United States
Environmental Protection
Agency
Control Technology
Center
Research Triangle Park NC 27711
EPA-600/R-93-044
March 1993
CHARACTERIZATION OF EMISSIONS
FROM THE SIMULATED OPEN-BURNING
OF NON-METALLIC AUTOMOBILE SHREDDER RESIDUE
-------
EPA REVIEW NOTICE
This report has been reviewed by the Control Technology Center (CTC) established by the Office
of Research and Development (ORD) and the Office of Air Quality Planning and Standards (OAQPS) of
the U.S. Environmental Protection Agency (EPA), and has been approved for publication. Approval does
not signify that the comments necessarily reflect the views and policies of the U.S. EPA nor does mention
of trade names or commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-6GQ/R-93-044
March 1993
FINAL report
CHARACTERIZATION OF EMISSIONS
FROM THE SIMULATED OPEN-BURNING
OF NON-METALLIC AUTOMOBILE SHREDDER RESIDUE
Prepared by:
Jeffrey V. Ryan and Christopher C. Lutes
Aeurex Environmental Corporation
4915 Prospectus Drive
P.O. Box 13109
Research Triangle Park, NC 27709
EPA Contract No. 68-DO-0141
Technical Directive Nos. 91-030/92-055
EPA Project Officer: Paul M. Lemieux
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Prepared for:
U.S. Environmental Protection Agency
Office of Research and Development
Washington, D.C. 20460
-------
ABSTRACT
The reclamation process for ferrous and non-ferrous metals from scrap automobiles generates a
non-metallic waste product called "fluff," consisting of a combination of plastics, rubber, glass, wood
products, and electrical wiring. The waste product is often stockpiled or landfilled. A number of these
stockpiles have caught fire, resulting in the emission of numerous air pollutants. To gain insight into the
types and quantities of these air pollutants, a study was conducted in which the open combustion of fluff
was simulated and the resulting emissions collected and characterized. Samples were collected and
analyzed for volatile and semivolatile organics, particulate, and metal aerosols. Typical combustion
process gases, carbon dioxide (CO,), carbon monoxide (CO), nitric oxide (NO), oxygen (02), and
unbumed hydrocarbons (THC) were monitored continuously. The respective samples were analyzed using
GC/MS, GC/FID, gravimetric, and atomic emission methodologies to identify and quantify the types of
compounds present in the open combustion process emissions. The resulting mass/volume concentrations
were related to the measured net mass of material consumed through combustion and known dilution air
volume to derive an estimate of overall emissions. Volatile and semivolatile organics characterized
included mono- and polyaromatic hydrocarbons, substituted alkanes and alkenes, aldehydes, nitriles,
phenols, chlorinated aromatics, heterocycles, and polychlorinated dibenzodioxins and furans. Of the 11
metal aerosols characterized, cadmium, copper, lead, and zinc were found in significant quantities. The
emission characterizations performed indicated that substantial quantities of air pollutants are emitted. For
the organic pollutants alone, the emission of more than 200 g/kg of fluff combusted was observed.
PREFACE
The CTC was established by EPA's Office of Research and Development (ORD) and Office of
Air Quality Planning and Standards (OAQPS) to provide technical assistance to state and local air
pollution control agencies. Three levels of assistance can be accessed through the CTC. First, a CTC
HOTLINE has been established to provide telephone assistance on matters relating to air pollution control
technology. Second, more in-depth engineering assistance can be provided when appropriate. Third, the
CTC can provide technical guidance through publication of technical guidance documents, development
of personal computer software, and presentation of workshops on control technology matters.
The technical guidance projects, such as this one, focus on topics of national or regional interest
that are identified through contact with state and local agencies.
ii
-------
TABLE OF CONTENTS
Section Page
ABSTRACT ii
PREFACE ii
LIST OF FIGURES iv
LIST OF TABLES v
1.0 INTRODUCTION . 1
2.0 EXPERIMENTAL APPROACH 3
2.1 General Project Description . 3
2.2 Experimental Apparatus 4
2.2.1 Bum Hut 4
2.2.2 Sample Shed 4
2.2.3 Hazardous Air Pollutants Mobile Laboratory (HAPML) 8
2.3 Experimental Methods and Procedures 8
2.3.1 Combustion of Fluff 8
2.3.2 CEMs 10
2.3.3 Volatile Organics 10
2.3.4 Semivolatile Organics and Particulate Matter 13
2.3.5 Metal Aerosols 17
2.3.6 PM10 Sampling 18
3.0 DATA, RESULTS AND DISCUSSION 20
3.1 General 20
3.2 Combustion Characteristics 20
3.3 Volatile Organic Emissions Data 28
3.4 Semivolatile Organics Data 37
3.5 PAH Analyses . 43
3.6 PCDD/PCDF Data 46
3.7 Metal Aerosols Data 50
3.8 Particulate Data . 50
3.9 Emission Data Summary 53
4.0 SUMMARY AND CONCLUSIONS 59
5.0 REFERENCES 62
APPENDIX A: QUALITY CONTROL EVALUATION REPORT 65
iii
-------
LIST OF FIGURES
Figure Page
2-1. Diagram of Burn Hut 5
2-2. Aerial View of PIC Facility .... 6
2-3. Sampling Systems in Sample Shed 7
2-4. PM10 Medium Volume Particulate Sampler 19
3-1. Burn Rates of Fluff 22
3-2. Temperatures Over Burn Pit 23
3-3. CO Concentrations in Three Fluff Tests . 24
3-4. C02 Concentrations in Three Fluff Teste 25
3-5. THC Concentrations in Three Fluff Tests 26
3-6. NO Concentrations in Three Fluff Teste . 27
3-7. Total Volatiles Vs. Burn Rate 29
3-8. Estimated Emissions for Selected CAAA HAPs 32
3-9. PAHs in the Vapor Phase 44
3-10. PAHs in the Particulate Phase 45
3-11. Estimated Emissions for Vapor Phase PCDDs/PCDFs 47
3-12. Estimated Emissions for Particulate-bound PCDDs/PCDFs . 48
3-13. Total PCDD/PCDF Estimated Emissions . .......... 49
3-14. Estimated Emissions for Selected Metals 51
3-15. Distribution of Organics by Sampling Method 54
3-16. Distribution of Organics by Boiling Point 55
3-17. Particle Mass Distribution 57
iv
-------
LIST OF TABLES
Table Page
3-1. Combustion Conditions in Three Fluff Tests 21
3-2. VOC Mass and Burn Rate Data 30
3-3. Estimated Emissions for Selected Volatile Organies 33
3-4. Volatile Organic Compounds Identified 35
3-5. Estimated Organic Emissions by Sample Fraction (g/kg) 38
3-6. Estimated Emissions for Organic Compounds Collected on XAD Resin and
Particulate Filters 39
3-7. Metals Emission Data 52
3-8. Summary of Estimated Particulate Emissions 52
3-9. Mass Balance Summation (g/kg) 56
A-l. Data Quality Summary 66
A-2. Results of PAH Performance Evaluation Audit 71
A-3. Recoveries from Spiked XAD-2 Samples 72
A-4. Recoveries of Isotopically Labeled PCDD/PCDF Internal Standards 73
v
-------
SECTION 1
INTRODUCTION
The reclamation process for retrieving recyclable ferrous and non-ferrous metals from scrap
automobiles generates a non-metallic waste product called "fluff." The fluff waste stream from
automobile reclamation facilities also often includes the non-metallic residue of major household
appliances which are known as "white goods*.1 The major constituents of fluff are plastics such as
polyethylene (PE), polypropylene (PP), acrylonitrile-butadiene-styrene (ABS), polyurcthane foam
(PUF), polyvinylchloride (PVC), rubber, glass, wood products, cloth, paper, dirt, and electrical
wiring.2,3 The actual composition of fluff depends on the type of separating technique used during the
reclamation process. In one such process, the fluff is separated from the desired, recyclable material
using a series of air blowers and yields the final waste product described above. Another common
reclamation technique uses water to separate floating, undesirable products from the denser material.4
With this process, the denser materials such as glass and electrical wiring are less likely to be present
in the fluff fraction.
In 1974, Mahoney et al., reported that a large percentage of automotive plastics are processed
at approximately 100 junk automobile shredders in this country. These reclamation facilities are
capable of processing 50,000-200,000 automobiles per year.5 Valdez et al., state that more than half
of the automobiles scrapped annually in the United States are now processed by shredding.2 Since
1960, the amount of plastic contained in automobiles has increased drastically, and this trend is
projected to continue. The average automobile contains more than 90.7 kg (200 lb) of plastics.6
1
-------
Conservatively estimating that each of the roughly 100 automobile shredding facilities handle 100,000
automobiles annually, these facilities can generate 9.1 x 10s kg (2 x 109 lb) of plastic waste per year.
This estimate of the number of autos shredded per year agrees closely with an industry estimate of 9
million autos processed per year.1
The resulting automobile fluff is discarded at landfills or, more commonly, stockpiled on site.
At several automobile reclamation facilities, these stockpiles have, for various reasons, caught fire.
One such stockpile fire, in Montvale, VA, burned for 38 days emitting unknown quantities of
potentially harmful air pollutants. It was estimated that 13,000-16,000 bales of fluff, weighing
1,360 kg (3,000 lbs) each, were burned in the fire.7 Over the course of the fire, several attempts were
made to extinguish the fire as well as to accelerate combustion.8 The Commonwealth of Virginia's
Department of Air Pollution Control contacted the Control Technology Center (CTC) of the U.S.
Environmental Protection Agency (EPA) requesting emissions data on combustion of this material.
Unfortunately, data pertaining to the open burning of fluff or any similar material were extremely
limited. As a result, the CTC felt that a study characterizing the emissions resulting from the open
combustion of fluff was warranted.
Under contract to EPA's Air and Energy Engineering Research Laboratory (AEERL), Acurex
Environmental Corporation performed a study to identify and quantify organic and inorganic emission
products produced during the simulated open combustion of fluff. Specifically, this study was
designed to determine emissions factors, accurate to one order of magnitude, for volatile and
semivolatile organics, particulate matter, and selected metal aerosols identified in combustion
emissions. Emphasis was placed on gaining a broad overview of the diversity of pollutants produced.
