001011
(MM5) or Method 231^ can be stored for a longer time after extraction of the sampling media. In
addition, since this effort was directed at identification of the multitude of unknowns in the
semivolatile and non-volatile fraction, the majority of the effort was directed at the higher molecular
weight compounds.
2.2 - Experimental Equipment
2.2.1 - Rotary Kiln Incinerator Simulator
The incineration tests were performed using the RKIS facility at the EPA's Air Pollution
Prevention and Control Division's (APPCD's) combustion laboratory in Wing-G of the EPA's
Environmental Research Center (ERC) located in Research Triangle Park, NC. The facility has a
RCRA Research, Development, and Demonstration (RD&D) permit to burn actual and surrogate
hazardous waste. The RKIS, shown in Figure 1, consists of a 73 kW (250,000 Btu/hr) rotary kiln
section, a transition section, and a 73 kW (250,000 Btu/hr) secondary combustion chamber
(SCC). The RKIS was designed for the testing of liquid and solid surrogate hazardous waste
materials.
The RKIS was designed to contain the salient features of full-scale kilns, but still be sufficiently
versatile to allow experimentation by varying one parameter at a time or controlling a set of
parameters independently. The rotating kiln section contained a recess which contains the solid
waste during incineration. The recess was designed with a length to diameter (L/D) ratio of 0.8,
which is 20 to 25% of a full-scale system. The main burner, based on an International Flame
Research Foundation (IFRF) variable swirl design, was the primary heat source for the system.
Natural gas was used as the primary fuel during startup and idle, then was switched over to the
surrogate waste feed used throughout testing.
From the kiln section, the combustion gases entered the transition section. The gases then flowed
into the SCC. The SCC consisted of three regions: the mixing chamber, the plug flow section, and
the stack transition section. A replaceable choke section separated the mixing chamber from the
plug flow section.
2-1
-------
Water Jacket Dilution Air Duct 2 TC Duct 3 TC
Duct 1 TC 11 1
Duct 4 TC
Liquid Waste
Rotary Leaf
Spring Seal
Afterburner
Liquid
Waste
Main
Burner
Ramrod
Kiln Section Transition Section
Figure 2-1. Rotary kiln incinerator simulator
A conical refractory insert was installed into the first plug flow sub-section to provide a gradual
divergence from the choke diameter to the plug flow section diameter and minimize recirculation
zones downstream of the choke. The afterburner, also based on an IFRF variable swirl design,
provided heat and flame to the SCC, and was also fired with natural gas during startup and idle
times, then switched to the liquid surrogate waste during the tests.
Combustion gases exiting the afterburner passed through a water-jacketed convective cooling
section of 20.3 cm (8-in nominal pipe thread [NPT]) diameter stainless steel (SS) ducting. Further
cooling was achieved by adding ambient dilution air via a dilution damper located upstream of the
9.9-m (35-ft) sampling duct. Emissions samples were collected at sampling locations 66.7-cm
(169.5-in) and 98.6-cm (250.5-in) downstream of the dilution damper. These sampling locations
were oriented to meet isokinetic sampling requirements.
2.2.2 - Flue Gas Cleaning System
All of the research combustors in the Wing-G combustion research facility were manifolded into a
common flue gas cleaning system (FGCS). The FGCS consisted of a 1.02 MW (3.5 x 10^
Btu/hr) afterburner followed by a water quench, baghouse, and wet scrubber. The purpose of the
2-2
-------
FGCS was to take exhaust gases from the research combustors, destroy any unburned organic
material, and remove any particulates and acid gases from the effluents prior to their release to the
atmosphere.
A roof-mounted induced-draft (ID) fan pulled exhaust gases from research combustors into a
manifold. Flow direction of emissions was then determined by the position of a three-way valve.
By-pass (vent fumes mode) flow feeds directly to the draft fan. The flow of fumes (permit mode)
feeds through the afterburner, quench, baghouse, scrubber and draft fan.
Exhaust gases were oxidized at temperatures of 1000 °C (1,832 °F) or greater for at least 2 s in a
natural-gas-fired Hirt afterburner. The exhaust gases of the afterburner were then cooled by a
controlled water spray that is air-aspirated through a nozzle in the quench section. Particulate
matter was then removed by filter cartridges in a baghouse. Acid gases were removed in the
scrubber by a sodium hydroxide caustic solution that is sprayed into the exhaust stream. After
exiting the draft fan, exhaust emissions are continuously monitored for CO2, CO, and oxygen
(O2). The FGCS is depicted in Figure 2-2.
Effluent
from
RKIS
Figure 2-2. Flue gas cleaning system.
2.3 - Waste Feed
The surrogate hazardous waste that was fed during tests was designed to possess representative
compounds from many common classes of organic hazardous wastes. The composition of the
surrogate hazardous waste feed was developed based on recommendations from members of OSW
and the Regional Permit Writers. Table 2-1 lists the composition of the surrogate waste feed. In
addition to the organic surrogate waste, an aqueous mixture of metal salts, including zinc
nitrate»hexahydrate, nickel nitrate»hexahydrate, and copper nitrate»hexahydrate, was also fed into
2-3
-------
the kiln. The purpose of the metals injection was to provide a representative supply of metal
catalyst to promote any heterogeneous reactions forming PCDDs/PCDFs. Copper (Cu), nickel
(Ni), and zinc (Zn) were fed as metal nitrate»hexahydrate compounds dissolved in 100 mL/hr of
water with sufficient metal present to reach the target gas-phase concentrations of 60 jig/m3 (Cu),
40 |ig/m3 (Ni), and 90 |ig/m3 (Zn).
Hazardous wastes are burned in blended mixtures of many waste streams. These tests were
designed to mimic this complexity. The principal purpose of this work was to establish a list of
possible compounds that should be investigated as PICs from hazardous waste incineration. In
order to have as many compounds on the list as possible, the feed stream was designed to have
several organic compounds of several different classes in its makeup. Additionally, since as much
of the effort as possible was to be directed at analysis, the cost of the waste feed was designed to
be held to as low a level as possible. In addition, it was required that personnel safety be
maximized.
With the exception of runs where batch feeding occurred, all runs were performed using the same
standard mix of compounds. The nominal chlorine (Cl) content of the waste was 10 % by weight.
The waste consisted of a mixture of several compounds co-fired with No. 2 fuel oil. Some
brominated organic compounds were substituted for a fraction of the chlorinated compounds. The
composition of the waste that was fed is shown in Table 2-1. Note that too much dibromoethane
was inadvertently added in Run 10, resulting in a bromine (Br) mass percent 3 times the intended
level.
In addition, some of the tests involved batch charging of containerized liquid wastes. The charges
consisted of 0.9 L (1 qt) polyvinyl chloride (PVC) containers filled with No. 6 fuel oil that had
been doped with hexachlorobenzene (1000 ppm). This waste was fed in 10 minute intervals with
the kiln rotating at 0.5 rpm.
During all runs, the kiln and afterburner burned the standard mix of wastes in both the primary and
secondary burners, by pumping the makeup fuel (No. 2 fuel oil) from 55 gal. drums, and mixing it
with the stream of waste compounds that are being pressure-fed from a 5 gal. container using
pressurized nitrogen. The entire system was tied into the flame safety interlock system so that any
flameout resulted in the waste feed's being cut off. Flow rates were measured using rotameters.
The nominal experimental descriptions that were used are listed in Table 2-2. The combustion
blanks consisted of samples taken while no waste was being fed.
The metals solution was injected into the primary combustion chamber using the apparatus shown
in Figure 2-3.
2.4 - Sampling Approach
2.4.1 - General Sampling Information
The sampling methodologies and procedures used to conduct this study followed EPA-
standardized test methods for the collection of volatile, semivolatile, and non-volatile organics. In
general, the test procedures were followed as described in the reference method. Analytical results
are not available for all runs for which sampling occurred. Table 2-3 lists the samples taken during
the tests for which analytical results are available. With the exception of the continuous emission
2-4
-------
monitors (CEMs), all extractive samples were taken at the sample ports in the horizontal duct
between the RKIS and the FGCS. As shown in Figure 2-1, one set of CEMs sampled at the port
located near the kiln exit; another set of CEMs sampled at the port located near the SCC exit; and
the HC1 CEM sampled just downstream of the sample port where all of the extractive organics
sampling trains were located.
Table 2-1. Waste Feed Composition
Class
carrier liquid
chlorinated non-aromatic
chlorinated aromatic
non-chlorinated aromatic
alcohol
ketone
nitrated waste
FAFF
brominated waste
Compound
No. 2 fuel oil
methylene chloride
chloroform
carbon tetrachloride
monochlorobenzene
dichlorobenzene
chlorophenol
toluene
xylene
isopropanol
methyl ethyl ketone
pyridine
naphthalene
bromoform
ethylene dibromide
Formula
n/a
CH2C12
CHC13
CC14
ceHsci
C6H4Cl2
ceHscio
C7H8
CgHlO
CsHgO
C4H80
CsHsN
CloHg
CHBrs
C2H4Br2
Mass %
50.0
8.0
4.5
2.4
3.3
3.8
1.5
5.2
5.2
2.4
4.8
5.9
1.5
0.75
0.75b
a - Polycyclic aromatic hydrocarbon.
b - On Run 10, too much ethylene dibromide was inadvertently added.
2-5
-------
Combustion
Air
Controller
/ /
T5
Pump
Natural Gas
Figure 2-3. Metal solution injection system.
Table 2-2. Test Conditions.
Run
Description
Date
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Combustion Blank
Combustion Blank
High Temperature
High Temperature
Baseline
Baseline
SCC Off
SCC Off
Low Temperature
Low Temperature
Fuel-Rich
Fuel-Rich
Fuel-Rich
Fuel-Rich
Batch Charging
Batch Charging
4/13/95
4/18/95
4/20/95
4/26/95
5/3/95
5/4/95
5/9/95
5/10/95
5/12/95
5/16/95
5/23/95
5/31/95
8/14/95
8/16/95
8/21/95
8/23/95
2-6
-------
Table 2-3. Samples taken during each test for which analytical results are available.
Run CEMs Method 0040 Method 0023 Method 0030 Method 0010 OLGC
TedlarBags Dioxins VOST MM5
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
16
X
X
X
X
X X
X X
X
X
X X
X
X
X
X
X X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
2.4.2 - Continuous Emissions Monitors
Two separate CEM benches provided simultaneous gas monitoring of O2, CO2, CO, nitric oxide
(NO), and THC before and after the SCC. In addition to the two CEM benches, a Perkin
Elmer/Bodenseewerk MCS 100 Emission Monitoring System (which is capable of measuring HC1,
CO2, and H2O simultaneously and continuously under wet conditions) was available throughout
most of the tests.
2.4.3 - On-Line GC
Volatile organic PIC emissions were measured on selected runs using an on-line gas
chromatograph (OLGC) system, shown in Figure 2-4. The OLGC analytical system 13,14
contained a heated sample delivery system, a purge and trap sample concentrating system, and the
GC analytical system. The sample concentrating device was a Tekmar LSC-2000 thermal
desorption unit that had been modified to accommodate the direct collection of combustion
samples. The GC analytical system was a HP 5890 series II GC equipped with both flame
ionization detector (FID) and an electron capture detector (ECD). The effluent of the column is
split (ratio 9:1, respectively) to deliver sample to both the FID and ECD simultaneously. Ninteen
individual volatile organic PICs can be quantified at concentration levels of about 1 ppbv. The
OLGC sampled at two different locations: 1) at the choke in the SCC, and 2) near the exit of the
SCC where the other CEMs sampled, in an attempt to measure changes in PICs as a function of
residence time.
2-7
-------
Vacuum
Pump
2.4.4 - Volatile Organics
Figure 2-4. On-line GC system.
Volatile organic compounds (VOCs) were collected using both the Volatile Organic Sampling Train
(VOST - SW-846 Method 0030)10 and Tedlar bags (SW-846 Draft Method 0040)9 The VOST
method is intended to be used for VOCs with boiling points (BPs) ranging from 30 to 1 10 °C. For
the more volatile VOCs (BPs < 30 °C), Tedlar bag samples were collected.
VOST samples were collected as described in SW-846 "Test Methods for Evaluating Solid Waste"
Method 0030 "Volatile Organic Sampling Train." Four sets of samples were collected for each test
condition (two sets per test day). A total volume of ~ 20 L was collected for each sample.
Sampling was performed at 0.5 L/min for 40 min. Liquid condensate samples were also collected
daily for separate analysis.
The VOST tube sets were quality control (QC) checked for background contaminants by GC/MS
under the same conditions used for actual sample analysis. The acceptable blank level was less
than 10 ng for any single target analyte per tube. There is no established level for total VOC
contamination. VOST tubes were conditioned in batches of seven sets. At least one set of tubes
out of each batch of seven (14.3%) was QC checked.
Once the tubes were QC checked, the Tenax-only tubes were spiked with known quantities o
labeled benzene and bromofluorobenzene (BFB) as part of the quality assurance (QA) procedure
for the sampling. The tubes were then individually placed in metal cigar tube-type containers which
were secondarily placed in a metal container or glass jar containing activated charcoal. The
secondary container was then kept in a refrigerator maintained near 0 °C until delivery for
sampling. Following sampling, the tubes were returned to their respective individual containers
2-8
-------
and then placed in a separate secondary container, also containing activated charcoal, and kept
refrigerated until analyzed. All samples were analyzed within 30 days of collection.