2
-------
SECTION 2
EXPERIMENTAL APPROACH
2.1 GENERAL PROJECT DESCRIPTION
The project consisted of a replicate study to collect and qualitatively and quantitatively
characterize organic and inorganic emissions resulting from the simulated open combustion of actual
automobile fluff. Small quantities (9.1-11.4 kg [20-25 lb]) of fluff were combusted in test facilities
specifically designed to simulate open-combustion conditions. The tests were conducted in triplicate to
allow for the heterogeneous composition of the fluff and to assess reproducibility. A portion of the
combustion effluent was diverted to an adjacent sampling facility via an induced draft duct Organics
were collected using the volatile organ ics sampling train (VOST) and a semivalatile
organics/particulale collection system using XAD-2 and particulate filters. Metal aerosols were
collected on particulate filters. The organic constituents were analyzed both qualitatively and
quantitatively using gas ehromatograph/mass spectrometer (GC/MS), gas chromatograph/flame
ionization detector (GC/F1D), and gravimetric methodologies. The metal aerosols were characterized
using an inductively coupled argon plasma (ICAP) method. Measured concentrations were related to
dilution air volumes and measured net mass of fluff combusted to derive emission rate.
3
-------
2.2 EXPERIMENTAL APPARATUS
2.2.1 Burn Hut
The bum hut (Figure 2-1) is an outbuilding with a 2.7 x 3.4 m (8.9 x 11.1 ft) floor area and a
sloping roof with a minimum height of 1.9 m (6.3 ft) and a maximum height of 2.2 m (7.3 ft),
modified for small-scale, open-combustion simulation experiments. The building has been fitted with
a conditioned air handling system delivering nominally 41.3 m3/min (1,460 ft3/min). At this flow rate,
the effective air exchange rate of the burn hut is 2.17 exchanges/min. The test material was
combusted in a 61 x 56 cm (24-in x 22-in) diameter steel, cylindrical vessel on a platform scale to
continuously monitor weight differential. A pyramidical deflector shield is located 0.9-1.2 m (3-4 ft)
over the pit to deflect flames, protect the ceiling, and enhance ambient mixing. The sample transport
duct, 17-cm (6.6-in) OD stove pipe, is located directly over the deflector shield. This duct transports a
representative portion of the burn hut environment to the sampling shed located adjacent to the burn
hut (Figure 2-2). To minimize heat loss and condensation of organics, the duct is insulated outside the
burn hut. The inner walls and ceiling of the bum hut are covered with 1.6-ram (1/16-in) aluminum
sheeting to provide an inert surface within the test facility.
2.2.2 Sample Shed
The sample shed (Figure 2-2, 2-3) contains the majority of the associated sampling equipment:
the volatile organic sampling train (VOST) system, the semivolatile organies/particulate sample
collection systems, and the particulate removal system for the continuous emission monitors (CEMs),
The digital readout for the platform scale is remotely operated from the sample shed.
All samples were extracted from a sampling manifold within the duct. The manifold consists
of 9,5-mm (3/8-in) OD stainless steel probes positioned in the sample transport duct so that the probe
orifice faces the direction of sample flow and that all samples are collected at the same axial and
radial location (see Figure 2-3). The sample stream was pulled from the burn hut into the sample shed
under vacuum by an induced draft (ID) fan located downstream of the sample manifold.
4
-------
Sample Duct
Fluff Combustion
Container
Air Met
F —-1 .T\l Su *—
Weighing Platfonn
I
Air Inlet
Figure 2-1. Diagram of bum hut
5
-------
Insulated
Sample
Duel
Sample Shed
Sampling Control Center
Particle Sampling
Volatile and Semivolatile
Organic Sampling
Airborne Metals Sampling
Bum Hut
Heated Sample Line
A
OEMS
O2 J 'THC !
, _ _ _<
• C02 1 I NO !
CO
Hazardous Air
Pollutants Mobile
Laboratory
•Data
: System
It
7
Figure 2-2. Aerial view of the PIC facility.
-------
~
Duct Cross Section
Sampling Shed
Heated TFE Tubing to HAP
Heated
Spun Glass
ID Fan
From Burn Hut
Condensers
VOST
System
142 MM A
TFE Coated^
r0%Pump
Dry Gas
Meter
Water
Cooled
Condenser
Tenax
Traps
J Vacuum
Dry Gas Pump
Meter
Vacuum
Pump Dry Gas
Meter
XAD-2
Canister
Figure 2-3. Sampling systems in sample shed.
7
-------
2.2.3 Hazardous Air Pollutants Mobile Laboratory (HAPML)
The HAPML (Figure 2-2) was used for the continuous monitoring of the fixed combustion
gases. A heated (121 °C [250 °F]), particulate-free gaseous sample was extracted from the sample
manifold and routed to individual analyzers for continuous measurement. A portion of the heated
sample was routed to the total hydrocarbon (THC) analyzer. The remaining portion of the sample
stream was further conditioned for moisture removal by a refrigeration condenser and silica gel before
being routed to the oxygen (O^), carbon dioxide (CO-,), and carbon monoxide (CO) analyzers. The
gas stream for nitric oxide (NO) was obtained from a location between the refrigeration condenser and
desiccant The analog output of the individual analyzers was recorded using a computerized data
acquisition system which recorded all readings at 30-s intervals.
2.3 EXPERIMENTAL METHODS AND PROCEDURES
2.3.1 Combustion of Fluff
Actual fluff obtained from an automobile reclamation facility was used for this study.
According to accompanying sample documentation, the fluff contained the following:
* A major portion of shredded carpet, upholstery fragments, and foam fragments from seat cushions
* A moderate portion of vinyl and other plastic fragments and fragments of rubber from hoses and
grommets
* A lesser portion of glass shards, soils and dirt (from wheel well, car underbody, and car interior -
not a sample contaminant), metal fragments, mostly small and thin, both steel and pot (zinc), wires
(copper), and chrome-plated plastic parts
* A very small portion of paper products, circuit boards (from car radios and other electronics) and
anything else of low density that might be put inside a car during its useful life9
The material was described as extremely non-homogeneous with particles from < 0,025 cm
(< 1/100 in) to > 30.48 cm (> 12 in). The material was collected from an air separation reclamation
facility.9
8
-------
A visual inspection of the as-received test material confirmed the presence of polyurethane
foam, various unknown plastics, coated and uncoated electrical wiring, compressed wood products, and
metal fragments. The fluff was combusted in a 0.61 m (24 in) x 0.56 m (22 in) diameter cylindrical,
steel vessel.
The fluff was contained in a wire mesh support placed within the combustion vessel. The
wire mesh support was used to allow adequate oxygen access within the combustion vessel.
Nominally, 11.4 kg (25 lb) of fluff was placed into the combustion apparatus for each of the three
tests. Before fluff ignition, the CEMs were operated for at least 15 min to establish background
levels. During this time, the conditioned air handling system was operating and continued to operate
throughout the test period.
After the baseline levels had been established, the fluff was manually ignited with a propane
torch. After 1-2 min, the torch was removed, and the burn hut door closed to leave a 102-mm (4-in)
opening for visual observation. After 5-15 min from the time the fluff was ignited, to allow sufficient
time for any propane combustion products to dissipate, the metal, dioxin, and organic semivolatile
sampling systems were activated. The volatile organic sampling train (VOST) was operated for three
separate intervals during each test to characterize volatile organic emissions under various fluff
combustion conditions. At the start and end of each sample condition, as well as continuously
throughout the duration of the test, the time and net weight of fluff in the combustion pit were
recorded digitally. Temperatures inside the bum hut (over the burn pit, at the deflector shield, and at
the entrance to the transfer duct), within the sample duct (in the bum hut and in the sample shed), and
dilution air temperatures were digitally recorded continuously throughout the duration of the test
In addition to these three tests, a "hut blank™ test was conducted between the second and third
actual tests. In this experiment, the test facility and sampling apparatus were operated as in an actual
test but no fluff was combusted. The purpose of this test was to assess the background levels of
9
-------
pollutants present in the test facility. In addition, field or laboratory blank samples were obtained for
many of the sampling trains.
2.3.2 CEMs
Fixed combustion gases were measured continuously using on-line process analyzers. Before
sampling, a four-point calibration (mm, span, and 2 mid range) was performed on each instrument to
verify analyzer linearity. Only three points (zero, span, and midrange) were available for the NO
analyzer. A zero and span check were performed before and after each sample period. This quality
control check was used to assess instrument performance and integrity and to validate collected data.
The analog output of each analyzer was converted digitally and then acquired on a computer using
30-s averages. The data were continuously stored and imported into a standard spreadsheet program.
Multiple copies of the electronic data are maintained in separate locations.
2.3.3 Volatile Organics
Volatile organics were collected using an unmodified VOST system operated according to
1 ft
Method 0030 found in SW-846. During this study, no stack probe was used. A heated section
(125 °C) of 0.63-cm (1/4-in) OD stainless steel tubing affixed to the sampling manifold was used to
transport the gaseous sample from the sample duct to the VOST system. The sample was drawn
through the Tenax® and Tenax®/eharcoal tubes at a flow rate at 0.25-0.5 L/min for 20-45 min. Total
sample volume collected was 4.5-10 L. Three sets of samples were collected during each test.
Sample rates and total volume collected were varied to maximize sampling time while minimizing the
risk of overloading the analytical system through excessive organic collection.
The sample tube sets were submitted to quality control contamination checks before use,
spiked with deuterated benzene, and were stored refrigerated at 4 °C in Teflon® bags both before and
after use. Daily field blanks were performed by transporting a tube set to the field and leaving it
exposed to the ambient atmosphere in the sampling shed for the entire period during which another
10
-------
tube set was installed in the sampling train and sampled. A laboratory blank sample was also
analyzed. All samples were analyzed within 30 days of collection.
The VOST samples were analyzed by GC/MS/F1D on a purge-and-trap thermal desorption
system. The effluent of the chromatographic column was split to each of the GC detectors for
simultaneous detection of eluting analytes from one sample. Method 5040 of SW-846 best represents
the procedure used for sample analyses.11 Compound identifications were accomplished using
multicomponent calibration standard comparisons, mass spectral library searches as well as investigator
interpretation. Identified analytes were quantified using GC/MS or GC/FID system responses. The
system selected for quantification was based on the characteristics of the compound identified.
Before calibrating the analysis of samples, the MS was tuned with perfluorotributylamine
(PFTBA) to linearize the MS over the total ion monitoring range of mass units (24-300 amu). The
MS was calibrated with a variety of volatile compounds, determining relative response factors between
the internal standard, D6-benzene, and the analyte of interest. A continuing calibration standard
containing known concentrations of acrolein, toluene, cyclohexane, hexanol, decane, D6 benzene, and
bromofluorobenzene (BFB) was analyzed daily to verify acceptable system performance. The FID
was calibrated by determining the linear response to varied concentrations of toluene containing
standards.
The Tenax® and Tenax®/charcoal samples were desorbed in a clamshell heater maintained at
220 °C, purging the organics for 10 min with helium at a nominal flow rate of 10 mL/min onto a
cryogenically cooled (0 °C) Tenax® trap. The tubes were analyzed in pairs and desorbed in reverse
direction from that sampled. The Tenax® trap was ballistically heated to 250 °C, and the carrier
directed onto a 30-m x 0.32-mm x 1.8-/«n film thickness DB-624 megabore column (J & W
Scientific). The column head pressure was 8 psig. The GC oven temperature was cryogenically
maintained at 0 °C for 5 min; then a temperature ramp was invoked at 3 °C/min until reaching
250 °C, where the temperature was held for 5 more min. As the sample constituents eluted from the
-------
column, they were split into two essentially equal streams and interfaced to the MS and FID,
respectively.
Compounds were identified using a combination of interpretative and confirmatory techniques.
Acquired mass spectra were compared to a mass spectral database. A probability-based matching
computer program assigned tentative identification based on the similarity of spectra. The computer-
generated match and sample spectra were evaluated manually. This information along with other
qualitative information such as compound boiling point and likelihood of presence, was used to aid in
tentative compound identification. In addition, acquired retention times and mass spectra were
compared to spectra and retention times of multi-component standard mixes analyzed under identical
conditions.
Identified and tentatively identified compounds were quantified using several approaches.
Individual compounds generated data based on both MS and FID responses as a result of splitting the
effluent of the column to the respective detectors. Therefore, the compounds could be quantified on
either system. Compounds present in multicomponent standards were quantified both by GC/MS and
GC/FID, in both cases using response factors specific to the individual compounds. The results of
these two methods were then averaged. Identified compounds not present in multicomponent
standards were quantified by GC/FID using the toluene response factor if they were identified as
hydrocarbons. All other identified compounds not present in multicomponent standards were
quantified by GC/MS using the response factor derived for the component of the multicomponent
standard with the most similar chemical composition in the judgement of the investigators. GC/MS
relative response factors (RRFs) were determined for individual compounds relative to the Dg-benzene
internal standard. The recovery of Dfi-benzene was determined relative to BFB. The quantification
method used for each analyte is presented with the analytical results in a later section of this report.
12
-------
2.3.4 Semivolatile Qrganies and Particulate Matter
Semivolatile organics and particulate matter were collected using a sample system modified for
use in this study to operationally separate semivolatile organics into gaseous and particulate-bound
fractions. A 0.95 cm (3/8-in) OD stainless steel tube connected the sampling manifold to a particulate
(Si
filter assembly. Particulate was collected on a 142-mm diameter Teflon -coated, glass fiber filter in
the filter housing. An ice water-cooled condenser was located between the filter assembly and the
XAD-2 sorbent module to cool the sample gas stream before contact with the XAD-2. The exit of the
sorbent module was connected to a pump and metering system. The gaseous sample was collected at
an average flow rate of 17.0-34.0 L/min (0.6-1.2 scfm) for approximately 3 h.
Two separate semivolatile organic/particulate collection systems were operated simultaneously
during the test period. One sample system was used to collect samples for the purpose of general
semivolatile organic and particulate characterization while the remaining system was used to collect
samples for polychlorinated dibenzodioxin (PCDD) and polychlorinated dibenzofuran (PCDF) analyses.
The only difference in operation between the two trains is the type of XAD-2 sorbent module used to
collect semivolatile organics. The stainless steel XAD-2 sorbent module used in the general organic
sampling system contained approximately 150 g of XAD-2 while the ice water-cooled glass XAD-2
module used for PCDD/PCDF collection contained approximately 40 g. Field and laboratory blanks
were collected for the general organic train and a field blank was collected for the dioxin train. Field
blanks consisted of filter and XAD-2 samples transported to the test facility along with the actual
samples. The laboratory blank consisted of a filter or XAD-2 module retained in the analytical
laboratory.
The Teflon®-coated, glass fiber filters used for particulate collection were desiccated, tared,
and placed in aluminum foil and a zip-lock bag before use. After sample collection, the particulate
samples were stored desiccated, weighed, and stored at 4 °C until extraction. Cleaned and quality
control checked (QC'd) XAD-2 resin was placed in sealed modules, sealed in teflon bap, and stored
13
-------
refrigerated before use. The XAD-2 was prepared and quality control checked as recommended in
typical source sampling procedures.12 The XAD-2 was subjected to the total chromatographable
organics (TCO) quality control check only. After sampling, the canisters were resealed in the Teflon®
bags and stored refrigerated at 4 °C until extracted. The tubing connecting the sampling manifold to
the filter housing and the front half of the filter housing were rinsed with the appropriate solvent
(acetone then dichloromethane for general analyses, acetone then toluene for PCDD/PCDF analyses),
collected, and combined with the particulate filter fraction. The back half of the filter housing, the
tubing connecting the filter housing to the condenser, the condenser, and the tubing connecting the
condenser to the XAD-2 sorbent module were also rinsed with the appropriate solvent, collected, and
combined with the XAD-2 fraction.
The semivolatile and particulate bound organics from the general organics samples were
retrieved from the collection media by soxhlet extraction using dichloromethane. The XAD-2 was
extracted separately from the particulate fraction. Following the > 16-h extraction, the samples were
concentrated using a 3-ball Snyder column system, filtered through a 0.45 /mi filter, and brought to
known volume. All organic concentrates were stored refrigerated. Both the particulate extracts and
the XAD-2 extracts were analyzed for TCO, organic compounds with boiling points between 100-
300 °C, and total gravimetric organics (GRAV), organic compounds with boiling points greater than
300 °C.12
To perform the GRAV analyses, aluminum weigh boats were prepared, desiccated, tared, and
then filled with 1.0 mL of the organic extract and allowed to evaporate. After evaporation, the boats
were again desiccated and weighed on two successive days until constant weight was reached.
Organic compounds with boiling points greater than 300 °C represent the net gain in mass. The
analyses were performed in duplicate and included performance evaluation samples.
The TCO analysis was done by GC/FID. The data from the GC/FID analyses were also used
for individual compound quantification as described below. A multipoint calibration was conducted
-------
using an alkane mix standard. The O,, C10, C12> C14, C17 n-alkane mix was used to establish the
boiling point retention window as well as the system response. All peaks with retention times falling
between, but not including, the C7 and C17 retention times, were quantified. The system response to
the CIO, C12, and C14 alkanes was used for quantification. The analysis was performed using a 30-m
x 0.25-mm x 0.25-^m film thickness DB-5 column (J & W Scientific). For the XAD-2 samples, the
1-fxL, injection was made with the oven temperature held at 40 °C for 3 min, ramped to 250 °C at
5 °C/mint and held for 5 min after reaching final temperature. For the filter samples, the 1-/*L
injection was made at 40 °C, which was held for 3 min. The oven was then ramped at 14 °C/min to
170 °C. The temperature was then ramped at 4 °C/min to 255 °C then at 2 °C/min to 300 °C where
it was held for 2 min. A separately prepared quality control standard was injected at the beginning
and end of each analytical day to verify instrument performance.
The GC/MS data were acquired on a system configured for capillary column use. A 30-m x
0.32-mm x 0.25-^m film thickness DB-5 column (J & W Scientific) was directly interfaced to the MS
source (interface temperature = 300 °C). Before calibrating or analyzing the samples, the MS was
tuned with perfluorotributylamine (PFTBA). Continuing calibration standards were injected at the
beginning and end of each analytical day to verify the consistency of instrument performance.
Injections of 1-2 pL were made into a splitless injector maintained a 300 °C. The XAD-2 and
particulate fractions were analyzed using different temperature programs. The XAD-2 fractions were
analyzed with die initial oven temperature of 40 °C maintained for 3 min. The oven temperature was
then ramped at 5 °C/min to 300 °C and held for 5 min. The particulate fractions were analyzed with
the initial oven temperature of 40 °C maintained for 3 min, at which time the temperature was ramped
at 14 °C/min to 170 °C. The temperature ramp was changed to 4 °C/min until reaching 255 °C, at
which time the temperature ramp was changed to 2 °C/min until reaching 300 °C where the
temperature was maintained an additional 2 min.
15
-------
Individual semivolatile organic compounds were identified and quantified using an approach
similar to that used for the volatile organies except that the GC/FID and GC/MS data were acquired in
two separate injections rather than in the same analysis. The XAD-2 and particulate extracts were
analyzed by GC/MS to obtain mass spectral information for qualitative identification. The mass
spectra of acquired data were compared to mass spectra of multicomponent standard mixes as well as
the mass spectral database to assign compound identifications. Multicomponent alkane standard mixes
were analyzed on both systems to aid in the comparison of retention times. Compound identifications
determined during the GC/MS analyses were cross referenced to the FID analyses using retention
time/retention indices comparisons.
Quantification of individual compounds, for which quantitative calibration data were available,
was based on the mass spectrometry data after determining RRFs between Dg-naphthalene and
compounds present in calibration standards. The internal standard, Dg-naphthalene, was added to each
sample extract before mass spectrometry analysis. Non-hydrocarbon compounds for which individual
standards were not available were quantified based on the mass spectrometry data using the response
factor between the standard most similar in chemical composition and Dg-naphthalene. Quantification
of hydrocarbon compounds for which no specific response factor was available was performed from
the FID data using the average response factor obtained on the FID system for the C10, C12» and C14
normal alkanes. The internal standard, Dg-naphthalene, was added to each sample extract before mass
spectrometry analysis. The quantitative method used for each analyte is presented with the analytical
results in Section 3.4 of this report
A portion of the XAD-2 and particulate extracts was submitted for separate polyaromatic
hydrocarbon (PAH) analyses to a contracted laboratory. The samples were analyzed by a GC/MS
procedure specific to the analysis of a group of 16 PAHs. The standards for this analysis were
prepared from commercially available PAH standards with the addition of 9,10-dichloroanthracene and
Djj-perylene as external standards. Initial calibration was performed at six different concentration
-------
levels and response factors from each level were averaged. The MS was calibrated with PFTBA and a
continuing calibration standard was run at the beginning of each analytical day. The mass
spectrometry was performed in the selected ion monitoring mode using the parent (M+) ion for
quantification and the (M+l+) ion for confirmation of analyte identification.13
The PCDD/PCDF samples were analyzed by a hybridized method utilizing techniques found in
SW-846 Method 8280 and 40 CFR Part 60, Appendix A, Method 23.14,15 The samples were analyzed
by low resolution GC/MS where isotopically labeled homologues for all congeners were used for
qualitative and quantitative purposes. The analytical method used does not identify individual isomers
within each congener group but does, however, quantify each isomer chromatographically resolved
within each congener group. The data are reported here in terms of the total analytical concentration
within each congener group.