Tedlar bag samples were collected as described in SW-846, "Test Methods for Evaluating Solid
Waste," and Draft Method 0040, "Sampling of Principal Organic Hazardous Constituents from
Combustion Sources Using Tedlar Bags." Only one sample was collected for each test condition.
A total volume of ~ 20 L was collected for each sample. The liquid condensate was also collected
for separate analysis.
The Tedlar bags were conditioned for use by sequentially filling the bags with nitrogen and then
evacuating them with a vacuum pump. This conditioning process was performed at least three
times or until the bags were demonstrated to be free of background contaminants. The bags
themselves were QC checked for background contamination as described above. The nitrogen
used for conditioning was also tested for background contamination. All bags used for sampling
were QC checked. Once the bags were demonstrated to be free from background contamination,
they were once again evacuated and stored at ambient temperature until used for sampling.
Following sampling, the bags were resealed. All samples were analyzed within 72 h of collection.
2.4.5 - Semivolatile and Non-Volatile Organics
Semivolatile organic compounds (SVOCs) were collected using the Modified Method 5 (MM5)11
train train as described in SW-846 "Test Methods for Evaluating Solid Waste" Method 0010
"Modified Method 5 Sampling Train." Two MM5 samples were collected for each test condition.
Samples were collected on separate test days. The trains were operated isokinetically as required by
the method. As stipulated in EPA 40 CFR Part 60 Method 1 A, the Pitot tube was not attached to
the probe. Radial sampling locations were based on the preliminary velocity traverse. A post-test
velocity traverse was also performed. The pre- and post-test velocity traverses were used to assess
isokinetic variation. The run times were increased to maximize the total volume sampled. A
nominal run time of 4 hours was used. As no particulate measurements were made from this train,
filters were not weighed. No other method deviations are anticipated.
The MM5 trains were recovered so as to generate five separate components for analysis:
1. The parti culate filter (labeled Container 1)
2. The front-half rinse (labeled Container 2)
3. The back-half rinse - all train components between filter and sorbent module (labeled
Container 5)
4. The XAD-2 module (labeled Container 3)
5. The condensate and condensate rinse of 1st empty impinger (labeled Container 4)
Note: Container labeling is consistent with Method 0010.
Given the high acid concentration of the sample stream, flushing the XAD-2 sorbent modules with
high performance liquid chromatography (FIPLC) grade water to remove the concentrated acid was
required. This rinse was combined with the contents of Container 5.
The XAD-2 was cleaned and QC checked as described in Method 0010 with several additional
solvents. The methylene chloride extraction was followed by acetone, toluene, and once again
2-9
-------
methylene chloride extractions, respectively. The cleaned XAD-2 was subjected to background
contamination quality control checks. Although the method requires that the XAD-2 blank exhibit
a TCO level less that 10 |ig/g, experience has shown that we can also outperform the recommended
level of 4 |ig/g, typically demonstrating background levels in the 1 |lg/g range. The XAD-2 was
also QC checked by GC/MS to screen for any target analyte background contaminants. No QC
acceptance criteria have been established for this additional QC check, although less than 5
jig/sample (based on -30 g sample) has been achieved for individual target analytes. Prior to
sampling, 40 g of XAD-2 was packed into the sorbent modules, capped with glass stoppers, the
ends wrapped in cleaned aluminum foil, and stored, refrigerated at 4° C until use. Following
sample retrieval, the XAD-2 modules were stored in an identical manner. All samples were
extracted within 30 days of sample collection.
2.4.6 - PCDDs/PCDFs
PCDDs/PCDFs were collected as described in 40 CFR, Part 60, Appendix A, Method 23
"Determination of Poly chlorinated Dibenzo-p-dioxins and Poly chlorinated Dibenzofurans from
Stationary Sources" 12. This method is virtually identical to California Air Resources Board
(CARB) Method 428 "Determination of Polychlorinated Dibenzo-p-dioxin (PCDD),
Polychlorinated Dibenzofuran (PCDF), and Polychlorinated Biphenyl Emissions from Stationary
Sources" 15. The only real differences are in the analytical approach. The MM5 sampling train
location and operation criteria presented above also apply to Method 23. The run times were
increased to maximize the total volume sampled. All samples were extracted within 45 days of
collection.
2.5 - Analytical Approach
2.5.1 - General Analytical Information
The analytical approach considered both screening and analyte-specific analytical techniques. A
literature review of bench-, laboratory-, and pilot-scale incineration studies was used to help
establish an expanded target analyte list. Similarly, target compound classes such as PAHs, that
are made up of many more than the 16 or so compounds routinely targeted, were expanded to
include alkylated, chlorinated, and nitrogenated PAHs that have harmful health effects.
Sulfonated, oxygenated, and nitrogenated heterocyclic compounds were also targeted.
2.5.2 - Volatile Organics
The VOST and Tedlar bag samples collected were analyzed by gas chromatography/mass
spectrometry (GC/MS) following the procedures described in SW-846 Methods 5040/824016,17
This method was suitable for the analysis of both sample types. Method 8240 quantifies
compounds with BPs ranging from —30 to ~ 200 °C, encompassing the capabilities of both
sampling methods. The Method 8240 target analyte list was modified/expanded to include
additional potential PICs.
The resulting GC/MS total ion chromatograms were analyzed to identify peaks that were not target
analytes. Nontarget PICs were identified by comparing spectral data of the unknown to spectral
data contained in the National Institute of Standards and Technology (NIST) and Wiley mass
spectral databases. A probability-based spectral matching algorithm assigned tentative
2-10
-------
identification. The quality of the match, along with investigator spectral interpretation and physical
data (e.g., boiling point vs. retention time) was used to assist in identification. Where possible,
additional standards containing tentatively identified compounds were prepared and analyzed to
confirm identification. Following Method 8240, these unknowns are quantified based on the
internal standard closest in retention time and a relative response factor (RRF) of 1. A
multiconcentration calibration was performed using standards of the identified compounds to
establish RRFs specific to each compound to enhance quantitative accuracy.
The Tedlar bag samples were also analyzed to characterize the highly volatile organic species. The
bag samples were analyzed by gas chromatography/flame ionization detector (GC/FID) to quantify
such compounds as methane, ethane, propane, chloromethane, and acetylene. The FID response to
nontarget analytes was also reported.
2.5.3 - Semivolatile and Non-Volatile Organics
A detailed chemical characterization was performed on the MM5 samples. MM5 analyses were
performed quantitatively; however, the main emphasis was on qualitative identification of major
emissions components.
Following collection, the MM5 samples were Soxhlet extracted sequentially with several solvents
of decreasing polarity. The samples were extracted sequentially with methylene chloride, acetone,
and toluene. The individual sample extracts were concentrated to a known volume and archived
for analysis. The five containers from each sample train were extracted so as to generate three
separate sample components. For each solvent, separate sample extracts were generated from each
train. The filter and front-half rinse (Containers 1 and 2) were composited as a single extract as
were the XAD-2 sorbent and back-half rinse (Containers 3 and 5). The condensate and condensate
rinse (Container 4) is the third sample component. For methylene chloride, the extractions were
performed as described in SW-846 Draft Method 5060, "Preparation of MM5 Train Components
for Analysis by SW-846 Method 8270." The acetone and toluene extractions were performed
similarly with only the filters and XAD-2 being extracted. Surrogates were added only to the MM5
train components.
After initial analyses were performed using conventional GC/MS, and significant unidentified
peaks were found, an alternative analytical approach was taken. The methylene chloride extracts
from Run 10 were sent to the University of Dayton Research Institute (UDRI), where the
technique of multi-dimensional GC/MS (MDGC/MS) was used to further characterize the samples.
The MDGC/MS system used!8,19 js shown in Figure 2-5. The uniqueness of the MDGC
technique lies in the ability to further resolve coeluting peaks from the primary column on a
secondary column. This system uses a "Deans switching mechanism" for obtaining narrow
fractions (heartcuts) from a primary chromatogram. It uses a low-temperature cryogenically cooled
trap at -80 °C and uses two 30 m X 0.25 mm open tube columns (OTCs) with a 0.25 |im film
thickness. The primary column contained a non-polar 5% phenylmethylsiloxane stationary phase,
while the secondary column used a moderately polar 1701 cyanosiloxane stationary phase. Using
the second column with a stationary phase of differing polarity enables better separation of
compounds that were not cleanly separated in the first column. The effluent from the secondary
OTC was passed directly into an HP 5970B mass selective detector. Both OTCs were mounted
inside an HP 5890 GC system.
2-11
-------
2.5.4 - PCDDs/PCDFs
PCDDs/PCDFs were quantified from the Method 23 sampling train. This procedure is described in
CARS Method 428. The PCDD/PCDF analyses were performed as described in Method 23 with
only one exception: the analyses were performed by low resolution mass spectrometry (LRMS) as
opposed to high resolution mass spectrometry (HRMS). The use of LRMS can generally quantify
only different PCDD/PCDF congener groups, rather than individual isomers within the congener
groups.
Figure 2-5. MDGC-MS Setup (Copyright © 1996; reproduced with permission of UDRI).
2-11
-------
3.0 RESULTS AND DISCUSSION
3.1 - Results from Continuous Measurements
Results from temperature measurements made during the incineration tests are shown in Table 3-1.
Note that the thermocouple at the kiln exit broke and was not operational for some of the tests.
Also note that we had only mixed success in maintaining constant temperatures in the transition
duct. This inability to hold the duct temperatures constant from run to run impacted our ability to
develop surrogate performance indicators for PCDDs/PCDFs that are explicitly based on only
combustion parameters. The temperatures labeled Duct 1, 2, 3, and 4 represent thermocouples
placed at axial positions in the duct leaving the SCC. The Duct 1 thermocouple is just downstream
of the water jacket, and the Duct 4 thermocouple is near where the extractive sampling was
performed.
Table 3-2 lists the results from the conventional gas CEMs. The columns labeled CO Low and CO
High represent the high- and low-range CO analyzers. For runs where CO values were within the
normal operating range of the CO Low CEM, the data for the high-range CO analyzer were labeled
n/a. The high concentration of acid gases damaged both THC CEMs, eventually resulting in the
failure of both instruments (note the n/a's near the end of the test matrix). The HC1 CEM was not
available for the test days during August 1995. Note that the Duct CO2 concentrations are
approximately 50% of the CO2 concentrations at the SCC Exit. This is due to dilution air's being
added in the transition duct leading to the FGCS. Extractive samples were sampled downstream of
the addition of dilution air.
Table 3-1. Temperature Results (°C)
Run Kiln SCC Mix SCC Exit Duct 1 Duct 2 Duct 3 Duct 4
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
16
886
865
698
618
470
554
532
495
n/a
n/a
n/a
n/a
911
939
925
945
890
851
796
778
592
632
517
492
485
457
697
695
567
612
574
562
1006
1054
1049
1007
863
932
497
459
624
578
845
899
856
867
836
848
674
701
712
681
548
589
339
313
387
352
488
548
524
520
554
559
534
552
568
543
433
462
280
264
302
272
369
406
370
362
397
403
327
334
348
333
259
277
218
210
228
211
280
297
279
263
312
312
301
305
320
307
236
251
193
186
193
181
243
258
240
230
272
273
3-1
-------
t«
3
t«
c
's
O
1
CO
H
O
es
a
la
Q tn
es
O
o
o
o
be
3
o
o
g3
o o
oo O
CS
O
r )
s^
es
O
Sw/
O
i
bfi
3
o
o
w ^
£ o
^H O
es
O
€-
I
I
g
I
^
^
1
I
b
|
^3-
^
^
Os en
in os
"° ^
O es
OS OO
*O OS
o o
oo o
II
^ VO
oq p
i—i en
cs in
0 0
O r-
OS ^H
^H es
C3
II
en t-
en es
oo
^H OO
es cs
en os
•*' en
os in
en cs
OS Tt
•n en
es o
en os
en es
^H O
•* t-
in ^o
II
o t-
en cs
en f-
^o in
0 ^-i
OS \O
o oo
0 ^H
O es
•n -st
II
•n oo
CS CS
OS OS
^H CS
0 0
en in
f- t-
t-' t-'
rr-i ^t-
in o
in \d
in f-
os t-
es in
cs cs
•* t-
in in
o o
^H OO
t- t-
II
es ^H
^H es
OS OO
oo oo
•*t OS
•n oo
oo oo
2^
OS Tt
II
t-
en o
oo oo
\o in
^ 0
OS ^H
in ^n
^H OO
es t-;
en en
en os
oo oo
f- cs
es en
CS OS
^d in
0
^H O
in in
es es
II
O OS
in oo
"* CS
en Tt
en es
OS O
II
•n in
en es
cs
2 S
I— H QKJ
oo en
^o r —
t-' OS
i — tin
OS ^H
**. °°.
-t ^
en Os
•* 00
es es
^d in
oo o
OS O
es os
cs es
in Tt
O ^H
O O
0
O en
^H es
in os
in \o
^H OO
en es
t- en
CS OS
^O OO
in
rt f^
^ 0
oo
r~~~ ^^
r*^ ^^
oo en
in f-
in ^^
t- cs
t-. p
en in
O
\o in
^H OS
^o oo
^H O
in Tt
in
^H OS
OS
in o
oo en
es
cs en
O O
00 f-
OS ^H
\O OS
es oq
en f-
en -st
en o
\o oo
^H O
•n es
^O OS
o o
t- •— 1
^H OS
OS OO
en oo
^t- os
es >— !
t-; OO
en en
^H es
^ ^
ss
o o
o o
en oo
oo o
^H es
•n ^H
es ^H
O O
o o
t- f-
i— ! en
^O es
CS r~^
0 0
^H es
es cs
oo es
^O OO
o o
O es
t- •— 1
t- OS
OS
CS ^O
en Tt
[-; OS
es o
en Tt
^ ^
ss
o o
II
in es
cs es
en in
O O
OS ^H
en es
t- oo
o in
^o o
"* t-
II
o in
t- •— i
cs es
•n oo
en es
O O
oo cs
•n in
en in
es es
oo in
p •*_
^h es
in ^o
CD
JD
CD
CD
"o
m
3-2
-------
3.2 - Volatile Organic Results
3.2.1 - On-Line GC Results
OLGC sampling results are shown in Table 3-3. There are several interesting observations made
from these measurements. First, all of the samples taken while no waste was being fed into the
RKIS still showed measurable levels of many of the OLGC target analytes. This is likely due to
residual contamination of the RKIS itself with some of the chlorinated PICs of interest.