2.3.5 Metal Aerosols
Particulate matter containing metal aerosols was collected using a separate sampling system to
characterize airborne metals emissions. A gaseous sample was drawn across a 142-mm diameter
quartz fiber filter under vacuum at an average flow rate of 17.0-42.5 L/min (0.6-1.5 cfm) for about
3 h. The quartz filters used were desiccated and tared, then placed in aluminum foil and a zip-lock
bag before use. A field blank sample, consisting of a filter transported to the test facility along with
the actual samples, was also obtained. Following sample collection, the samples were again desiccated
and weighed. Ultimately, the samples were delivered to a contracted analytical laboratory for metals
quantification. Metals potentially present in fluff samples were chosen for quantification. The
samples were analyzed by inductively coupled argon plasma (ICAP) for aluminum, arsenic, barium,
cadmium, total chromium, copper, lead, magnesium, selenium, and zinc.16
17
-------
2.3.6 PM,„ Sampling
A Texas A&M medium volume ambient particulate sampler (Figure 2-4), similar to the
Andersen Series 254 medium flow air sampler, was used to collect particulate of 10 wm in diameter
and less. The sampler is designed so that when a flow of 0.11 actual m/min (4 acfm) is maintained
across the system, only particulate of 10 pm in diameter and less is collected on the filter. The 142-
mm diameter glass fiber filters used for particulate capture were desiccated and tared, then placed in
aluminum foil and a zip-lock bag before use. A field blank sample, consisting of a filter transported
to the test facility along with the actual samples, was also obtained. Following sampling, the filters
were desiccated and weighed to determine total mass of particulate matter collected.
18
-------
Air Flow
142 MM
Filter Holder
To Sample Shed
(Vacuum Pump and Dry Gas Meter)
Figure 2-4. PM^medium volume particulate sampler.
19
-------
SECTION 3
DATA, RESULTS AND DISCUSSION
3.1 GENERAL
As previously mentioned, the first of the two objectives of this study was to identify the
poorly characterized emissions resulting from the simulated open combustion of Buff. The second
objective was to obtain estimations for these emissions accurate to within one order of magnitude.
This necessitated an approach in which the breadth of information was given the greatest emphasis and
quantitative data were obtained using methods with a degree of analytical variability appropriate to
generating estimated emissions accurate to within an order of magnitude. It was hoped that this
qualitative approach would provide the data and insight to direct subsequent, more specialized
investigations.
3.2 COMBUSTION CHARACTERISTICS
Table 3-1 presents a summary of the data pertaining to combustion performance over the
course of the three combustion tests. Nominally, 11.4 leg (25 lb) of fluff was evaluated during each
test As Table 3-1 indicates, not all of the material tested was actually combusted during testing.
Indeed, only approximately 40 percent of the mass of fluff placed in the combustion apparatus was
actually combusted in the 3-h test. Because the dilution air added to the burn hut during testing was
cut off at the end of each test day and the door sealed, it is unknown whether all combustible material
present in fluff was completely burned. The remaining ash and incombustible material were not
characterized.
20
-------
TABLE 3-1. COMBUSTION CONDITIONS IN THREE FLUFF TESTS
Day 1
Day 2
Day 3
Mass fluff at start (kg)
11.3
10.7
11.2
Weight after 200 min of combustion (kg)
5.8
5.8
6.2
Fluff mass lost because of combustion in 200 min (%)
48.7
45.8
44.6
Average burn rate over 200 min test (kg/hr)
1.65
1.47
1.50
Average bum rate over sampling period (kg/hr)
1.56
1.50
1.50
Length of sampling period (min)
180
183
174
Total mass of fluff combusted was determined by subtracting the final mass of material in the
test apparatus from the initial mass. The net mass of fluff combusted divided by the duration of the
test period determined the average burn rate for each test Table 3-1 also presents the average burn
rate of fluff for each 3-h test. Given the non-homogeneity of the composition of the fluff, there is
excellent agreement (less than 20 relative percent difference) between the average burn rates of the
three tests.
Bum rates were also determined for smaller elapsed periods of time. Burn rates were
determined by relating the mass of fluff combusted to the length of time the mass was burned. Figure
3-1 represents the burn rates relative to elapsed time for each of the three tests. Maximum bum rates
were observed within 20 min of material ignition. After this time, burn rates gradually decreased
throughout the burn. Figure 3-2 presents the temperatures observed by a thermocouple placed directly
over the combustion apparatus. Peak temperatures correlate well with peak burn rates.
Figures 3-3 to 3-6 present the continuous emissions monitoring data for CO, C02, THC, and
NO, respectively. Peak concentrations correlate reasonably well with peak burn rates. However, the
THC data reveal peak emissions at periods farther into the test than the observed peak burn rates. The
oxygen data are not presented graphically. Over the course of the bums, 02 concentrations remained
greater than 19 percent indicating that conditions adequately simulated an open combustion
environment where combustion would not be expected to be oxygen starved. For purposes of clarity,
-------
PRETEST BURN RATE SET = 0
11
10
9
8
7
6
5
4
3
2
1
0
ELAPSED TIME SINCE IGNITION (min)
¦ FIRST TEST * SECOND TEST » THIRD TEST
Figure 3-1. Bum rates of fluff.
22
-------
600
500
400
300
200
00
-20
20
60
140
180
220
100
time (min since start of test)
first fluff test + second fluff test o fluff hut blank A third fluff test
Figure 3-2. Temperatures over bum pit
23
-------
600
TIME SINCE FLUFF IGNITION (mm)
O FIRST TEST + SECOND TEST O THIRD TEST
Figure 3-3. CO concentrations in three fluff tests.
24
-------
0.9
0.6 i-
0.3 >~
0.2 -
-30 -10
10
130
150
170
30
50
90
1 10
190
70
TIME SINCE FLUFF IGNITION (min)
O FIRST TEST + SECOND TEST o THIRD TEST
Figure 3-4.
C02 concentrations in three fluff tests.
25
-------
POST TEST QC FAILED TESTS 1-«-2
-tj
TIME SINCE FLUFF IGNITION (rnin)
FIRST TEST + SECOND TEST O THIRD TEST
Figure 3-5, THC concentrations in three fluff tests.
26
-------
NOTE FIRST "EST FAILED PC ?T TEST QC
-
i
_
~
~
o
c>
%
0
sS
o
-
~ ,
0 +
GO B
°
1
0
f>
1
i
10 30 50 70 90 no 130 150 170 190
TIME SINCE FLUFF IGNITION (mm)
~ FIRST TEST + SECOND TEST o THIRD TEST
Figure 3-6. NO concentrations in three fluff tests.
27
-------
hut blank CEM data are not shown graphically. Hut blank concentrations held quite stable at
background values throughout the test (CO between 10-25 ppm, C02 between 0.02 and 0.06 percent,
NO at 1 ppm, 02 > 21 percent, THC < 10 ppm).
3.3 VOLATILE ORGANIC EMISSIONS DATA
GC/MS analysis of the VOST samples collected yielded the identification of more than 50
compounds. However, for the range of volatile compounds characterized (retention times up to and
including benzaldehyde), more than 100 peaks were evident in the GC/FID chromatograms. The
majority of the compounds identified were alkanes, alkcnes, cycloalkanes, and alkyl substituted
aromatics. However, aldehydes, ketones, alcohols, nitriles, and chlorinated aromatics were also
identified. The types of volatile compounds identified are consistent with compounds identified during
thermal decomposition studies of individual plastics.18*21 Because of the diversity of the plastics
present in fluff, it is not possible to attribute the identified products of incomplete combustion (PICs)
to any one type of plastic.
The mass of volatile organic emissions was characterized in several ways. Volatile organics
were characterized as total mass emitted based on the FID response to toluene. The integrated areas of
peaks up to and including toluene were summed and used to give an estimate of volatile organic
matter with boiling points up to 110 °C. Table 3-2 presents summary data of total volatile organic
emissions for the respective burn rates observed. The observed burn rates were a function of how
long into the test the samples were collected; burn rates decreased as the elapsed time of the test
increased. Figure 3-7 graphically represents the observed linear relationship between burn rate and
observed volatile organic emissions. As burn rates decreased, the mass of volatile organics relative to
fluff mass combusted increased. This observation can be further confirmed by observing that the total
hydrocarbon emissions (Figure 3-5) peaked much later in the test than the burn rate (Figure 3-1).
28
-------
90
80
70
60
50
40
30
20
10
0
Notes:
Linear regression does not include samples in which trap freezing was noted.
Linear regression r3 = 0.959,
¦Trap Frozen
¦Trap Frozen
J i i i i i i i i i 1 1 i
0 0.02 0.04 0.06 0.08 0.1 0.12
Bum Rate (kg/min)
Figure 3-7. Total volatiles vs. bum rate.
29
-------
TABLE 3-2. VOC MASS AND BURN RATE DATA
Experiment
Sample
Sequence
Total
Volatiles
(ng)
Estimated Volatile
Organic Emissions
(g/kg)
fluff Combustion
Rate (kg/min)
1
1
85347
6.47
0.1134
1
2
42471
21.19
0.0184
1
3
74876
69.40
0.0091
1
EB
1417
NAP
NAP
Avg. of experiment 1 samples
67565
3236
0.0469
2
1
223007
47.72
0X5389
2
2
115491
77.67
0.0124
2
3
NAV
NAV
0.0084
2
EB
888
NAP
NAP
Avg. of experiment 2 samples
169249
62.69
0.0256
HB
1
1787
NAP
NAP
HB
2
3805
NAP
NAP
HB
3
1719
NAP
NAP
HB
FB
2977
NAP
NAP
Avg. of hut blank samples
2437
NAP
NAP
3
1
166277
29.11
0.0475
3
2
129743
79.61
0.0130
3
3
56971
75.44
0.0063
3
FB
234
NAP
NAP
Avg. of experiment 3 samples
117664
61.39
0.0223
Lab Blank
Lab Blank
224
NAP
NAP
Avg. of all Actual Test Samples
111773
50.83
0.0324
Note: HB = Hut Blank, FB = Field Blank, NAV = Not Available, NAP = Not Applicable
The estimated emissions presented are estimations based on several variables. They were
calculated by assuming that the dilution air added to the bum hut was at a known, constant volume
and that the volume of air added to the burn hut equalled the volume exiting the hut. It was also
assumed that the gas mixture collected in the sample duct was well mixed and representative of the
gas mixture found throughout the burn hut. The average volatile organic gaseous concentration,
30
-------
determined by dividing the mass collected by the volume sampled, was multiplied by the volume of
air added to the burn hut per unit time. This represents the mass of organic material emitted per unit
time. Dividing by the respective fluff burn rate yields the mass of volatile organics emitted relative to
the mass of fluff combusted.
Table 3-3 presents emissions data for selected individual compounds identified in the VOST
samples, and Table 3-4 presents a list of all individual compounds tentatively or positively identified
in VOST samples. Emphasis was placed on establishing emissions data for compounds found on the
Clean Air Act Amendment (CAAA), Title III, Hazardous Air Pollutants (HAP) list. Figure 3-8
graphically depicts the major volatile organic compounds (VOCs) found on this list and their relative
contributions. Benzene is one of the top two volatile organic compounds emitted, generating nearly 10
g for every kilogram of fluff consumed in combustion. Benzene is also one of the more toxic
compounds identified.
A comparison of the total VOC mass to the mass of individually identified compounds with
boiling points within this range, indicates that greater than 70 percent of the detectable mass has been
characterized. The remaining fraction was not identified because of the complexity of the sample,
limitations of the analytical system, and budgetary constraints.
Very few detectable peaks were seen in the field and laboratory blank samples. Compounds
present in field and laboratory blanks included methylene chloride, acetone, trichlorofluoromethane,
toluene, and benzaldehyde. All compounds present in these blanks were either not present or present
at more than an order of magnitude greater concentration in the actual samples. The compounds that
were detected at low concentrations in the blanks but not in the samples may well have been also
present in the samples but obscured by interfering analytes at much higher concentrations.
31
-------
40
35
30
25
20
IS
10 -
R
1
/
_
i
$
TEST 1, SAMPLE 1 I TEST2, SAMPLE 1 I TEST 3 SAMPLE 2
TEST 1, SAMPLE 3 TEST 2, SAMPLE 2 TEST 3, SAMPLE 3
BENZENE TOLUENE
CHLOROBENZENE [ZZJ ETHYL BENZENE ~~ M/P XYLENE
Figure 3-8. Estimated emissions for selected CAAA HAPs.
32
-------
TABLE 3-3. ESTIMATED EMISSIONS FOR SELECTED VOLATILE ORGAN1CS (continued on next page)
(g PollutantAg Fluff combusted)
Pollutant
Identification confirmed
by comparison to
standard?
Test 1
Sample 1
Test 1
Sample 3
Test 2
Sample 1
Test 2
Sample 2
Test 3
Sample 2
Test 3
Sample 3
Test 1
Avg.
Test 2
Avg.
Test 3
Avg.
ACET ALDEHYDE
NO
0.315
0.437
1.295
1.507
1.016
ND
0.376
1.401
0.508
n-Pentane
NO
ND
7.481
2.052
5.191
4.700
3.955
3.741
3.622
4.328
ACROLEIN
YES
0.258
1.437
1.044
3.220
2.344
1.766
0.848
2.132
2.055
ACETONITRILE
NO
0.106
0.416
0.441
0.000
0.964
2.900
0.261
0.221
1.932
ACRYLONITRILE
NO
0.304
0.558
0.886
0.714
0.912
1.260
0.431
0.800
1.086
C5H60 Methyl Furan
NO
ND
ND
0.076
0.113
0.137
ND
0.000
0.095
0.068
BENZENE
YES
2.846
9.989
9.138
9.811
9.083
16.635
6.417
9.475
12,859
1-Heptene
NO
ND
1.485
0.695
1.572
1.395
0.862
0.743
1.134
1.128
n-Heptane
YES
ND
1.019
ND
1.076
0.944
0.720
0.510
0.538
0.832
CSH802, Carboxylic Acid,
Methyl Ester
NO
0.007
ND
ND
0.072
ND
ND
0.003
0.036
0.000
TOLUENE
YES
0.564
6.634
4.739
8.626
8.674
34.920
3.599
6.683
21.797
n-Octane
YES
ND
ND
ND
NAV
0.999
0.725
0.000
0.000
0.862
C9H18
Trimethylcyclohexane
NO
ND
0.859
ND
NAV
0.542
0.365
0.429
0.000
0.454
CHLOROBENZENE
YES
0.053
0.966
0.891
NAV
1.851
5.176
0.510
0.891
3.513
ETHYL BENZENE
YES
0.076
4.086
2.188
NAV
4.242
15.974
2.081
2.188
10.108
m/p-XYLENE
YES
0.059
1.477
1.038
NAV
2.200
3.675
0.768
1.038
1938
Styrene
NO
0.943
5.650
7.128
NAV
10.055
6,801
3.296
7.128
8.428
n-Decane
YES
ND
ND
ND
NAV
1.423
0.889
0.000
0.000
1.156
Bcnzaldehyde
YES
0.219
ND
ND
NAV
2.749
3.409
0.110
0.000
3.079
CAAA HAPi lifted in <11 apt
NAP - Not Applicable
NAV - No; AviiliMe
ND - Not Detected
-------
TABLE 3-3. ESTIMATED EMISSIONS FOR SELECTED VOLATILE ORGANICS (concluded)
Pollutant
Identification confirmed
by comparison to
standard?
Sample 1
Avg.
Sample 2
Avg.
Sample 3
Avg.
Overall
Avg.
Method used
for
quantification
Compound used to generate
MS response factors
ACETALDEHYDE
NO
0.805
1.262
0.218
0.762
MS
Acrolein
n-Pentane
NO
1.026
4.946
5.718
3.897
FID
NAP
ACROLEIN
YES
0.651
2.782
1.601
1.678
FID + MS
Acrolein
ACETONITRILE
NO
0.274
0.482
1.658
0.805
MS
Aciylonitrile
ACRYLONITRILE
NO
0.595
0.813
0.909
0.772
MS
Acrylonitrile
C5H60 Methyl Furan
NO
0.038
0.125
0.000
0.054
MS
Dibenzofuran
BENZENE
YES
5.992
9.447
13.312
9.584
MS
Benzene
l-Heptene
NO
0348
1.484
1.173
1.001
FID
NAP
n-Hepiane
YES
0.000
1.010
0.869
0.626
FID
NAP
C5H802, Carboxylic Acid,
Methyl Ester
NO
0.003
0.036
0.000
0.013
MS
Benzaldehyde
TOLUENE
YES
2.651
8.650
20.777
10.693
FID + MS
Toluene
nOctane
YES
0.000
0.999
0.362
0.345
FID
NAP
C9H18
Trimethylcyclohexane
NO
0.000
0.542
0.612
0.353
FID
NAP
CHLOROBENZENE
YES
0.472
1.851
3.071
1.787
MS
Chlorobenzene
ETHYL BENZENE
YES
1.132
4.242
10.030
5.313
FID + MS
Ethyl Benzene
m/pXYLENE
YES
0.548
2.200
2.576
1.690
FID + MS
p-Xylene
Styrene
NO
4.035
10.055
6.226
6.115
FID
NAP
n-Decane
YES
0.000
1.423
0.445
0.462
FID +MS
Decane
Benzaldehyde
YES
0.110
2.749
1.704
1.275
MS
Benzaldehyde
CAAA HAPi lined in all capa
NAP - Not Applicable
NAV - Not Available
HD - Noi Detected
-------
TABLE 3-4. VOLATILE ORGANIC COMPOUNDS IDENTIFIED (continued on next
(Listed in order of increasing retention time)
Compound Name or Class
Molecular Formula
Alkene
Diene or Alkyne
Alkene
Diene or Alkyne
Acetaldehyde
Alkene
Alkene
Alkene or Cyclic
n-Fentane
Alkene or Cyclic
Diene
Alkylated Cyclopropane
Alkene
2-Propenal (Acrolein)
Propanal
Diene
Unsaturated Hydrocarbon
Acetonitrile
Alkane
Alkene
2-Propenenitrile (Acrylonitrile)
Alkene or Cyclic
I iexane
2-Methyl-2-Pro penal
Alkene
Branched Alkene
Alkene
Methyl Furan
Diene or Alkyne
Alkene
Nitrile
Alkene Substituted Cyclic or Diene
Substituted Cyclopropane
Benzene
Alkene
1-Heptene
n-Heptane
Probably a Heptene
Alkene
Alkene
Alkyne or Diene
Alkene
Alkene
Carboxylic Acid, Methyl Ester
Alkyl Substituted Cydopentane
Branched Alkane
C4H8
C4H6
C4H8
C4H6
C2H40
C4H8
C4H8
C5H10
C5H12
C5H10
C5H8
C5H10
C5H10
C3H40
C3H60
C5H8
C5H6
C2H3N
C6H14
C6II12
C3H3N
C6H12
C6H14
C4H60
C6H12
C6H12
C6H12
C5H60
O6H10
C7H14
C4H5N
C6H10
C7H14
C6H6
C7H14
C7H14
C7H16
C7H14
C7H14
C7H14
C7H12
C7H14
C7H14
C5H802
C8H16
C8H18
35
-------
TABLE 3-4. VOLATILE ORGANIC COMPOUNDS IDENTIFIED
(concluded)
Compound Name or Class
Molecular Formula
Toluene
C7H8
Alkene or Cyclic
C8H16
Alkene
C8H16
n-Octane
C8H18
Alkene or Possibly Cyclic
C8H16
A Trimethylcydohexane
C9H18
Cydopentanonc
CSH80
Oxygenated Hydrocarbon
C8H140
A Trimethyl Cydohexane
C9H18
Chloro benzene
C6H5Q
Ethyl Benzene
C8H10
m-Xylene and or p-Xylene
C8H10
1-Nonene
C9H18
Ethynyl Benzene
C8H6
Styrene
C8H8
Alkylated Aromatic
C9H12
Alkylated Aromatic
C9H12
Alkylated Aromatic
C9H12
Alkylated Aromatic
C9H12
Alkene
C10H20
n-Decane
C10H22
Alkylated Aromatic
C9H12
Alkylated Aromatic
C9H10
Benzaldehyde
C7H60
36
-------
Few detectable peaks were seen fa the hut blank samples. The compounds that were present,
included dichlorodifluoromethane (Freon 12), allyl alcohol, propanol, methylene chloride,
trichloroethane, 1,4-dioxane, and toluene. All compounds present in these blanks were either not
present or present at more than an order of magnitude greater concentration in the combustion test
samples. The one exception was a single sample (the second obtained during the hut blank test) in
which toluene was detected at a level only a factor of two lower then the lowest sample toluene
concentration (the third sample of the first combustion test). In summary, the volatile organic blank
data support the validity of the sample data presented.