Another observation is that the measurements made at the SCC choke are generally higher than the
measurements at the SCC exit, particularly with respect to the chlorinated target analytes. The
exception is on Runs 7 and 8, where the SCC's afterburner was off. It is likely that some ring
growth was occurring as the gases from the kiln passed through the SCC when no flame was
present in the SCC. This observation is illustrated in Figure 3-1, showing the concentration of
tetrachloroethylene. Measured values of tetrachloroethene at the choke are consistently higher than
at the SCC exit. Figure 3-2 shows this observation for 1,2-dichlorobenzene, a potentially
important precursor to PCDDs/PCDFs. Note how the concentration of 1,2-dichlorobenzene is
higher at the SCC exit for those runs where the afterburner was off. This shows the potential for
significant ring growth to occur in the moderate temperature region of incinerators after the
combustion sections, but prior to any heat recovery or rapid quenching.
80-
70-
co
O)
c
_CD
"CD
60-
0-
Choke
SCC Exit
Run
Figure 3-1. OLGC results of tetrachloroethylene concentrations at choke and SCC exit.
3-3
-------
o
HH
fin
O
'8
ctf
1
cn
-^—*
3
.£5
ex
sj
ob
o
ta
o
J3
O
o
6
9U9zu9qojojipip £'j
Y& d'ra
9U9ZU
gpuoyqo
O O
CS i—
CS CS CS
O
om
om
O
oomm
o
^
CL>
W
I
o
o
I
o
I
t
C3
o
1
8
O
O
CO
3-4
-------
160
7 7
7 8
Choke
SCO Exit
Run
Figure 3-2. OLGC results of 1,2 dichlorobenzene concentrations at choke and SCC exit.
3.2.2 - VOST and Tedlar Bag Results
Analyses of the Tedlar bag samples for Ci and C2 non-halogenated alkanes, alkenes, and alkynes
resulted in none of those compounds being detected. Estimated minimum detection limits are on
the order of 1 - 2 ppm, and apparently none of these compounds were present at these levels. This
is consistent, however, with our measured THC concentrations on the order of 1-2 ppm.
The VOST and Tedlar bag analytical results indicate that a significant number of VOC PICs have
been identified both as target analytes and as tentatively identified compounds (TICs). Table 3-4
shows the Tedlar bag results for the target analytes, and Table 3-5 shows the Tedlar bag TIC
results. VOST results are shown in Table 3-6. The VOST target analyte results are displayed
qualitatively in Table 3-7, showing which of the VOST target analytes were detected. Table 3-8
qualitatively lists the VOST TICs. Although differences exist in quantitation levels between VOST
and the Tedlar bags, it must be remembered that VOST samples are taken over longer periods of
time. Of the 44 target analytes, 38 were detected. It should be noted that several of these
compounds are POHCs. Over 50 nontarget analytes were tentatively identified as PICs.
However, a large number of PICs present in the VOST samples were not identified. To aid in
perspective, at least 82 compounds were detected in a single sample. Of those, 28 were identified
as target analytes, 21 were tentatively identified, and 33 remained unidentified.
3-5
-------
Table 3-4. Tedlar Bag Results: Target Compounds (jig/m^)
Run
chloromethane
vinyl chloride
bromomethane
chloroethane
1,1-dichloroethene
iodomethane
carbon disulfide
acetone
methylene chloride
1,2-dichloroethene
1,1 -di chloroethane
chloroform
1,2-di chloroethane
2-butanone
1,1,1 -tri chloroethane
carbon tetrachloride
benzene
trichloroethene
1 ,2-dichloropropane
dibromomethane
bromo dichloromethane
cis-l,3-dichloropropene
2-hexanone
trans- 1 ,3 -dichloropropene
1 , 1 ,2-tri chloroethane
dibromochloro-methane
1,2-dibromoethane
bromoform
4-methyl-2-pentanone
toluene
tetrachloroethene
chlorobenzene
ethylbenzene
1,1,1 ,2-tetrachloroethane
m,p-xylene
o-xylene
styrene
1 , 1 ,2,2-tetrachloroethane
1,2,3 -tri chloropropane
trans- l,4-dichloro-2-butene
pentachloroethane
1 ,2-dibromo-3 -chloropropane
5
18742
ND
8304
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
569
ND
75
232
19
ND
ND
15
ND
ND
ND
ND
ND
ND
15
ND
27
ND
ND
9
ND
30
12
14
ND
ND
ND
ND
ND
6
2827
ND
667
18
ND
ND
ND
1703
116
ND
ND
50
ND
232
ND
65
64
ND
ND
5
16
ND
ND
ND
ND
17
7
22
ND
16
ND
20
3
ND
8
4
4
ND
ND
ND
ND
ND
9
1
ND
ND
ND
ND
ND
6
137
ND
ND
ND
ND
ND
46
ND
ND
5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
10
ND
ND
ND
ND
ND
14
55
ND
29
ND
ND
ND
33
288
28
ND
ND
ND
ND
28
ND
ND
11
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
21
20
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3-6
-------
Table 3-5. Tedlar Bag Results: Tentatively Identified Compounds (jig/
Run 5 69 14
1,1-dimethoxy ethane ND 489 ND ND
1,2-dichlorobenzene ND 71 ND ND
1-pentene 3-methyl 2-ethyl ND ND 19 357
1-phenyl ethanone 443 296 ND ND
2-methyl 1-propene ND ND 9 175
2-nitrophenol 81 ND ND ND
3-methylene pentane ND ND ND 130
3-methyl heptane ND ND 6 80
3-methyl pentane ND ND 20 2490
3-methylene nonane ND ND 7 ND
acetaldehyde ND ND ND 164
benzaldehyde 481 218 ND ND
benzoic acid methyl ester 325 68 ND ND
benzonitrile 122 194 ND ND
cyclohexane ND ND ND 339
dodecane ND ND 17 ND
hexane 2402 ND 661 3904
methyl cyclopentane 4850 ND ND 4896
nitromethane 1223 ND ND ND
tetrahydrofuran 7836 ND 96 ND
tridecane ND 61 ND ND
trimethyl hexane ND ND 77 ND
undecane 103 ND 32 ND
3-7
-------
Table 3-6
dichlorodifluoromethane
chloromethane
vinyl chloride
bromomethane
chloroethane
trichlorofluoromethane
1,1-dichloroethene
iodomethane
carbon disulfide
acetone
methylene chloride
1 ,2-dichloroethene
1,1-dichloroethane
chloroform
co 1,2-dichloroethane
00 2-butanone
1,1,1 -trichloroethane
carbon tetrachloride
benzene
trichloroethene
1 ,2-dichloropropane
dibromomethane
bromodichloromethane
cis- 1 , 3 -dichloropropene
2-hexanone
trans- 1 ,3 -dichloropropene
1 , 1 ,2-trichloroethane
dibromochloromethane
1 ,2-dibromoethane
bromoform
5
819.6
1066
1.7
197.4
7.5
5.9
6.6
0.3
17.4
15.8
26.9
1.0
ND
15.9
0.7
2.6
0.3
6.7
1.9
1.7
ND
1.4
2.4
ND
ND
ND
0.1
3.1
1.6
16.7
6
418.5
2.0
0.1
6.2
0.4
12.3
ND
0.3
2.5
8.4
42.7
ND
0.1
26.7
0.4
2.4
0.5
22.7
1.1
0.3
0.1
0.8
2.6
0.1
0.2
0.2
ND
3.3
8.0
19.4
7
50.2
0.8
0.7
4.1
0.1
0.6
0.9
ND
1.3
7.1
10.3
0.2
ND
5.5
ND
0.6
0.1
5.0
1.2
0.7
ND
4.0
5.3
ND
ND
ND
ND
6.4
0.3
37.8
8
257.9
4.8
0.9
9.9
1.7
0.7
2.0
ND
4.7
6.0
15.8
0.4
0.1
7.7
0.2
0.6
0.2
29.6
2.1
2.0
0.1
12.2
9.5
ND
ND
ND
ND
14.2
1.9
44.1
9
2.3
12.6
11.3
27.6
0.8
0.3
46.3
ND
1.7
2.4
3.3
15.9
ND
9.9
0.2
ND
0.1
152.4
1.4
40.1
ND
3.8
35.1
0.1
ND
0.4
ND
43.4
1.9
73.0
. VOST Results (Jlg/m3)
Run
10
0.1
2.7
42.3
1.7
0.1
ND
15.4
ND
0.3
0.3
2.3
1.9
ND
6.3
1.0
1.0
0.1
297.9
268.9
96.2
ND
201.1
165.2
ND
ND
ND
ND
223.2
109.7
817.4
11
0.6
5.6
1.8
12.5
1.1
0.03
1.9
0.1
4.6
3.2
3.8
3.6
0.03
11.7
1.1
0.7
0.1
8.5
26.0
13.8
0.03
7.4
12.6
0.2
ND
0.3
ND
11.2
1.9
40.7
12
14.7
142.8
94.8
74.3
2.3
0.2
56.1
0.2
3.6
37.3
59.5
44.3
0.1
19.5
11.0
0.7
0.2
13.9
7.0
54.2
0.1
11.5
13.0
0.2
ND
0.3
0.7
5.1
7.6
18.0
13
2.3
16.2
4.5
2.3
0.1
ND
0.7
0.1
1.2
1.6
10.5
1.1
ND
2.1
1.4
0.7
ND
1.4
36.6
24.4
1.7
1.7
2.2
ND
ND
ND
0.1
0.4
1.2
0.2
14
11.4
3.5
3.6
2.4
0.4
ND
0.6
ND
1.5
2.4
13.5
4.6
ND
4.8
1.0
0.7
ND
1.9
226.7
32.3
1.9
0.8
4.4
ND
0.4
ND
ND
2.1
0.5
0.4
15
169.0
41.5
44.0
12.0
1.7
0.1
5.6
ND
81.7
4.6
73.6
26.6
0.1
19.9
5.5
ND
ND
24.5
218.6
52.4
0.1
8.3
19.7
0.7
ND
0.6
ND
15.0
62.8
31.7
16
1.9
19.1
20.0
11.0
0.9
0.1
4.8
0.6
31.6
3.5
39.0
23.5
0.1
30.8
8.4
1.3
0.1
28.8
221.2
54.8
0.2
9.2
27.5
0.9
ND
0.8
ND
26.4
39.6
43.9
(continued)
-------
Table 3-6 (cont). VOST Results (Jig/m3)
I
CD
4-methyl-2-pentanone
toluene
tetrachloroethene
chlorobenzene
ethylbenzene
1,1, 1,2-tetrachloroethane
m,p-xylene
o-xylene
styrene
1 , 1 ,2,2-tetrachloroethane
1,2,3 -trichloropropane
trans- 1 ,4-dichloro-2-butene
pentachloroethane
l,2-dibromo-3-chloropropane
5
ND
15.6
5.4
3.5
1.6
ND
5.4
1.6
0.2
ND
ND
ND
ND
0.1
6
0.9
32.6
1.5
36.3
7.2
0.1
23.2
9.7
ND
0.2
0.3
ND
0.1
0.4
7
ND
2.0
1.0
2.1
0.2
ND
0.6
0.2
0.03
0.1
ND
ND
ND
ND
8
ND
2.2
5.8
3.9
0.1
ND
0.4
0.1
0.2
0.1
ND
ND
ND
0.4
9
ND
1.4
97.0
46.3
0.4
0.1
1.5
ND
ND
0.3
ND
ND
ND
0.8
Run
10
ND
2.6
750.5
8530
ND
ND
3.0
1.2
ND
ND
ND
ND
ND
237.1
11
ND
0.8
20.7
4.3
0.2
ND
0.8
0.3
0.1
0.2
ND
ND
ND
ND
12
ND
24.1
90.2
15.3
0.2
ND
0.6
ND
ND
0.03
ND
ND
ND
1.3
13
ND
2.8
25.1
5.8
1.3
ND
5.8
0.9
1.8
ND
ND
0.2
ND
ND
14
ND
1.5
31.7
4.7
ND
ND
0.6
ND
1.4
ND
ND
ND
ND
ND
15
ND
23.1
52.5
75.6
0.9
ND
1.3
0.6
12.0
0.1
ND
4.8
ND
ND
16
ND
25.1
59.4
72.0
1.3
ND
2.0
0.9
23.3
ND
ND
13.6
ND
ND
-------
Table 3-7. Target Volatile Organic Compounds Detected
Dichlorodifluoromethane
Chloromethane
Vinyl chloride
Bromomethane
Chloroethane
Trichlorotrifluoromethane
1,1 -Dichloroethene
Carbon disulfide
Acetone
Methylene chloride
1,2-Dichloroethene
1,1 -Dichloroethane
Chloroform
1,2-Dichloroethane
2-Butanone
1,1,1 -Trichloroethane
Carbon tetrachloride
Benzene
Trichloroethene
1,2-Dichloropropane
Dibromomethane
Bromodichloromethane
cis-l,3-Dichloropropene
trans-l,3-Dichloropropene
Dibromochloromethane
1,2-Dibromoethane
Bromoform
4-Methyl-2-pentanone
Toluene
Tetrachloroethane
Chlorobenzene
Ethylbenzene
1,1,1,2-Tetrachloroethane
Xylene (m, p)
Xylene (o)
Styrene
trans-l,4-Dichloro-2-butene
1,2-Dibromo-3 -chloropropane
3-10
-------
Table 3-8. Tentatively Identified VOST Compounds
Bromotrichloromethane
Chloroethyne
Bromoethyne
Bromochloroethyne
Dichloroethyne
Bromoethene
Bromochloroethene
Dibromoethene
Bromodichloroethene
Dibromochloroethene
Tribromoethene
Bromotrichloroethene
Tribromochloroethene
Dibromodichloroethene
Tetrabromoethene
Bromochloroethane
Bromopropyne
Bromochloropropyne
Bromodichloropropyne
Bromopropene
Pentachloropropene
Dibromopropane
Hexachlorobutadiene
Pentachlorobutadiene
Chlorobutane
Bromoheptane
Chlorooctane
Benzylchloride
Bromobenzene
Bromomethylbenzene
Bromdimethylbenzene
Bromochlorobenzene
Dibromobenzene
Bromodichlorobenzene
Propene
Methyl propene
Methyl butane
Butadiyne
Butadiene
Pentene
Pentane
Hexene
Hexane
Methylcyclohexane
Heptane
Methylheptane
Dimethylheptane
Octane
Nonane
Decane
Methyldecane
Undecane
Methylfuran
Benzaldehyde
Methylpentenal
Benzonitrile
Chlorothiophene
Tetrachlorothiophene
Dibromothiophene
3-11
-------
An interesting comparison was made of the Ci and C2 halogenated alkanes, alkenes, and alkynes.
A table was made of the possible chloro, bromo, and mixed bromochloro organics with one and
two carbons (Table 3-9). With only a few exceptions, each compound was detected in at least one
sample. These Ci and C2 compounds are of particular interest: they are considered to be
precursors in aromatic ring propagation reactions leading to higher molecular wei
Table 3-9. Combinations of Detected Ci and C2 Compounds
Target Analyte
Compound Detected
Hydrocarbons
C2 Alkynes
C2 Alkenes
chloromethane
bromomethane
dichloromethane
dibromomethane
bromochloromethane
trichloromethane
tribromomethane
bromodichloromethane
dibromochloromethane
tetrachloromethane
tetrabromomethane
bromotri chl oromethane
dibromodi chloromethane
tribromochl oromethane
chloroethyne
bromoethyne
dichloroethyne
dibromoethyne
bromochloroethyne
chloroethene
bromoethene
dichloroethene (total)
dibromoethene
bromochloroethene
trichloroethene
tribromoethene
bromodichloroethene
dibromochloroethene
tetrachloroethene
tetrabromoethene
bromotrichloroethene
dibromodichloroethene
tribromochloroethene
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
Yes
No
Yes
No
No
Yes
No
No
No
Yes
No
No
No
No
.a
(continued)
3-11
-------
Table 3-9 (cont). Combinations of Detected Ci and C2 Compounds
Target Analyte Compound Detected
C2 Alkanes
chloroethane Yes
bromoethane No
dichloroethane Yes •
dibromoethane Yes •
bromochloroethane No •
trichloroethane Yes •
tribromoethane No
bromodichloroethane No
dibromochloroethane No
tetrachloroethane Yes
tetrabromoethane No
bromotrichloroethane No
dibromodichloroethane No
tribromochloroethane No
a - Detected, but not quantified.
The results of analysis (from both VOST and Tedlar bags) for halogenated Ci and C2 VOCs are
listed in Table 3-10. The list contains possible chloro, bromo, and bromochloro organics with one
or two carbons. Note that di chloromethane (CH2C12) was found as a contaminant in some of the
blanks, possibly as a laboratory contaminant. Also, it is not known why the chloromethane
concentration was so high on one of the VOST tubes for Run 5. The Tedlar bag measurements of
chloromethane were also very high for that run. An interesting observation is that, with few
exceptions, almost all of these possible compounds were detected in at least one of the runs. If the
data are further analyzed, by simply averaging the concentrations of all identified compounds for
all of the reported runs, Figure 3-3 can be constructed. Figure 3-3 shows the concentrations of
some of the halogenated Ci and C2 compounds grouped together, with the chlorinated and
brominated analogs compared side by side. Note that the concentrations of the brominated and
chlorinated analogs are similar in most cases, even though Br was present in the feed at a mass
fraction of only about 10 % of the level of the CI. This observation indicates that the presence of
relatively small amounts of Br can potentially produce quantities of brominated PICs at levels
comparable to those of the chlorinated PICs. Table 3-10 also shows that significant quantities of
mixed bromochloro PICs were also measured. These low-carbon halogenated PICs are
participants in aromatic ring growth reactions leading to the larger organic PIC molecules, such as
the chlorinated benzenes and phenols, and possibly PCDDs/PCDFs.
Table 3-11 lists the concentrations of the aromatic VOCs found in the tests. Although the aromatic
compounds are not identified as commonly throughout all the runs as the smaller molecules were, a
similar pattern is found. The data from Run 10, which had the increased Br feed concentration,
show the highest concentration and highest number of identified aromatic brominated and
bromochloro PICs. The concentrations of brominated compounds are generally on the same order
of magnitude as their chlorinated analogs.
3-13
-------
Table
Compound
chloromethane
bromomethane
dichloromethane
dibromomethane
bromochloromethane
trichloromethane
tribromomethane
bromodichloromethane
dibromochloromethane
tetrachloromethane
tetrabromomethane
bromotrichloromethane
dibromodi chloromethane
tribromochloromethane
chloroethyne
bromoethyne
di chloroethyne
dibromoethyne
bromochloroethyne
chloroethene
bromoethene
dichloroethene (total)
dibromoethene
bromochloroethene
trichloroethene
tribromoethene
bromodichloroethene
dibromochl oroethene
tetrachloroethene
tetrabromoethene
bromotrichloroethene
dibromodichloroethene
tribromochloroethene
chloroethane
bromoethane
dichloroethane
dibromoethane
bromochloroethane
3-10. Ci & C2 Halogenated Hydrocarbons (|ig/m3)
Run 5
1066
197
26.9
1.35
0
16.0
16.7
2.4
3.1
6.7
0
0
0
0
0
0
0
0
0
1.7
0
7.5
0
0
1.7
0
0
0
5.5
0
0
0
0
7.5
0
0.8
1.6
0
Run 6
2.6
7.7
69.3
1.3
0
24
30.7
2.2
6.0
19.2
0
0
0
0
0
0
0
0
0
0.2
0
0.15
2.1
0
0.5
0
0
0
1.9
0
0
0
0
0.5
0
0.5
6.3
2.3
Run 9
12.7
27.6
3.3
3.8
0
9.9
73
35.1
43.4
152
0
12.2
0
0
0
0
0
0
0
11.25
0
62.2
0
2.3
40.1
0
0
0
0
0
2.4
0
0
0.8
0
0.2
1.9
7.5
Run 10
2.8
1.8
2.3
208
0
6.6
846
171
231
308
0
42.1
0
0
0
0.8
0
0
0
43.8
2.6
17.9
0.8
46.7
99.6
3.1
0
0
0
0
0
5.2
0
0.1
0
1
114
187
Run 13
16.3
2.3
10.5
1.7
0
2.1
0.17
2.2
0.4
1.4
0
0
0
0
21.1
13.4
0
0
6.5
4.5
0
1.8
0
1.5
24.4
8.5
0
14.8
0
0
0
32.4
0
0.1
0
1.4
1.8
2.8
Run 14
3.5
2.4
13.6
0.8
0
4.8
0.4
4.4
2.1
1.9
0
0
0
0
10.6
9.1
0
0
2.3
3.6
0
5.2
6.7
0
32.5
9.1
38.6
25.9
31.9
0
0
28.9
9.5
0.4
0
1
0.5
0
(continued)
3-14
-------
Table 3-10 (cont). Ci & C2 Halogenated Hydrocarbons (|ig/m3)
Compound
Run 5
Run 6
Run 9
Run 10 Run 13 Run 14
trichloroethane
tribromoethane
bromodichloroethane
dibromochl oroethene
tetrachloroethane
tetrabromoethane
bromotrichloroethane
dibromodichloroethane
tribromochloroethane
0.5
NDa
ND
ND
ND
ND
ND
ND
ND
0.6
ND
ND
ND
0.2
ND
ND
ND
ND
0.1
ND
ND
ND
97.1
ND
ND
ND
ND
0.2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
11.2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
a - none detected
¥ 1RO :
^_
"r- 140- "
O :
CD 120.
"c inn ~
CD IUU -
£ Rn:
o
o fin "
CD
CD 40 .
CD on "
II r
Jl
n
._ • 1
0~T i i i i i i i i i i i i i i i i
CDCD CDCD CDCD CDCD CDCD ^
cc cc cc cc cc CD
CD CD CD CD CD CD >, >• CD CD "Q
"CDCD CDCD CDCD CDCD "CDCD .,,
EE EE EE 2° 2g c
OO OO OO O t o ^ CD
oE o E o E Z2 Z2 £
z 2 z 2 z 2 "^ ^
1
CD
CD
r
aibromoet
,
i i
CD
CD
r
I
CD
C
CD
trichloroel
tribromoel
|
1 1
CD
C
CD
.C
0
O
_0
o
o
1
CD
C
CD
.C
nochloroel
,-, , , 1
i i i i i
CD CD CD CD
C C C C
CD CD CD CD
-C-C .C.C
"CD CD CD CD
o o o o
£E feE
o^ o^
T3T3
- J=
E2
O
T3
>_i
Figure 3-3. Average concentrations of analogous Q and C2 halogenated compounds
3-15
-------
Table 3-11. Halogenated Aromatic VOC Results (|ig/m3)
Baseline
Test Condition
chlorobenzene
bromobenzene
dichlorobenzene
bromochlorobenzene
dibromobenzene
trichlorobenzene
bromodichlorobenzene
dibromochlorobenzene
tribromobenzene
bromomethylbenzene
bromodimethylbenzene
5
4.2
0
0
0
0
0
0
0
0
3.0
0
5
2.7
0
0
0
0
0
0
0
0
0
0
6
36.3
0
3.8
0
0
0
0
0
0
1.3
5.7
6
14.8
0
8.1
0
0
0
0
0
0
0
1.6
Lov
9
16.8
0
0
0
0
2.6
0
0
0
0
0
/SCC1
9
75.8
0
36.7
3.9
0
8.7
0
0
0
0
0
"emp
10a
8828
13.2
37.4
38.1
4.0
32.7
5.7
0
0
0
0
|| SCC Fuel Rich |
13
3.7
0
3.5
0
0
0
0
0
0
0
0
13
0
0
0
0
0
0
0
0
0
0
0
13
13.7
1.6
2.8
0.6
0
0
0
0
0
0
0
14
4.7
0
0.9
0
0
0
0
0
0
0
0
a - on Run 10, too much ethylene dibromide was inadvertently added
3.3 - Semivolatile and Non-Volatile Organics
3.3.1 - Conventional GC/MS Analytical Results
The semivolatile organic analytical results of the methylene chloride extracts indicate that a
significant number of PICs have been identified both as target analytes and as TICs. For the
analytical data evaluated, PICs identified as target analytes and TICs are presented in Tables 3-12
and 3-13, respectively. Many of the target analytes were detected. It should be noted once again
that several of these compounds were in the original surrogate waste feed. Over 50 nontarget
analytes were tentatively identified as PICs. Many of the PICs present in the MM5 samples were
not identified. Also, the mix of PICs found on the filter sample fraction differed from that of the
XAD-2 sample fraction. For a selected filter sample, at least 174 compounds were detected: 25
were identified as target analytes, 11 were tentatively identified, and 138 remained unidentified.
For a selected XAD-2 sample, at least 194 compounds were detected: 18 were identified as target
analytes, 17 were tentatively identified, and 159 remained unidentified. Identification of non-target
analytes was particularly complicated by coeluting compounds. Coeluting compounds result in
combined mass spectra that cannot be compared easily to reference spectra.
Many of the TICs were oxygenated compounds, such as esters, aldehydes, diones, and carboxylic
acids. There were also many brominated TICs. There were also a significant number of
unidentifiable aliphatic hydrocarbons, silanes, and phthalates that were not reported in Table 3-13.
Silanes are frequently found as chromatographic artifacts from degradation of GC columns.
Phthalates are commonly found in combustor emissions, but it is not well-established whether they
are actual PICs or artifacts resulting from sampling and analytical treatments.
Analysis of the acetone and toluene sample extracts did not result in the identification of additional
compounds. These analyses do verify the acceptable performance of methyl ene chloride as the
single extraction solvent.
3-16
-------
CO
Table 3-12.