3.4 SEMIVOLATILE ORGANICS DATA
The characterization of semivolatile organic emissions collected on both the XAD-2 and
particulate filters used an approach similar to that used to characterize the volatile organics emissions.
TCQ and GRAV analyses were performed separately on the XAD-2 and particulate filter extracts to
determine total organic content based on boiling point range.
Table 3-5 provides a summary of these data expressed as estimated emissions. As expected,
the XAD-2 sample fractions contained more chromatographable (vapor phase) mass than did the
particulate filter fractions. Similarly, the particulate filter fractions contained more GRAV
(condensable) mass than did the XAD-2 fractions. Of the total extractablc semivolatile and non-
volatile organic mass emitted, an average of 49.9 percent was contained on the XAD-2 fraction while
the remainder was contained on the particulate. The TCO and GRAV values for the field, hut, and
laboratory blanks were quite low compared to the sample values. No blank was greater than 21
percent of the lowest corresponding sample concentration.
GC/MS analyses were performed on each of the sample fractions to aid in identification of
individual compounds. Table 3-6 lists the more than 45 compounds identified in the XAD-2 and
particulate fractions. The compounds identified are similar to the types of compounds identified in the
37
-------
VOST samples. In addition, phenols, PAHs, phthalales, and heteroeycles were identified. Again, the
types of compounds identified were consistent with the types of organics identified in various studies
of the thermal decomposition of individual plastics.18"21
TABLE 3-5. ESTIMATED ORGANIC EMISSIONS BY SAMPLE FRACTION (g/kg)
Day 1
Day 2
Day 3
Average
Volatiles - Early in Test
5.93
43.06
25.73
24.91
Volatiles - Mid test
17.28
69.16
70.87
52.44
Volatile® - Late in Test
62.54
NA
58.44
60.49
XAD-2 TCO
56.99
50.05
90.72
65.92
XAD-2 GRAV
6.68
10.12
23.58
13.4
Particulate TCO
0.61
1.27
0.85
0.91
Particulate GRAV
53.78
69.11
113.71
78.87
Particulate - General Organic Train
91.25
116.17
183.44
130.28
Particulate - Dioxin Train
85.82
115.68
188.63
130.04
Particulate - Metals Train
81.55
89.93
174.49
115.32
Average Particulate - 3 Trains
86.20
107.26
18219
125.22
P | Q
66.03
NA
41.11
53.57
Combustion Rate (kg/min)
0.028
0.024
0.024
0.025
38
-------
TABLE 3-6. ESTIMATED EMISSIONS FOR ORGANIC COMPOUNDS COLLECTED ON XAD RESIN AND PARTICULATE FILTERS
(Listed in increasing order of retention time) (continued cm next page)
1 Confound
XAD
Tea 1
XAD
Ten 2
XAD
Tea 3
XAD
Ayg.
XAD
Hut
Blink
Phiticulitc
Tea 1
Putawbtc
Tea 2
Putkulate
Tea 3
hniculaie
Avg.
Paniculate
lltx Blank
Identification
caafiimod by
campariacn to
known aundarda?
Quantification
Method
Compound uicd
for MS RFs
1 ETHYL BENZENE
2.26
205
190
2.40
ND
ND
ND
ND
0.000
ND
YES
MS
EthyBmzene
rn/p-Xfmm
1.03
1.11
1.72
1.29
ND
ND
ND
ND
0.000
ND
YES
MS
p-Xylone
Eihynyl Benzene
0.38
0.39
0.61
0.44
ND
ND
ND
ND
0.000
ND
NO
FID
NA
STYRENE
«JT7
6.49
11.81
8.19
ND
ND
ND
ND
0.000
ND
NO
FID
NA
l-MethyI-2-Pentyl Cyclopropane
0.46
ND
0.05
0.17
ND
ND
ND
ND
0.000
ND
NO
FID
NA
C9H12 AkyUied Benzene
0.95
1.05
1.15
1.05
ND
ND
ND
ND
0.000
ND
NO
HD
NA
BennldchyJe
1.20
1.53
2.34
1.69
ND
ND
ND
ND
0.000
ND
YES
MS
Benzildehyde
Ediyhaloene
o,
-------
TABLE 3-6. ESTIMATED EMISSIONS FOR ORGANIC COMPOUNDS COLLECTED ON XAD RESIN AND PARTICULATE FILTERS
(Listed in increasing order of retention time) (continued on next page)
Compound
XAD
Tat 1
XAD
Ten 2
XAD
Ten 3
XAD
Avg.
XAD
Hut
Blink
Paniculate
Ten 1
Paniculate
Ten 2
PuticulMc
T««t 3
Paniculate
Avg.
Articulate
Hut Blank
Metniiieatiim
ccofiamsd by
companion Id
known aundaitb?
Qualification
Method
Compound sued
for MS RFt
CI 11110 Alkyl Subfdtiited
Artmitic
0,14
0.14
0.1S
0.16
ND
ND
ND
ND
0.000
ND
NO
FID
NA
CI 1H10 Alkyl SubctUutod
Aramitic
ND
0.33
0,41
0.25
ND
ND
ND
ND
0.000
ND
NO
HD
NA
PTHAUC ANHYDRIDE
ND
0.41
0.23
ND
ND
ND
ND
0.000
ND
NO
MS
Bn [2-EthyIhcxyl} Phthaktc
Benzenebuianmithle
3,40
2.48
4. IS
134
ND
ND
ND
ND
0.000
ND
NO
MS
Benzoruirile
BIPHENYL
0.29
0.30
0.39
0.33
ND
ND
ND
ND
0,000
ND
YES
MS
Biphcnyl
C14H28 Alkaw at Cyclic
0.44
ftl3
0.19
125
ND
ND
ND
ND
0.000
ND
NO
FID
NA
ACENAPHTHYLENE
0.06
0.2?
0J1
o.»
ND
ND
ND
ND
0.000
ND
YES
FID
NA
CAJPROLACTAM
ND
ND
ND
0.00
ND
ND
0.380
0.15
0.177
ND
YES
MS
Cipjoltcum
lH-Isoindole-1,3(2H)-Dicii«
ND
ND
asi
0.27
ND
ND
0.628
0.81
0.479
ND
NO
MS
Bcmonitrilc
C11H7K Arocnttk
0.55
0.36
0.64
0.51
ND
ND
ND
ND
0.000
ND
NO
MS
Bcnzoniirile
I,IXh3-Prop*nodiyl)Bi»-B«tiiei*;
0.71
0.47
0l75
0.64
ND
0.063
0.225
0.156
0.148
ND
NO
FID
NA
HcpUdectne
ND
ND
ND
0.00
ND
0.072
0094
ND
0.055
ND
YES
FID
NA
PHENANTHRENE
0.395
0.241
0.45
0.36
ND
0.053
0,098
0.119
0.090
ND
YES
FID
NA
C18H36 Alkenc
ND
ND
ND
0.00
ND
0.0185
ND
ND
0.026
ND
NO
FID
NA
Ocudecuw
ND
ND
ND
0.00
ND
0.087
0.105
0.0851
0.092
ND
YES
FID
NA
C16M2 PAH
ND
ND
ND
0.00
ND
ND
0.0681
ND
0.023
ND
NO
FID
NA
Hontdecne
ND
ND
ND
0.00
ND
0.121
0.151
0,1688
0.147
ND
YES
FID
NA
C20H40 Alkene
ND
ND
ND
0.00
ND
0.147
0.131
0.181
0.153
ND
NO
FID
NA
Ekaaue
ND
ND
ND
0.00
ND
0.154
0.154
ND
0.103
ND
NA
NA
NA
CAAA HAPj listed in all capi
NA = Nat Available
ND = Nat Detected
-------
TABLE 3-6. ESTIMATED EMISSIONS FOR ORGANIC COMPOUNDS COLLECTED ON XAD RESIN AND PARTICULATE FILTERS
(Listed in increasing order of retention time) (concluded)
Compound
XAD
Te* 1
XAD
Teat 2
XAD
To* 3
XAD
A**
XAD
Hi*
Blink
Paniculate
Te«t 1
Particulate
Te*2
PuticuLite
Ten 3
Paniculate
Avg.
Putktilate
Hut Blank
Idoitifkatian
confirmed by
canpuucn to
known naadarda?
Qiutificatioo
Method
Compound used 1
for MS RFi |
Teiphenyl
ND
ND
ND
o.oo
ND
0.050
0.070
ND
0.040
ND
NO
FID
NA
Docohoc
ND
ND
ND
0,00
ND
0.112
0.U8
0.1489
0.126
ND
YES
FID
NA
Tricoune
ND
ND
ND
0.00
ND
0.094
0.112
0.1348
0113
ND
YES
FID
NA
Tcuicohiu
ND
ND
ND
0.00
ND
0.091
0.116
ND
0.069
ND
YES
HD
NA
C22H3404 Benzene Dicurimiyiic
Add
ND
ND
ND
0.00
ND
ND
ND
0.26
0.087
ND
NO
MS
Bu{2-Gth)rlhei}?l)Phlha]ale I
C243S04 Benzene Dicutoiylic
Add or Phthalals
ND
ND
ND
0.00
ND
ND
ND
0,156
0052
ND
NO
MS
Bia(2-Ethylhexyl)Phthalate 1
1,2Ben2en«ii£»rt>ooiylic Acid,
Dibeptyl Eiier
ND
ND
ND
0.00
ND
ND
0.130
ND
0.043
ND
NA
NA
NA
ND
ND
ND
0.00
ND
am
0,098
ND
0.070
ND
YES
FID
NA
Bi»(2-Elhylheiyl)Ptub*lHe
ND
ND
ND
(LOO
ND
0.761
1.995
3.418
2058
ND
YES
MS
Bi»(2-Ethylheiyl)Phth«laic
Quntaphenyl
ND
ND
ND
0.00
ND
01073
0.039
ND
0.037
ND
NO
FID
NA
Triconurie
ND
ND
ND
OlOO
ND
0.023
ND
ND
0.008
ND
YES
FID
NA
CAAA HAPa lilted in all caps
NA = Not Available
ND = Not Detected
-------
Similarly, Table 3-6 presents emissions data for individual organic compounds. The majority
of the compounds identified common to the HAP list were PAHs. Good agreement exists between
estimated emissions of compounds identified in both the VOST and XAD-2 sample fractions. Because
of the complexity of the sample and its components, identification of all compounds present in the
organic fractions would be costly and not within the scope of this study. More than 200 peaks were
integrated during the GC/F1D analyses, many of which contained coeluting compounds.