Run
N-methyl-N-nitroso-ethanamine
Bis(2-chloroethyl)ether
Aniline
Phenol
2-Chlorophenol
1,3-Dichlorobenzene
1,4-Dichlorobenzene
1,2-Dichlorobenzene
Benzyl alcohol
Bis(2-chloroisopropyl) ether
2-Methylphenol
Acetophenone
Hexachloroethane
N-Nitrosodipropylamine
Nitrobenzene
1-Nitrosopiperidine
Isophorone
2,4-Dimethylphenol
Bis(2-chloroethoxy)methane
2,4-Dichlorophenol
1,2,4-Trichlorobenzene
Naphthalene
2-Nitrophenol
2,6-Dichlorophenol
Hexachloropropene
4-Chloroaniline
Hexachlorobutadiene
N-Butyl-N-nitroso-butanamine
4-Chloro-3 -methyl-phenol
2-Methylnaphthalene
1,2,4,5-Tetrachlorobenzene
Hexachlorocyclopentadiene
1
ND
ND
ND
0.3
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
2
ND
ND
ND
0.8
ND
ND
ND
ND
0.5
ND
0.2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.3
ND
ND
ND
ND
ND
ND
ND
0.2
0.2
ND
Semivolatile Organic
3
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
4
ND
ND
ND
0.5
ND
ND
ND
0.6
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.2
1.6
ND
ND
ND
ND
ND
ND
ND
0.8
0.7
0.7
5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Target Results (jig/m3)
6
ND
ND
ND
0.5
ND
ND
ND
21.9
ND
ND
1.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.2
7.7
ND
ND
ND
ND
ND
ND
ND
1.3
0.2
ND
9
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
10
ND
ND
ND
ND
ND
ND
ND
1.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.0
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
11
ND
ND
ND
4.0
ND
0.4
0.3
0.7
ND
ND
0.3
6.2
ND
ND
ND
ND
ND
ND
ND
ND
2.5
110
0.2
ND
ND
ND
ND
ND
ND
1.6
4.5
ND
13
ND
ND
ND
ND
ND
1.8
1.0
29.3
1.6
ND
0.3
1.4
ND
ND
0.4
ND
ND
ND
ND
ND
3.6
21.9
ND
ND
ND
ND
ND
ND
ND
0.4
1.0
0.2
14
ND
ND
ND
ND
ND
0.5
0.2
0.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.0
3.5
ND
ND
ND
ND
ND
ND
ND
ND
1.2
ND
15
ND
ND
ND
ND
ND
1.3
1.0
2.7
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3.1
107
ND
ND
ND
ND
0.4
ND
ND
ND
39.5
7.7
(continued)
-------
CO
00
Table 3-12. Semivolatile
Run
2,4,6-Trichlorophenol
2,4,5-Trichlorophenol
2-Chloronaphthalene
2-Nitroaniline
3-Nitroaniline
Acenaphthylene
1 ,4-Naphthoquinone
Dimethylphthalate
2,6-Dinitrotoluene
Acenaphthene
4-Nitroaniline
2,4-Dinitrophenol
Dibenzofuran
Pentachlorobenzene
2,4-Dinitrotoluene
2,3 ,4,6-Tetrachlorophenol
4-Nitrophenol
Fluorene
Diethyl phthalate
4-Chlorophenyl phenyl ether
2-Methyl-4,6-dinitrophenol
Diphenylamine
4-Bromophenyl phenyl ether
Phenacetin
Hexachlorobenzene
Pentachlorophenol
Pentachloronitrobenzene
Phenanthrene
Anthracene
Dibutyl phthalate
Fluoranthene
Pyrene
1
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
35.4
ND
ND
2
0.3
0.5
ND
ND
ND
ND
ND
0.2
ND
ND
ND
ND
ND
1.2
ND
0.3
ND
ND
2.2
ND
ND
ND
ND
ND
ND
ND
ND
0.2
ND
41.9
0.2
ND
3
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
58.8
ND
ND
Organic Target Results (continued) (jlg/m^)
4
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.8
ND
0.2
ND
ND
1.9
ND
ND
ND
ND
ND
0.5
ND
ND
ND
ND
91.3
ND
ND
5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.8
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
49.9
ND
ND
6
0.3
0.5
ND
ND
ND
ND
ND
7.4
ND
ND
ND
ND
ND
1.2
ND
0.3
ND
ND
3.1
ND
ND
ND
ND
ND
2.6
ND
ND
ND
ND
69.4
ND
ND
9
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
153
ND
ND
10
0.4
0.5
0.7
ND
ND
ND
4.4
1.6
ND
ND
ND
ND
5.9
0.8
ND
1.8
ND
ND
0.3
ND
ND
ND
ND
ND
2.3
1.4
ND
8.1
ND
42.3
0.8
ND
11
13.1
ND
4.2
ND
ND
1.1
ND
ND
ND
0.7
ND
ND
5.7
29.3
ND
30.9
ND
3.2
4.1
ND
ND
ND
ND
ND
9.4
31.3
ND
26.7
ND
50.0
6.8
0.4
13
2.8
ND
0.9
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.3
6.9
ND
5.4
ND
ND
0.8
ND
ND
ND
ND
ND
3.4
5.5
ND
4.1
ND
12.9
0.7
ND
14
1.2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
2.8
ND
2.5
ND
ND
1.2
ND
ND
ND
ND
ND
2.2
1.4
ND
0.3
ND
13.5
ND
ND
15
30.9
30.5
199
ND
ND
ND
ND
34.6
ND
ND
ND
ND
74.6
128
ND
36.9
ND
ND
ND
ND
ND
ND
ND
ND
18.5
13.8
ND
2.2
ND
3.0
ND
ND
(continued)
-------
Table 3-12. Semivolatile Organic Target Results (continued) (jig/m^)
Run
P-Dimethylaminoazobenzene
Benzyl butyl phthalate
Chrysene
Benzo(a)anthracene
Di-N-octyl phthalate
Benzo(b)fluoranthene
7, 12-Dimethylbenz(a)anthracene
Benzo(k)fluoranthene
Benzo(a)pyrene
3 -Methylcholanthrene
Indeno( 1,2,3 -cd)pyrene
Dibenz(a,h)anthracene
Benzo(ghi)perylene
1
ND
0.2
ND
ND
1.4
ND
ND
ND
ND
ND
ND
ND
ND
2
ND
0.3
ND
0.2
1.4
0.2
ND
0.2
0.2
0.2
0.3
0.2
0.3
3
ND
ND
ND
ND
3.6
ND
ND
ND
ND
ND
ND
ND
ND
4
0.5
0.2
ND
ND
38.9
0.2
ND
0.2
0.3
0.2
0.2
0.3
0.3
5
ND
0.9
ND
ND
0.8
ND
ND
ND
ND
ND
ND
ND
ND
6
ND
ND
ND
ND
278
ND
ND
ND
ND
ND
ND
ND
ND
9
ND
ND
ND
ND
0.3
ND
ND
ND
ND
ND
ND
ND
ND
10
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
11
ND
0.3
ND
ND
448
ND
ND
ND
ND
ND
ND
ND
ND
13
ND
0.8
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
14
ND
ND
ND
ND
146
ND
ND
ND
ND
ND
ND
ND
ND
15
ND
ND
ND
ND
6743
ND
ND
ND
ND
ND
ND
ND
ND
CO
CD
-------
CO
k
o
Table 3-13. Semivolatile Organic Tentatively Identified Compounds (jig/m^)
Run
1, '-Biphenyl
1, '-Biphenyl, 2-phenoxy
1, ,2,2-Tetrabromoethylene
1, ,2-Tribromo-2-chloro-ethylene
1, ,2-Trichloroethane
1 ,2,3 ,4-Tetrachlorobenzene
1 ,2-Dibromo- 1 ,2-dichloroethylene
1 ,2-Dibromo-trans-cyclohexane
1 ,3 -Isobenzofurandione
1 ,4-Dibromo-cyclohexane
1,4-Dimethyl benzene
1,8-Naphthalic anhydride
1 -Bromo- 1 ,2,2-trichloroethylene
1 -Bromo-2-methoxy-,cis-cyclohexane
1 -Bromo-2 -methyl-benzene
1 -Bromo-4-methyl-benzene
1 -Bromo-naphthalene
1-Hexanol, 2-ethyl
lH,3H-Naphtho [1 ,8-cd]pyran- 1 -3 -dione
lH-Isoindole-l,3(2H)-dione
2,2,3 -Tribromobutane
2,4,5-Tribromotoluene
2,5-Cyclohexadiene- 1 ,4-dione
2,5-Dibromothiophene
2,6-Dibromo-p-chlorophenol
2-Butoxy-ethanol
2-Chloro-pyridine
2-Ethyl hexanoic acid
2-Ethyl- 1 -hexanol
4-Bromo-benzonitrile
6-Bromo-l,l,a,6-cycloprop[a]indene
1
ND
0.3
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
2
0.5
ND
ND
ND
0.7
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3
ND
ND
ND
ND
ND
ND
ND
ND
0.7
ND
ND
ND
ND
1.6
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
4
ND
ND
ND
ND
ND
ND
ND
ND
ND
2.4
ND
ND
ND
0.9
ND
ND
ND
3.3
ND
ND
5.9
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.5
5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
6
ND
ND
ND
ND
ND
ND
ND
ND
ND
25.5
5.8
ND
ND
ND
ND
ND
ND
9.6
ND
ND
ND
ND
ND
ND
7.1
ND
ND
0.5
ND
ND
ND
9
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.4
ND
ND
ND
ND
ND
ND
ND
ND
10
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
29.9
ND
ND
ND
ND
ND
12.2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
11
3.2
ND
ND
ND
ND
3.1
ND
ND
ND
23.7
ND
ND
ND
ND
ND
ND
ND
ND
0.5
0.5
ND
7.6
ND
ND
ND
0.4
0.4
ND
11.6
3.1
ND
13
ND
ND
ND
ND
ND
ND
ND
60.0
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
14
ND
ND
5.0
6.0
ND
ND
12.4
ND
ND
42.4
ND
ND
ND
67.6
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
15
3.3
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
8.6
ND
6.7
7.4
7.6
ND
ND
ND
ND
ND
ND
0.5
3.7
ND
ND
ND
ND
ND
ND
(continued)
-------
CO
Table 3-13.
Run
9, 10-Anthracenedione
9-Bromo-anthracene
9H-Fluoren-9-one
Benzaldehyde
Benzaldehyde, ethyl
Benzole acid
Benzole acid, methyl ester
Bromobenzene
Butanoic acid, methyl ester
Chlorobenzene
Cyclohexadecane
Cyclopentanecarboxaldehyde
Decane
Dichlorobromoethene
Ethyl benzene
Hexanedecanoic acid
Hexanedecanoic acid, methyl ester
Hexanedioic acid
Hexanedioic acid, bis(2-ethylhexyl)
Hexanedioic acid, dimethyl ester
Hexanedioic acid, dioctyl ester
Hexanedioic acid, mono(2-ethylhexyl)
Hexanoic acid
Methyl benzene
Nonane
Octadecanoic acid
Octadecanoic acid, methyl ester
Styrene
Tetrachloroethane
Tetrachloroethene
Triacetin
Tribromoethene
Tribromomethane
Trichlorobromobenzene
Semivolatile Organic Tentatively Identified Compounds (continued) (jlg/m^)
1
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
147.5
ND
ND
0.7
ND
ND
ND
ND
ND
ND
ND
2
ND
ND
ND
2.5
0.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.6
ND
ND
ND
ND
ND
61.1
ND
ND
577.3
ND
ND
ND
ND
5.0
ND
ND
ND
ND
ND
3
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
5.4
ND
ND
ND
ND
2.9
ND
ND
ND
ND
ND
4
ND
ND
ND
2.3
ND
5.9
0.9
ND
0.5
ND
ND
ND
2.2
ND
6.1
ND
ND
ND
ND
0.8
ND
ND
ND
840.5
106.4
ND
0.7
ND
ND
ND
ND
ND
ND
ND
5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.4
ND
ND
0.5
ND
ND
ND
20.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
6
ND
ND
ND
8.3
ND
5.7
33.8
ND
ND
6.0
ND
ND
ND
ND
11.3
ND
12.3
ND
ND
ND
ND
ND
ND
1446.6
ND
ND
31.1
ND
ND
ND
ND
ND
ND
ND
9
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.8
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
10
10.4
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
22.0
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
11
ND
0.5
ND
ND
ND
11.2
4.6
ND
ND
ND
ND
0.4
ND
ND
ND
ND
9.9
ND
ND
ND
ND
ND
0.4
202.2
ND
ND
8.2
ND
ND
ND
ND
ND
ND
ND
13
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
4.9
ND
15.1
ND
37.3
18.3
9.7
12.9
ND
ND
ND
ND
ND
ND
21.4
17.2
ND
ND
ND
155.1
11.1
ND
ND
14
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
15
ND
ND
ND
ND
ND
ND
4.2
ND
ND
9.1
ND
ND
ND
ND
ND
ND
10.4
ND
ND
ND
ND
ND
ND
ND
ND
ND
4.8
12.7
ND
4.6
ND
ND
ND
ND
-------
3.3.2 - Multi-Dimensional GC/MS Analytical Results
Figure 3-4 shows the results from the MDGC/MS analysis of the extract from Run 10. A 15
second heartcut (showing a doublet, two closely spaced peaks) and a 10-second heartcut from a
single peak (a singlet) were trapped at low temperature, and then both collected fractions were re-
chromatographed by the secondary OTC (the more polar phase column). Both heartcuts were
chosen by the difficulty of compound identification through MS spectral library searches. The
lower section of Figure 3-4 shows the chromatograms and MS identifications for these two
heartcuts. It is readily apparent that many more compounds were present in both of the heartcuts
than would appear from examination of the primary chromatogram. More importantly, good
separation was obtained by using the second chromatography step, resulting in reliable MS
identifications. This technique verifies that complex samples, such as incinerator emissions,
cannot be fully characterized using conventional techniques due to the problem of compound
coelution. Table 3-14 shows a list of compounds identified from the two heartcut fractions. The
potential benefits of using this technique for detailed examination of each peak of the primary
chromatogram are obvious.
UJ
CO
z
o
Q_
V)
LU
et
" 10 second: singlet
15 second: doublet
D
n
D
i™1
JL
yn
i
OS
Q.
5
6
o
e
o
£
S
£
/
&
83
—
D
S
\
^J
S
a
o
£
?
"^
w
»
i
a
n
c
2
L
s
•M
c
u
a
\
J(
z
Q.
D
/
jM
§
i
g
D
ist
E
X^^^^i^i^^M-we
-
(j
1
32<& 3&B 3ij33 as!aa 3&Q3 a/joa "^san " «am
I
S
'5
5T
i
2
%
^^w.
28.03
D
1
O
s
Jj3
<
g
D
'c
1
IA*^*^
59ffl
I
03 ,
O P?
S 0
1 i
I I
33£»3 31 m
QJ
W
a
1
\l
Jl
SJ
M
c
QJ
|
O
£
IS
S)
3SM
s
o
c
as
|
1
M
C
|
Q
Js
JL,
I
QJ
C
_
, 1
i a.
ro
li.
33$3 MQQ
Figure 3-4. MDGC/MS Analysis of Methylene Chloride Extract from Run 10. The upper trace is
for the single-column, "one-dimensional" analysis. "Two dimensional" resolution of a singlet and
a doublet are shown. (Copyright © 1996; reproduced with permission of UDRI).
3-22
-------
Table 3-14. Compounds identified via MDGC/MS
fluorobiphenyl bromothiophene
bromonaphthalene dibenzofuran
dichlorinated PAH benzopyran-2-one
benzoquinoline tribromophenol
nitrated PAH fluoren-9-one
pentachlorobenzene bromophenoxy benzene
tribromobenzene-diol naphthalene dicarboxylic acid
3.4 - PCDDs/PCDFs and PBDDs/PBDFs
Table 3-15 lists the PCDDs/PCDFs and PBDDs/PBDFs found in the tests. Some congeners were
found in all tests. These values reflect data taken at duct temperatures ranging from approximately
200 to 350 °C, and reflect short residence time in-flight formation of PCDDs/PCDFs and
PBDDs/PBDFs and emissions of those compounds as PICs rather than formation at longer
residence times, such as those found in paniculate control devices. In general, the low temperature
and high Br process conditions tended to yield higher levels of PICs than the baseline and even the
fuel-rich conditions. Of particular interest is the observation of the very high levels of
PCDDs/PCDFs that were found during Run 10, when the Br was at the high feed concentration.
Tripling the concentration of Br in the feed resulted in an order of magnitude increase in
PCDD/PCDF emissions, plus measured quantities of PBDDs/PBDFs were much higher. It may
be that the presence of Br inhibits reactions that reduce the production of PCDDs/PCDFs. It may
also be that Br may enhance some of the reactions that produce PCDDs/PCDFs. Further work is
planned to investigate this phenomenon. It is also of interest that variations between the different
run conditions produced a wide variation in concentrations of PCDDs/PCDFs. These data are
undergoing further analyses to evaluate differences between run conditions. In spite of efforts to
maintain a constant duct temperature, variations did occur, and this may be sufficient to account for
some of the variations. HC1 concentrations in the sampling duct were on the order of 5000 ppmv,
which could provide more than sufficient gas-phase Cl to achieve these levels of PCDD/PCDF
emissions. This is not typical of normal incinerator operation, since typically the HC1 is removed
prior to passing the flue gases through the optimal PCDD/PCDF formation temperature window.
3-23
-------
Table 3-15. Polychlorinated and
Baseline
Monochlorodibenzofuran
Monochlorodibenzodioxin
Dichlorodibenzofuran
Dichlorodibenzodioxin
Trichlorodibenzofuran
Trichlorodibenzodioxin
Tetrachlorodibenzofuran
Tetrachlorodibenzodioxin
Pentachlorodibenzofuran
Pentachlorodibenzodioxin
Hexachlorodibenzofuran
Hexachlorodibenzodioxin
Heptachlorodibenzofuran
Heptachlorodibenzodioxin
Octachlorodibenzofuran
Octachlorodibenzodioxin
Totals
Bromotrichlorodibenzodioxin
Bromotrichlorodibenzofuran
Dibromodichlorodibenzodioxin
Tetrabromodibenzodioxin
Pentabromodibenzodioxin
Bromotrichlorodibenzofuran
Tetrabromodibenzofuran
Pentabromodibenzofuran
Run
5
(ng/m3)
0
0
0
0
28.87
0
28.30
0
54.34
0
39.81
0
47.55
7.74
16.42
56.04
279
0
0
0
0
0
0
0
0
Run
6
(ng/m3)
0
0
0
0
0
0
0
0
8.33
0
0
0
7.04
0
0
0
15
0
0
0
0
0
0
0
0
Polybrominated Dioxins and Furans
| Low SCC Temp |
Run Run
9 10a
(ng/m3) (ng/m3)
693.79
16.89
1145.83
35.53
957.67
53.01
421.75
43.11
358.83
43.69
310.49
78.64
230.10
73.98
535.92
553.98
5553
0
0
0
0
0
9.71
0
0
10944.93
1770.38
16640.27
3671.38
8940.27
4677.70
1332.95
29.78
659.23
373.04
470.22
386.36
206.49
289.18
306.49
96.51
50795
90.52
0
32.45
0
0
295.51
8.49
0
| SCC Fuel Rich |
Run
13
(ng/m3)
67.38
0
52.04
0
48.35
0
39.22
0
57.28
0
177.09
5.63
126.99
29.13
27.96
79.81
711
0
.b
0
0
0
0
0
0
Run
14
(ng/m3)
1.22
1.57
0
0
0
0
0
0
0
0
0
0
10.43
0
14.43
8.00
36
0
0
0
0
0
0
0
0
a - On Run 10, too much ethylene dibromide was inadvertently added
b - Detected, but not quantified
3-24
-------
3.5 - Surrogate Performance Indicators
A surrogate performance indicator is an easily measured compound or group of compounds whose
variance can account for the variance in the measurements of a more difficult-to-measure
compound, such as PCDDs/PCDFs. In light of that, the data from the VOST analyses were
compared to the emissions of total PCDD and total PCDF, both singly and in combination using
the STEPWISE regression in the SAS JMP software package. STEPWISE first looked at all
compounds measured by VOST and determined whether variance in those compounds could
account for any of the variance in PCDDs/PCDFs. Then individual analytes were compared (using
the statistical correlation coefficient, R2) to see if an R2>0.5 was possible by correlating the
concentration of that pollutant vs LOG(PCDD) and LOG(PCDF).
Figure 3-5 shows trichloroethylene vs. total PCDD, Figure 3-6 shows trichloroethylene vs total
PCDF, and Figure 3-7 shows trichloroethylene vs. total PCDD+PCDF. Trichloroethylene was
chosen because it showed the highest correlation coefficient (R2) for any single compound. These
are remarkably good correlations considering that these data points span a wide range of
combustion conditions and temperatures, particularly in the transition duct where the maximum
formation temperature window for PCDDs/PCDFs can be found. The fact that one of the
chlorinated ethenes was found to be the best indicator is also promising. Chlorinated ethenes have
been implicated as some of the primary precursors to ring growth reactions resulting in the
formation of chlorinated benzenes and chlorinated phenols, the suspected precursors to
PCDDs/PCDFs.21
100000-
co
E
~
Q
Q
0
Q_
10000-
1000-
100-
!
10-
1
0 10 20 30 40 50 60 70 80 90 100
Trichloroethylene (|jg/m3)
Figure 3-5. Trichloroethylene vs. Total PCDDs; R2=0.6476 [based on LOG(PCDD)].
3-25
-------
100000-
co
E
~
10000-
1000-
0 10 20 30 40 50 60 70 80 90 100
Trichloroethylene (|jg/m3)
Figure 3-6. Trichloroethylene vs. Total PCDFs; R2=0.6956 [based on LOG(PCDF)].
100000-
10000-
Q 1000^
O
o_
Q
Q
O
Q_
100-
0 10 20 30 40 50 60 70 80 90 100
Trichloroethylene (|jg/m3)
Figure 3-7. Trichloroethylene vs. Total PCDDs+PCDFs; R2=0.6915 [based on LOG(TOTAL)].
3-26
-------
It must be reiterated that these are preliminary findings, based on pilot-scale tests performed on a
single waste stream. However, if it is borne out by further investigation that trichloroethylene is an
appropriate surrogate for PCDDs/PCDFs, there is good potential that this information could be
used for compliance assurance or system optimization, since trichloroethylene is one of the
OLGC's target analytes, and could be easily measured in the stack of an incinerator.
Potential surrogate indicators were further investigated by evaluating linear combinations of
multiple VOST analytes. JMP was used to do a principal component statistical analysis on all
VOST analytes, excluding brominated and fluorinated compounds. Principal component analysis
is a statistical tool that is used to transform data to group interrelated variables. It is not statistically
valid to directly use many VOST targets simultaneously to predict variance in PCDDs/PCDFs
since, with a limited number of measurements such as are present here, you can explicitly predict
virtually all of the variance in PCDDs/PCDFs by using a large enough group of VOST targets.
However, principal component analysis can allow you to reduce the number of predictors by
transforming their axes.The principal components represent variables that take into account the
interrelations between similar VOST targets since, for example, it is not possible to use benzene
and toluene as completely separate predictors, since their concentrations in the stack are related to
each other. This statistical analysis yielded interesting results, indicating that 72% of the
variability in the VOST PICs can be accounted for by the first three principal components, which
are linear combinations of the various VOST analytes. Performing a least squares regression using
the first three principal components vs total PCDD yielded an R2 of 0.8182, and an R2 of 0.8450
when correlated against total PCDF, and an R2 of 0.8487 when correlated against total
PCDD+PCDFs.
Using a principal component analysis of multiple volatile PICs may be a useful method with which
to derive a surrogate indicator of PCDDs/PCDFs that is based on several analytes rather than a
single analyte. It is unknown, however, how site-specific this approach might be. It would be
worthwhile to explore this possibility on other existing incinerator datasets to see if this method
holds promise.
3-27
-------
4.0 CONCLUSIONS
The conclusions are divided into sections relating to the various primary and secondary goals of
this study.
4.1 - Target Analyte List
Pilot-scale incineration tests have been performed under varied combustion conditions feeding a
mixed surrogate waste, resulting in the generation of numerous PICs. While many of these PICs
were identified as target analytes using standardized sampling and analytical methods, the majority
of PICs present in the incineration emissions were not target analytes. Although a substantial
number have been tentatively identified, a considerably larger number have not been identified at
this time. It can be concluded from these experiments that the current sampling and analytical
schemes for characterizing HWC emissions are inadequate and provide an incomplete picture of the
emission profile. This is primarily due to the presence of an extremely complex mixture of organic
compounds in the HWC emission samples. This is particularly evidenced in the semivolatile
organic samples. Nearly 200 chromatographic peaks were resolved through conventional
methodologies, many of which were coeluting peaks. These coeluting peaks could not be
identified due to combined spectra. The complexity of the samples was further illustrated by the
MDGC technique. Heartcuts of single, conventional peaks resulted in the resolution and
identification of 10 times the number of compounds initially evident. As a result, the number of
compounds suspected to be present in incinerator emissions may be an order of magnitude greater
than initially suspected. Other techniques, such as fractionation with HPLC, may provide similar
benefits for identification of coeluting peaks.
A very promising technique for enabling identification of the complex mixtures present in
combustion emissions is multi-dimensional GC/MS. This technique of performing an additional
chromatographic separation on chromatographic peaks that confound mass spectral identification,
enabled significant additional identification of unknowns on the limited sample for which it was
performed. The authors believe that a much more complete listing of PICs could be generated by
performing a careful analysis of complex samples such as these using MDGC/MS. However,
although MDGC/MS may eventually lend itself to routine analyses, in its current incarnation it is
still an experimental technique.
As a result of these experiments, an expanded list of PIC target analytes has been developed. This
list is by no means complete or comprehensive. This list should be viewed in context with this
particular set of experiments; i.e., waste mix. The PICs resulting from other varied waste streams
have not been evaluated.
The PICs identified fall into several chemical classes. A wide variety of chloro, bromo, and mixed
bromochloro alkanes, alkenes, alkynes, aromatics, and polyaromatics were detected. In addition,
nonhalogenated hydrocarbon homologues along with oxygenated, nitrogenated, and sulfonated
organics were detected. MDGC/MS detected chlorinated PAHs. Analytical methods specifically
suited to these chemical classes are needed to enhance PIC characterizations.
For this facility burning this particular waste stream, conventional Ci and C2 hydrocarbons were
present in levels below 1-2 ppm. Since THC analyzer readings were on the same order of
magnitude as the detection levels for Ci and C2 hydrocarbons, no carbon balance was attempted.