Quantitatively, 37 percent of the total mass of the XAD-2 fractions were characterized by identifying
individual compounds.
The particulate fraction did not demonstrate the same level of organic characterization. A
large portion of the particulate (69 percent) was methylene chloride extractable and thus can be
assumed to be organic; however, a substantial portion of the particulate catch was not methylene
chloride extractable (31 percent). In addition, because many of the organics had very high boiling
points, characterization of these organics was not practical by GC techniques used. Finally, the spectra
of many of the compounds that were amenable to GC/MS analysis did not give an adequate library
match and the resources available to this project did not allow investigator interpretation of all of these
unknowns from first principles. For this fraction, only 3 percent of the extractable organic mass (2
percent of the total particulate mass) was characterized by individual compounds. The estimates of the
percentage extractable organic mass characterized were based on a ratio of the GC/FID concentration
of all identified compounds to the total extractable organic mass (GRAV + TCO). The estimates of
the percentage total particulate mass characterized were based on a ratio of the GC/FID concentration
of all identified compounds to the total particulate catch in the general organic train.
There were very few compounds detected in the hut, field, and laboratory blanks collected for
the XAD-2 and particulate organic samples. Any compounds that were detected in blanks were either
42
-------
not detected in the actual samples or were at least an order of magnitude less concentrated than in the
actual samples.
3,5 PAH ANALYSES
The results of PAH analyses conducted on the XAD-2 and particulate extracts by an
independent laboratory are shown in Figures 3-9 and 3-10. These analyses confirm the emission of
significant quantities of PAHs from the fluff combustion process. These analyses detected additional
PAH compounds that were not detected in the GC/MS full-scan analyses, perhaps because of the
presence of interfering compounds or the lesser sensitivity of full-scan methods as compared to
selected ion monitoring methods. All 16 PAHs analyzed for were detected in at least some of the
combustion experiments and most were detected in all three experiments. PAH concentrations in field,
laboratory, and hut blanks were generally non-detectable and were, in no case, greater than 10 percent
of the observed sample concentrations.
The vapor/particle distribution evidenced in these results shows the expected preponderance of
lighter species in the vapor phase. For the analytes measured both in this analysis and in the general
organic analyses, the level of agreement was encouraging. The estimated emissions for naphthalene
agree quite closely in all three tests—within 30 percent relative percent difference—(see Table 3-6 and
Figure 3-9) as would be expected because a compound specific MS response factor was used for
napthalene in the general organic analyses. There was less agreement for the other analytes reported
in both data sets (acenaplhylene, phenanthrene, fluoranthene, and pyrene). This is not surprising, since
the analytical methods differed greatly (GC/MS with selected ion monitoring vs GC/F1D calibrated as
TCO), However, a comparison of the values (see Table 3-6 and Figures 3-9,10) obtained in the two
separate methods would suggest that quantification accurate to within one order of magnitude (the
stated goal of the project) was obtained.
43
-------
I
CO
H
?
d Naphthalene
*Tl
gj Acenaphthylene
1-3 Acenaphthene
g Fluorene
Phenanthrene
c/s
I? S n Anthracene
£ O o
In o § Fluoranthene
a
US
vb
S3
§* r3 S
Pyrene
S 9 S
"I j Benzo(a)anthracene
•o R3
Chrysene
Benzo(b&k)fluoranthene
Benzo(a)pyrene
Indeno( 1,2,3-cd)pyrene
Dibenzo(a,h)anthracene
Benzo(g,h,i)pyerylene
Estimated emissions (g/kg)
OOOOOOOOO —' ——4
o'-'iou^intD^iDiD-'^Nu
I 1 1 1—1 1 1 1 1 1 1 r-
-L
-------
Jk
t/l
o
u»
o
!"0
5r
o
•o
i.
I
¦a
cs
I
Vi
H
Estimated emissions (g/kg)
C/5
w
n
o
S5
O
8
TJ a
pi C
oo
H
o
o o
s s
o o
2 8
o o
s s
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene —bawwwm
Fluoranthene
Pyrene
T
T
T
^T*T*I^*T*T*T^C^T*TOT«T*TOT^T3»rO&T*TCO?A7*T«T»T»T«T*T*T«T^T*T«T*T*3S^%
w^v*v«v:v*v:v»VAv
-------
3.6 PCDD/PCDF DATA
Separate samples were also collected specifically to characterize PCDD/PCDF emissions.
Separate analyses were performed on the XAD-2 and particulate filter samples. The results are
summarized in Figures 3-11 through 3-13. It is important to point out that the analyses performed do
not determine specific isomers, and only present the total mass for each congener group. Therefore, it
is not possible to determine the toxicity of these samples because the 2,3,7,8-substituted isomers (with
the exception of OCDD/OCDF) were not positively confirmed. However, these more toxic isomers
may indeed be present.
Qualitatively, the resulting emissions favored the formation of the less-substituted chlorinated
dibenzofurans. The tetrachloro and pentachloro dibenzofurans (TCDF/PeCDF) were roughly an order
of magnitude greater in concentration than the dioxin homologues. Generally, an order of magnitude
higher concentration (on an emission factor basis) was found in the particulate than in the vapor phase.
No field or hut blank particulate dioxin sample contained greater than 12 percent of the concentration
observed in the corresponding test samples. The vapor phase data and congener distribution should be
viewed with suspicion, however, because the concentrations in the hut blank sample were similar to
the concentrations in the actual test samples. The fact that the concentrations observed in the field
blank sample were non-detectable may indicate vapor phase dioxins and furans formed in earlier
experiments desorbing from the walls of the test facility rather than sample contamination.
The congener profiles observed are similar to those observed from soil samples collected from
scrap automobile incineration sites in the Netherlands.22 The congener profile, although not typical of
municipal waste combustion (MWC) fly ash, is similar to some MWC processes, * On a mass/mass
basis (ng/g), the concentration of PCDD/PCDF material on the particulate is at least an order of
magnitude greater than MWC fly ash PCDD/PCDF concentrations.25"26
46
-------
ALL ISOMERS WITHIN CONGENERS SUMMED
0.00015
0.00014
0.00013
0.00012
0.00011
0.0001
0.00009
0.00008
0,00007
0.00006
0.00005
0.00004
0.00003
0.00002
0.00001
0
TCDDs
FIRST FLUFF TEST
n
-4-
T
PeCDDs 1 HxCDDs ' HpCDDs ' OCDDs
TCDFs PeCDFs HxCDFs HpCDFs OCDFs
SECOND FLUFF TEST
THIRD FLUFF TEST I I AVERAGE
Figure 3-11, Estimated emissions for vapor phase PCDDs/PCDFs.
47
-------
ALL ISOMERS WITHIN CONGENERS SUMMED
0.0032
0.003
0.0028
0.0026
0.0024
0.0022
0.002
0.0018
0.0016
0.0014
0.0012
0.001
0.0008
0.0006
0.0004
0.0002
0
TCDDs
PeCDDs
IlxCDDs
[
Si Ini
TCDFs
PeCDDs I IlxCDDs I HpCDDs OCDDs
PeCDFs HxCDFs HpCDFi OCDFs
FIRST FLUFF TEST
SECOND FLUFF TEST
THIRD FLUFF TEST I I AVERAGE
Figure 3-12. Estimated emissions for particulate-bound PCDDs/PCDFs.
48
-------
W§;
TCDDs
PeCDDs I HxCDDs
TCDFs PeCDFs
CXm
HpCDDs OCDDs
HxCDFs HpCDFs OCDFs
FUST FLUFF TEST SECOND FLUFF TEST
THIRD FLUFF TEST I I AVERAGE
Figure 3-13. Total PCDD/PCDF estimated emissions.
49
-------
The majority of PCDD/PCDF material was found on the particulate filters. Nearly 30 times
more total PCDD/PCDF material was contained on this fraction relative to the XAD-2 fraction.
Similar congener profiles were observed in the particulate and XAD-2 fractions.
3.7 METAL AEROSOLS DATA
A separate particulate collection system was used to characterize potential metal aerosol
emissions. Eleven metals, potentially present in fluff and of environmental concern, were targeted for
analysis. Table 3-7 summarizes the estimated emission rates for these targeted metals. Concentrations
of analyzed samples were often at, or near, analytical method detection levels. Therefore, many of the
estimated emissions are presented as less than levels. Of the 11 metals targeted, cadmium, copper,
lead, and zinc where found in concentrations significantly greater (2-3 orders of magnitude) than blank
and/or method detection levels. Figure 3-14 graphically represents the relative contributions of these
metals. Copper is present in relatively large concentrations and copper compounds have been
investigated as catalysts in the formation of PCDDs/PCDFs in municipal waste incineration
T7
processes.
3.8 PARTICULATE DATA
Particulate matter was collected by several different sampling systems—the semivolatile
organics systems, the metal aerosol system, and the PM10 ambient sampler. The estimated particulate
matter emissions for these systems are presented in Table 3-8. For the sampling systems operated in
the sample shed and connected to the sampling manifold, excellent agreement exists between sampling
systems when comparing the particulate emissions for a given test day. Greater variation exists when
comparing the particulate emissions of different test days. Possible reasons for this variation include
significant differences in the composition of the fluff, or more likely, isokinetic sampling variation
between test conditions. The configuration of the test facility at the time of testing made isokinetic
sampling impractical and unnecessary, given the scoping nature of the studies performed. The bias
50
-------
ONLY DETECTED METALS SHOWN
0.8
A»,V
CADMIUM COPPER LEAD ZINC
FIRST FLUFF TEST SECOND FLUFF TEST M THIRD FLUFF TEST I I AVERAGE
Figure 3-14. Estimated emissions for selected metals.
51
-------
realized by excessive variance from isokinetic sampling rates would tend to underestimate the
particulate and particulate bound organic emissions as opposed to inflating them.