In addition, below 10 ppm, THC analyzer readings are not accurate due to biases introduced by the
4-1
-------
presence of common flue gas constituents.22
4.2 - Effect of Presence of Bromine
Brominated Ci and C2 PICs were present at higher-than-expected concentrations than their
chlorinated analogs, in spite of Br's being present at only 10% of the mass concentration of Cl in
the feed. This phenomenon was also observed with aromatic halogenated PICs such as
brominated and chlorinated benzenes. A large number of chlorinated, brominated, and
bromochloro semivolatile organics were also detected. Even though the sampling was performed
upstream of a particulate matter control device, and samples were taken after a fairly short
residence time in the optimal formation window between 600 and 200 °C, chlorinated, brominated,
and bromochloro dioxins and furans were detected, and some congeners of the PBDDs/PBDFs
were detected. During Run 10, with an erroneously high level of Br in the feed, emissions of
PCDDs/PCDFs were increased dramatically, and significant emissions of PBDDs/PBDFs and
bromochloro dioxins and furans were found. It is not known whether the presence of Br enhances
production or inhibits destruction of PCDDs/PCDFs. Additional experiments are needed to
confirm these results.
It is also unknown whether bromination increases or decreases the relative amounts and toxicities
of the PCDD/PCDF, PBDD/PBDF, and PXDD/PXDF PICs. If bromination of PICs is additive,
then brominated compounds (e.g., PBDDs/PBDFs) could add significantly to risk assessment
calculations, especially if emissions of PBDDs/PBDFs are at a similar concentration as
PCDDs/PCDFs. If the process is substitutive, Br could bring into question trial burn and
compliance test PCDD/PCDF results due to bromination of chlorinated PICs resulting in
brominated or bromochloro PICs that aren't considered in risk assessment calculations.
4.3 - Surrogate Performance Indicators
Based on these tests, on this facility, burning this particular waste stream, emissions of
trichloroethylene give a very good correlation with emissions of total PCDD and total PCDF, even
though PCDD/PCDF emissions varied over several orders of magnitude. Trichloroethylene is a
relatively easily measured compound in the stack of incinerators, and because of its importance as a
ring growth precursor, has a scientific basis for its use as a surrogate for PCDDs/PCDFs, as well
as other chlorinated aromatic PICs of interest. It is not known whether trichloroethylene correlates
with PCDDs/PCDFs in practical systems, although the authors will investigate whether this is the
case. Likely, if trichloroethylene is a viable surrogate in full-scale systems, it will correlate with
PCDDs/PCDFs prior to flue gas cleaning equipment, and would need to be coupled with flue gas
cleaning equipment temperatures in order to be a viable surrogate for stack emissions of
PCDDs/PCDFs.
It is possible to account for most or all of the variance in the PCDD/PCDF data by using linear
combinations of several common volatile PICs, using a principal component statistical analysis to
account for the interrelationships between the volatile PICs of interest. The first three principal
components of the VOST analytes, when correlated against PCDDs/PCDFs, were able to generate
R^s in excess of 0.80. It is not known how broadly applicable or facility specific this observation
is.
Measurement of surrogate performance indicators via OLGC appears to have good promise. Not
4-2
-------
only can the OLGC system make stack measurements, but can measure PICs at intermediate
locations within the combustor, to gain insight into PIC formation processes and for system
optimization. The analytes that gave good promise for potential surrogates for PCDDs/PCDFs
were also OLGC targets. Observations made with the OLGC system show formation of
chlorinated aromatics as gases passed through moderate temperature regions.
4.4 - Implications of These Results
The results from these tests have implications regarding incinerator trial burns and compliance
tests. Although it is not within the scope of this report to make recommendations related to EPA
policy, it is within ORD's charter to bring scientific implications of our results to OSW's attention.
This study raises the following questions:
• Can compliance with potential PCDD/PCDF emission limits that have been demonstrated in
a trial burn, using a synthetic POHC feed with no Br in the system, be ensured during
actual operation when Br is present in the feed?
If a facility will eventually burn Br-containing wastes during operation, should Br be added
to the system during trial burns to challenge the system, even though brominated organics,
including PBDDs/PBDFs, are not included in the regulations or the risk assessment
calculations?
• How can PICs such as PBDDs/PBDFs be accounted for if their sampling and analytical
methodologies have not been validated?
Is it possible to use a common volatile PIC, such as trichloroethylene, as a surrogate for
PCDDs/PCDFs and other chlorinated aromatic compounds? If one can be found, what is
an appropriate level to control to?
How facility specific would it be to use linear combinations of multiple volatile PICs as a
surrogate for PCDDs/PCDFs?
4.5 - Recommendations
Much was learned analytically attempting to expand the target analyte list. Foremost is the obvious
conclusion that conventional analytical methodologies and approaches are inadequate to
characterize the inherently complex emissions samples. This is evidenced by the small number of
target analytes observed relative to the large number of compounds present. Part of the problem
lies in the fact that existing methodologies focus on the identification and quantification of
hazardous waste components and not PICs. The greater problem is that, with complex samples,
chromatographic interferences inhibit the ability to identify unknowns as well as confirm target
analytes. Complex samples often result in significant numbers of coeluting peaks. The mass
spectral fragmentation patterns of coeluting peaks are combined and additive, making individual
spectral identifications difficult. This phenomenon would exhibit itself in the form of large
numbers of tentatively identified compounds with poor identification probabilities from the mass
spectral search. Fortunately, techniques were identified and demonstrated that were capable of
deconvoluting the complex samples. The authors strongly believe that improved analytical
methodologies emphasizing identification and quantification of unknown compounds would
4-3
-------
provide the greatest opportunity to reduce uncertainty in risk assessment calculations with minimal
expenditure.
Additional testing is recommended that incorporate these techniques. This additional testing should
use as a foundation, EPA's Total Organics Approach (TOA). Particular emphasis should be placed
on characterization of the semivolatile and nonvolatile fractions. This would equate to total
chromatographable organic (TCO) and gravimetric organic (GRAY) fractions of the TO A. Each
sample fraction should be segregated or fractionated, based on polar characteristics, to provide a
first step towards deconvoluting the sample. This can be quantitatively accomplished using High
Performance Liquid Chromatography (HPLC). Each segregated fraction should then be re-
subjected to the TCO and GRAY analyses to ensure mass recovery. Then each sample fraction
should be reanalyzed by GC/MS as well as MDGC/MS. This will not only improve compound
identification and quantitation, but also demonstrate this particular approach as a potential method
for characterizing incinerator emissions.
This testing should also include separate efforts to identify the components present in the GRAY
fraction. Theoretically, the GRAY fraction includes primarily nonvolatile organics possessing high
molecular weight compounds. It is possible, even probable, that a considerable portion of these
compounds are not amenable to conventional GC analyses. However, the ability to characterize
this fraction has met with mixed results. This fraction typically remains uncharacterized, with only
a small percentage of the mass being identified.
It is the authors' strong contention that the GRAY fraction may consist of organic and/or inorganic
mass not directly attributable to organic incinerator emissions. This artifact may be comprised of
inorganic salts, super-fine particulate, fractured XAD-2 resin, or some other unknown. This
artifact may account for the inability to identify a significant percentage of the GRAY fraction.
Experiments can be designed to further determine the representativeness of the GRAY fraction.
Based on these results, more efficient analytical approaches can be devised to characterize the
GRAY fraction, thereby improving the potential for identifying a larger percentage of the GRAY
fraction.
Finally, it may be possible to develop a multi-tiered approach to measuring PICs from incineration
systems. Some incineration systems may exhibit a relatively small number of identifiable PICs,
whereas others may have an exceedingly complex mixture in the stack. This multi-tiered approach
could be performed by commercial analytical laboratories on a routine basis. The multi-tiered
approach would consist of the following:
Tier 1: First Pass Analysis
The first pass analysis would focus on using existing analytical methodologies that
focus more on potential PICs. The MM5 samples would be extracted and analyzed
conventionally using a Method 8270C analysis, directed at the Method 8270C
targets. The existing target list should be expanded to include common PICs that
are amenable to GC/MS analysis. Aliquots from these same extracts would be
subjected to further analyte-specific analyses for chlorobenzenes and chlorophenols
(Method 8041), PAHs (CARB Method 429), and nitroaromatics and cyclic ketones
4-4
-------
(Method 8091). These are more analyte-specific analyses and offer greater
sensitivity, particularly through the use of selective ion monitoring techniques.
High resolution mass spectrometry (HRMS) may also be used to improve
sensitivity, if needed.
Greater emphasis should be placed on the identification and accurate quantitation of
unknowns. Guidelines should be developed that standardize this approach. These
guidelines should include spectral library searching and spectral interpretation
requirements, confirmation of unknowns with known standards where possible,
and other criteria that add to the quality of the identification (e.g., retention time,
boiling point). In addition, tentatively identified unknowns should be quantified
using a response factor of a compound similar to the characteristics of the unknown
rather than an unrelated compound closest in retention time.
The Method 23 samples would be analyzed for PCDDs/PCDFs and PCBs using
HRMS. The PCB analysis should include both totals and the co-planar isomer
specific analyses. Two PCB Methods exist which can accomplish this method:
CARB 428 and Draft Method 1668. The Method 23 target analyte list should be
expanded to include the mono-, di-, and tri- substituted dioxin and furan congeners
as well as the tetra- through octa- as are normally measured. Limited laboratory and
field data suggest that the lower chlorinated congeners may be suitable surrogates
for the higher chlorinated congeners, and measurement of the lower chlorinated
congeners with a CEM may be practical in the near future. It is necessary to
develop a database of the lower substituted congeners to develop correlations for
different facility and feed types.
Based on the results from the Tier 1 analysis, it will be decided whether the sample
was sufficiently complex to merit further investigation (e.g., number of peaks
identified relative to total number of peaks). Again, complex samples would result
in significant numbers of coeluting peaks, making spectral identifications difficult.
This would ultimately result in a large number of unidentified compounds. If the
samples analyzed using the Tier 1 approach indicate that a significant number of
coeluting peaks exist, then Tier 2 should be used.
Tier 2: Sample Deconvolution
For Tier 2, the MM5 extracts would be run through an HPLC fractionation system.
A solvent gradient would be used to partition the material eluting off an HPLC
column according to elution time. Separating the MM5 extracts into multiple
fractions of varying polarity, then running those fractions back through a GC/MS
analysis, dramatically reduces the problems of coeluting peaks. This reduction is
due to the fact that GC and HPLC use different techniques to differentiate
compounds: GC separates primarily based on compound boiling points; whereas,
HPLC separates primarily based on compound polarity.
The fractionated samples could also be run on a GC with atomic emission detection
(AED). This detector is element specific and would aid in the interpretation of mass
spectral data by confirming the presence of elements such as halogens, oxygen,
4-5
-------
nitrogen, and sulfur.
Finally, the fractionated extracts could be analyzed by multidimensional gas
chromatography (MDGC). The power of this technique has been demonstrated
through this study.
4-6
-------
5.0 REFERENCES
1 40 CFR Part 261 Identification and Listing of Hazardous Waste - Appendix VIII-Hazardous
Constituents. July 1, 1990, Government Printing Office, Washington, DC.
2 Clean Air Act Amendments of 1990, P.L. 101-549, U.S. Congress, Washington, DC,
November 15, 1990.
3 Resource Conservation and Recovery Act (RCRA), as amended by the Hazardous and Solid
Waste Amendments of 1984, 40 CFR.
4 Hartzell, G.E., "Overview of Combustion Toxicology," Toxicology, 115: 1-3, 7-23,
December 1996.
5 Dellinger, B., P.H. Taylor, and D.A. Tirey, "Minimization and Control of Hazardous
Combustion By-Products," EPA/600/S2-90/039 (NTIS PB90-259854), Risk Reduction
Engineering Laboratory, Cincinnati, OH, August 1990.
6 Rigo, G.H., AJ. Chandler, and W.S. Lanier, "The Relationship Between Chlorine in Waste
Streams and Dioxin Emissions from Waste Combustor Stacks," ASME Research Report
CRTD-Vol 36, 1996.
7 Okey, A.B., D. Riddick, and P. Harper, "The Ah receptor: Mediator of the Toxicity of 2,3,7,8
TCDD and Related Compounds," Toxicology Letters, 70: 1-22, 1994.
8 Ryan, J., P. Lemieux, C. Lutes, and D. Tabor, "Development of PIC Target Analyte List for
Hazardous Waste Incineration Processes," Paper presented at the International Incineration
Conference, Savannah, GA, May 6-10, 1996.
9 EPA Test Method 0040 "Sampling of Principal Organic Hazardous Constituents from
Combustion Sources Using Tedlar Bags" in Test Methods for Evaluating Solid Waste, Volume
II, SW-846 (NTIS PB88-239223). Environmental Protection Agency, Office of Solid Waste,
Washington, DC. (August 1994).
10 EPA Test Method 0030 "Volatile Organic Sampling Train" in Test Methods for Evaluating
Solid Waste, Volume II, SW-846 (NTIS PB88-239223). Environmental Protection Agency,
Office of Solid Waste, Washington, DC. (September 1986).
11 EPA Test Method 0010 "Modified Method 5 Sampling Train" in Test Methods for Evaluating
Solid Waste, Volume II, SW-846 (NTIS PB88-239223). Environmental Protection Agency,
Office of Solid Waste, Washington, DC. (September 1986).
12 EPA Test Method 23 "Determination of Polychlorinated Dibenzo-p-dioxins and
Poly chlorinated Dibenzofurans from Stationary Sources" in Code of Federal Regulations, Title
40, Part 60, Appendix A, U.S. Government Printing Office, Washington, DC. (July 1991).
13 Richards, M.K., L.R. Waterland, and E. Whitworth, "Innovative Continuous Emission
Monitors: Results of the EPA/DOE Demonstration Test Program," Paper presented at the
International Incineration Conference, Savannah, GA, May 1996.
5-1
-------
14 Ryan, J.V., P.M. Lemieux, W.T. Preston, and L.R. Waterland, "Field Demonstration of a
Prototype On-Line Gas Chromatograph for Near-Real-Time Measurement of Trace Volatile
Organic Products of Incomplete Combustion from Incinerators," AWMA J., in press.
15 California Air Resources Board Method 428 "Determination of Poly chlorinated Dibenzo-p-
dioxin (PCDD), Polychlorinated Dibenzofuran (PCDF), and Polychlorinated Biphenyl
Emissions from Stationary Sources" in Methods for Determining Emissions of Toxic Air
Contaminants from Stationary Sources, Volume III, State of California Air Resources Board,
December 1991.
16 EPA Test Method 5040 "Protocol for Analysis of Sorbent Cartridges from Volatile Organic
Sampling Train" in Test Methods for Evaluating Solid Waste, Volume II, SW-846 (NTIS PB-
239223). Environmental Protection Agency, Office of Solid Waste, Washington, DC,
September 1986.
17 EPA Test Method 8240 "Gas Chromatography/Mass Spectrometry for Volatile Organics" in
Test Methods for Evaluating Solid Waste, Volume II, SW-846 (NTIS PB-239223).
Environmental Protection Agency, Office of Solid Waste, Washington, DC, September 1986.
18 Rubey, W.A., B. Dellinger, and R.C. Striebich, "Chemical Analysis of Combustion Samples
Using Multi-Dimensional Gas Chromatography," Paper presented at the International
Conference on Incineration and Thermal Treatment Technologies, May 12-16, 1997, Oakland,
CA.
19 Anderson, S., J. Garver, W. Rubey, R. Striebich, and R. Grinstead, "The Separation and
Identification of Trace Heteroatomic Species in Jet Fuel by Sample Enrichment and
Multidimensional Gas Chromatography with Mass Selective Detection (MDGC-MSD),"
Proceedings of the 17th ISCCE, Wintergreen, VA, May 1995.
20 Tsang, W., "Mechanisms for the Formation and Destruction of Chlorinated Organic Products
of Incomplete Combustion," Combustion Science and Technology, 74:99-116 (1990).
21 Tirey, D.A., P.H. Taylor, J. Kasner, and B. Dellinger, "Gas Phase Formation of Chlorinated
Aromatic Compounds from the Pyrolysis of Tetrachloroethylene," Comb. Sci. Tech.., 74,
137-157, 1990.
22 Ryan, J.V., P.M. Lemieux, and P.W. Groff, "Evaluation of the Behavior of Flame lonizati on
Detection Total Hydrocarbon Continuous Emission Monitors at Low Concentrations," Paper
presented at the International Conference on Incineration and Thermal Treatment Technologies,
May 12-16, 1997, San Francisco, CA.
5-2
-------
APPENDIX A
QUALITY CONTROL EVALUATION REPORT
This project was conducted under the guidance of an EPA-approved QA Test Plan (APPCD
Category III). The Test Plan describes the intended experimental approach and procedures.
The Test Plan also presents Data Quality Objectives (DQOs) for this study: to collect data of
sufficient quality to develop a qualitative list of organic compounds present in HWC emissions.
This list is not meant to be representative of all incineration configurations, conditions, or waste
mixtures. Data Quality Indicator (DQI) goals were established to meet DQOs.
Table A-l presents the DQI summaries for accuracy, precision, and completeness achieved
during testing along with the planned DQI goals for each measurement or analysis performed. In
general, the intended RKIS operational DQI goals were achieved. However, DQI goals for
quantitative organic measurements generally were either not achieved or could not be assessed
from the available data. Quantitative DQI goals were not met primarily due to poor surrogate
and/or internal standard recoveries. As a result, the analytical data should be viewed as
semiquantitative at best. While it is not appropriate to report organic emissions concentrations as
absolute, the data are of sufficient quality to make rough order of magnitude quantitative
comparisons between test condition data sets. It should be stressed, though, that qualitative
identification was the primary goal of this project, not quantitative. The recovery problems have
negligible impact on the qualitative identification of the PICs. As a result, the data are of
sufficient quality to meet project objectives to develop a qualitative list of organic compounds
present in HWC emissions.
Case narratives for specific analytical activities are included in the following subsections.
A.I - Continuous Measurement Results
The THC analyzers failed after Run 11 due to the high HC1 content of the flue gas. Their data
were not available for Runs 12 through 16, resulting in a completeness of 69.4%, which was
slightly below the desired 70% completeness. In addition, the Bodenseewerk HC1 CEM was not
available during Runs 13 through 16 due to its redeployment on other facilities, resulting in a
75% completeness of data. Other CEMs passed QC criteria.
The thermocouple at the kiln exit failed after Run 8 and was not available during Runs 9 through
12, resulting in a completeness of 75% for that thermocouple. A replacement thermocouple was
installed at that point, and kiln exit temperatures were measured during subsequent tests. All
other thermocouples operated normally within QC guidelines.
A-l
-------
There was an anomaly in the bromine feed concentration in the batch of feed used in Run 10.
All of the other runs were prepared at 449.8 g Br per batch, while run 10 was at 1589.8 g Br per
batch.
Table A-l. Data Quality Indicator Summary for Critical Measurements
Measurement Accuracy Accuracy Precision Precision Completeness Completeness
Goal Achieved Goal Achieved Goal Achieved
02
CO2
CO
THC
NO
Temperature
HC1
VOCs (VOST)
VOCs(TedlarBag)
SVOCs
PCDDs/PCDFs
VOCs (OLGC)
±5
±5
±5
±5
±5
±2
±5
50-150
50-150
18-120
40-120
NA
pass
pass
pass
passa
pass
NA
NA
fail
fail
fail
*a
NA
5
5
5
5
5
±2
5
NA
30
30
30
NA
pass
pass
pass
passa
pass
±2
pass
NA
NA
NA
NA
NA
70
70
70
70
70
100
70
75
75
75
70
NA
100
100
100
68.75
100
100 (75)
75
100
100
100
100
NA
a - see additional information in text.
A.2 - Volatile Organic Compound Analyses
A.2.1- VOST Samples
The 30 day holding times to analysis for these samples were generally adhered to.
The surrogate recoveries for the VOST compounds were mostly below the pass/fail criterion of
50-150%. The insufficient recoveries do not, however, impact the qualitative analysis of the
data. A possible reason for the failure of the recoveries of internal standards is that the extremely
high HC1 content (several thousand ppmv) of the flue gas may have degraded the Tenax' ability
to adsorb VOCs. The VOST method is intended for application downstream of particulate and
acid gas control systems and not in the highly corrosive environment during these tests.
The first internal standard (bromochloromethane) was identified as a PIC. Because of this, the
A-2
-------
second internal standard was used to quantify the targets that are normally referenced to the first
internal standard.
The blanks showed a general trend of having common ketones, solvents, and chloromethane
present (as is common for VOST samples) as contaminants. There were also a few instances of a
minor carryover from the daily standard. But with target hits as high as 5000 ng per tube of
benzene and many other compounds being near 1000 ng per tube, the contaminant levels were
insignificant relative to sample levels.
Many of the VOST samples exhibited concentrations higher than the calibration range. The
concentrations of these compounds will tend to be over-estimated due to non-linear responses of
the mass spectrometer at regions above the calibration range. The nature of the VOST
sampling/analysis does not allow reanalysis or dilution to bring these compounds into the
calibration range. Data exceeding calibration levels are flagged as estimates. Given the
semiquantitative nature of reported results, these estimates do not pose a problem.
A.2.2 - Tedlar Bag Samples
Hold times did not exceed 1 day, which is acceptable.
The blank samples were generally clean with only a few compounds reported above the practical
quantitation limit (PQL). Few PICs were found in the blank samples.
There were inconsistencies in the reported recoveries of surrogate standards, which make it
difficult to assess the quality of the quantitations. Based on careful examination of available
data, in both hard copy and disk form, it is believed that the qualitative results are correct, but
that the quantitative results may be in error by a factor of 2.5. Since these data are compared
only to other test conditions, relative differences are not affected.
A.3 - Semivolatile Organic Compound Analyses
Semivolatile analysis by SW-846 Method 8270 was completed for eight samples. Filter and
XAD-2 fractions were extracted separately. In general, filter extract surrogate recoveries were
low, with many being just barely acceptable. The XAD samples, generally showed acceptable
recovery. In all analyses, the surrogate recovery is worse for the earlier eluting (lower boiling
point) compounds. A contributor to poor recovery was that some sample extracts were
concentrated on a rotary evaporator (Roto-Vap) instead of the Kaderna-Danish concentrating
apparatus which is specified in the method. This technique is less efficient and would result in
greater azeotroping and, therefore, the preferential loss of the more volatile surrogate standards.
A-3
-------
After these samples exhibited the poor recoveries, the laboratory stopped using the Roto-Vap
apparatus for semivolatile samples and resumed using the Kaderna-Danish apparatus. Volatile
surrogate standard recoveries improved somewhat. Matrix effects, due to the extremely high
HC1 content of the sample collected, also likely impacted surrogate recoveries. Fortunately, the
poor volatile surrogate standard recoveries were associated primarily with the filter extracts.
During sampling, the more volatile species would tend to be collected on the XAD-2 rather than
the filter. While the poor surrogate recovery problem impacts quantitative capabilities,
qualitative information should not be compromised. This tenet is supported by the independent
identification of overlapping PICs in both the VOST and MM5 samples.
Due to the high concentrations of nontarget analytes in the initial MM5 analyses, many reactive
compounds responded poorly. Initial MM5 extracts, once concentrated to 1 mL, were dark and
non-transparent. It is likely that the cumulative effect of injection of these corrosive, complex
samples caused active sites to develop in the injection port and entrance of the column causing
poorer responses for these more reactive compounds. Frequent injector and guard column
maintenance reduced this problem. To verify acceptable MS and chromatographic performance,
the decafluorotriphenylphosphine (DFTPP) tuning criterion was met prior to sample analyses
each day, and the degradation products of dichlorodiphenyltrichloroethane (DDT) (compound in
the DFTPP tuning solution) demonstrated less than 6% degradation prior to sample analyses for
each day. DDT is a typical example of a labile compound used by the method to determine the
condition of the chromatographic system. If degradation of DDT was greater than 20%, GC
maintenance was performed. In an effort to improve chromatographic separations, GC
conditions were modified to reduce the oven temperature ramping rate and to optimize column
carrier flowrate from levels used during the initial analyses.
The five-point calibration ranged from 10 to 120 ng injected on column (except for the acid
surrogates which ranged from 20 to 240 ng). Poorer responding compounds' PQLs — defined
here as the lowest point on the calibration curve - were raised to 30 and sometimes 60 ng to
obtain good response correlation throughout the calibration range.
All continuing calibration check compounds (CCC) and System Performance Check Compounds
(SPCCs) had less than 30 % relative standard deviation and greater than 0.05 relative response,
respectively (prior to daily sample analyses), which satisfies Method 8270 cutoff values.
A.4 - PCDD/PCDF and PBDD/PBDF Analyses
Both chlorinated and brominated DD/DF analyses were performed. As described earlier, the
PCDD/PCDF analyses were performed following standardized procedures. A significant portion
of the internal standard surrogate recovery results were outside of the method criteria (40-120%)
A-4
-------
with many recoveries in the 20-30% range. Recoveries were highly variable but didn't seem to
have a pattern. Run ll's filter fraction, Run 12's XAD fraction, and Run 13's filter fraction
exhibited below 1% recoveries of the internal standards. Run 14's filter was lost and no extract
was produced. These results, although not quantifiable within method criteria, are still usable to
evaluate trends between test conditions.
Formalized methods for identifying and quantifying brominated DD/DF do not exist. As a
result, the analyses performed were essentially a screening technique attempting to verify the
presence or absence of select PBDD/PBDF congeners for which limited standards are available.
For the brominated compounds, the ion ratio was the only definitive criterion available to
confirm presence: no window defining mixes are available. The retention time was evaluated
compared to the 1-^C labeled TBDD/F standards. We used a general rule that a compound with a
bromo substitution would correspond roughly to the retention time area of the same compound
with a dichloro substitution. The fully brominated penta, hexa, and hepta diphenyl ethers were
monitored for, but none were detected. This indicates that there was no interference between the
fully brominated furans and these compounds. This approach is sufficient to screen for the
presence of PBDD/PBDF PICs.
A.5 - Online GC Samples
On-line GC measurements were performed primarily to evaluate performance as a potential
VOC monitor. No DQI goals were established. Each day a system bias check was performed to
verify that recoveries of a 200 ppb sample were within the range of 50-150% by injecting a VOC
standard mix into the probe at the stack and comparing the measured concentrations to the same
mix injected directly into the sparge vessel of the OLGC. The system passed the system bias
check each day. In addition, system blanks were performed to verify that no targets were present
in the system prior to each run day, and a calibration was performed each day to verify retention
times and concentrations.
A-5
------- |