TABLE 3-7. METALS EMISSION DATA
Particulate
Bound
Cadmium
(g/kg)
Particulate
Bound
Copper
(g/kg)
Particulate
Bound
Lead
(g/kg)
Particulate
Bound Zinc
(g/kg)
Total
Metals
(g/kg)
Particulate
Bound
Silver
<8*8)
Particulate Bound
Aluminum fg/kg)
First
0.0249
0.2073
0.5390
0.4561
1.2272
<0.0002
*0.0104
Second
0.0344
0.6878
0.6878
0.6349
2.0448
<0.0003
*0.0079
Third
0.0312
0.3686
0.6872
0.4248
1.5118
*0.0037
*0.0187
Average
0.0302
0.4212
0.6380
0.5052
1.5946
**0.0014
**0.0123
Parti culate
Bound
Arsenic
(g/kg)
Parti culate
Bound
Barium
(g/kg)
Particulate
Bound
Chromium
(g/kg)
Particulate
Bound
Magnesium
(g/kg)
Part culate
Bound
Selenium
(g/kg)
llllllSHllllll
First
<0.0021
*0.0021
*0.0041
*0.0041
<0.0021
• i'i'w i'lviv i" '• ¦!
Second
<0.0026
<0.0026
*0.0053
*0.0026
<0.0026
Third
<0.0062
<0.0062
*0.0750
<0.0062
<0.0062
Average
**0.0037
**0.0037
**0.0281
**0.0043
**0.0037
Ifiiilill
Note: < = analytical nondetects listed at a value derived from the analytical detection limit
* = method nondetects, values derived from analytical results that do not significantly exceed corresponding blank
values, but do exceed the analytical detection limit
** = values derived from averaging values including one or more analytical or method nondetects
TABLE 3-8. SUMMARY OF ESTIMATED PARTICULATE EMISSIONS
Test
General
Organic Train
Total
Particulate
(g/kg)
Dioxin Train
Total
Particulate
(g/kg)
Metals Train
Total
Particulate
(g/kg)
Average of 3
Total
Particulate
Measurements
(g/kg)
PM10 Sampler
Particulate
(g/kg)
First
91.2
85.8
81.5
86.2
66.0
Second
116.2
115.7
89.9
107.3
FRE
Third
183.4
188.6
174.5
182.2
41.1
Average of 3
days
130.3
130.0
115.3
125.2
53.6
Note: FRE = Flow rate error, the flow rate on this run for the PM10 sampler differed by more than 20
percent from the design flow rate.
52
-------
A separate medium vol 113 L/min (4 cfm) PM10 ambient sampler was operated within the
bum hut to characterize particulate matter 10 //m in diameter and less. Because this sampling did not
occur in a duct or stack, there is no concern with regards to isokinecity for these samples. The
particulate matter emissions for this system are also presented in Table 3-8. For qualitative purposes,
a comparison of the PM10 to total particulate has been made based on total averaged values. The
PM10 comprises roughly 43 percent of the total particulate matter collected.
Levels of particulate collected in the hut, field, and laboratory blank samples collected were an
order of magnitude less then the lowest values observed in actual test samples for the general organic,
dioxin, and PM10 trains. The metals train hut blank particulate concentration is 20 times less than any
actual test sample and the field blank was 3.7 times less than the lowest test sample.
3.9 EMISSION DATA SUMMARY
The tests and subsequent analyses performed during this study were selected to characterize, as
broadly as possible, the diverse emissions resulting from the open combustion of fluff. A considerable
body of data were generated as a result Because some approaches to understanding these data can be
overwhelming and time consuming, the individual data sets have been summarized to illustrate their
respective relative contributions to total mass emissions.
To assess the overall organic emissions, the volatile and semivolatile organics emission data
were summarized. The total organics emitted, volatile (volatile determined using the FID response
factor for toluene and totaling all compounds up to but not including the retention time of toluene),
vapor-phase semivolatile, and particulate-bound semivolatile, averaged more than 200 g/kg fluff
combusted. The actual mass contribution from each fraction is summarized in Table 3-9. Figure 3-15
graphically presents the relative mass contributions of various sample fractions to the total organic
mass emissions. Figure 3-16 depicts the relative mass distribution based on the boiling point of
53
-------
VOST (23.2%)
PARTICULATE (38.5%)
XAD-2 (38.3%)
Figure 3-15. Distribution of organics by sampling method.
54
-------
VOLATILES: <110 C (23.2%)
TCO: 100-300 C (32.3%)
GRAY: >300 C (44.6%)
Figure 3-16. Distribution of organics by boiling point.
55
-------
organic emissions. The actual mass of organics identified by individual compounds or compound
classes have been previously addressed in the respective organics characterization sections.
TABLE 3-9. MASS BALANCE SUMMATION (g/kg)
Day 1
Day 2
Day 3
Average
Volatile Organics
28.58
56.11
51.68
45.46
Total Chromatographable
Organics on XAD
56.99
50.05
90.72
65.92
Gravimetric Organics on XAD
6.68
10.12
23.58
13.46
Particulate (all trains average)
86.20
107.26
182.19
125.22
CO as C
67.93
71.54
72.09
70.52
C02 as C
915.71
746.30
771.93
811.31
NO as N
NAV
2.54
2.43
2.49
Sum
1162.10
1043.92
1194.62
1134.37
Note: NAV = Not available
The particulate matter collected was also quantitatively characterized to the greatest extent
possible. On average, nearly 60 percent of the total particulate mass was found to be
dichloromethane-extractable (as quantified by the TCO and GRAV methods). Similarly, the metals
content of the particulate matter comprised nearly 1.2 percent. The remaining particulate matter may
be comprised of organic compounds not extracted by dichloromethane, e.g., strongly polar organics,
inorganics which were not analyzed for, or carbonaceous matter. Figure 3-17 depicts the relative
distribution of identified emissions present in particulate matter.
Finally, as a general evaluation of the estimated mass emissions characterizations, a total mass
balance was performed. The diversity of the measurements performed during testing was felt to
sufficiently enable the determination of total mass emissions. The majority of the detailed emissions
analyses characterized the PICs. In addition, the continuous emission monitoring data collected could
56
-------
TCO- ORGANICS (0.7%) METALS (1.2%)
UNKNOWN (38.0%)
GRAY - ORGANICS (60.1%)
Figure 3-17. Particle mass distribution.
57
-------
be used to characterize the gaseous products of combustion. The total mass of emissions would
include the total vapor-phase organics (volatile and semivolatile), particulate (includes organ ics and
metal contributions), C02 (as carbon), CO (as carbon), and NO (as nitrogen). For the purposes of this
analysis, total volatiles were estimated by totaling the area of all compounds eluting before toluene and
applying the toluene FID response factor. Semivolatile organics were estimated as the total of GRAV
and TCO organics captured on XAD-2. The total mass emissions would also likely include HC1,
which was not measured. Not accounting for this mass would lead to an underestimation of total mass
emissions. Conversely, a number of oxygenated organic compounds were identified in the various
organic sample fractions. Similarly, the metals aerosols captured on particulate filters were likely
present as metal oxides. The source of the oxygen in these compounds would likely be ambient
molecular oxygen as opposed to fluff-bound oxygen. Because it was not corrected for, the presence of
oxygen in these identified organic compounds could lead to overestimating total mass emissions. The
mass emissions of C02, CO, and NO were corrected for oxygen based on the assumption that all of
this oxygen was ambient molecular oxygen.
Nitrogen, as NO, was included in the mas balance because several of the plastics present in
fluff contain nitrogenated compounds. It is very possible that the NO measured resulted from
conversion of atmospheric N2 during combustion. However, the contribution of N to the mass balance
is small, and its inclusion or deletion from the mass balance would have minimal impact on the quality
of the mass balance.
The actual mass balances, based on individual and overall test average emission rate values,
are presented in Table 3-9. The result of the mass balances reveals that the total estimated mass
emissions were roughly 15 percent greater then the measured weight loss of the solid fluff. Given the
semiquantitative nature of the tests performed and the confounding factors in the mass balance
previously noted, this result appears to be well within reason.
58
-------
SECTION 4
SUMMARY AND CONCLUSIONS
This laboratory study successfully achieved its goals of characterizing the broad spectrum of
emissions from open fluff combustion and of generating emission factors accurate to within one order
of magnitude for these emissions. As previously discussed, the results from this study were used to
perform a mass balance, the quality of which indicates that it is likely that the major emissions from
this source have been detected by the sampling and analytical methods chosen. A large number of
PICs have been identified, many of which are previously known products of combustion of various
individual plastics. However, not all of the organic compounds present in the sample have been
identified and quantified. The chromatograms obtained from these emissions samples were often
highly complex, and all the chromatographably resolved peaks were not able to be identified especially
in the semivolatile and particulate-bound organic fractions. The use of compound specific analytical
methods, such as were used here for PCDDs/FCDFs, should reveal the presence of low but potentially
significant concentrations of analytes not detected in this study. For example, given the prevalence of
brominated flame retardants in commercial plastics and the identification of PCDDs and PCDFs in our
•jo "?Q
sample, the production of the brominated analogues of PCDDs and PCDFs seems likely. ' The use
of bioassay directed fractionation in future studies of emissions from similar processes is suggested to
determine if the unidentified organics are of any human health concern and to focus identification and
quantification efforts.
-------
The relationship of this experiment to actual uncontrolled fluff combustion has not been
established. It is difficult to quantitatively state how accurately these combustion experiments,
performed on tens of pounds of material for hours, simulate a combustion process which, in real life,
may involve millions of pounds of fluff burning for months. There are, however, encouraging
similarities between our data set and the small existing data sets collected at actual fluff fires. The
major compounds measured using Method TO-14 at the Montvale, VA fluff fire were (in descending
order of concentration in the "smoke" sample) benzene, chlorobenzene, ethylbenzene, toluene, styrene,
an
and xylenes. These were among the compounds found in the highest concentration in the volatile
organic portion of our work (all of these compounds had estimated emissions greater then 1 g/kg as an
average across all our samples). In addition, the three metals detected in the particulate matter in the
highest concentration in our experiment (copper, lead, and zinc) are the three metals found in highest
concentration in the "Ash Composite" sample from the Montvale fire and in the PM10 particulate
matter collected at a school near that fire. The fourth metal we detected in the particulate matter,
cadmium, was the sixth most concentrated of 10 metals tested for in the Montvale sample.8,30 In die
absence of air dispersion modeling of the Montvale site, quantitative comparisons cannot be made but
the qualitative agreement of our data to one example of an open fluff fire appears good. Similarities
also exist between our data set and data on metals and semivolatiles in the debris pile left behind from
the Washoe County, NV fluff fire except that wipe tests for dioxins and furans at the Washoe County
site proved negative.9 This result might be explained by differences in the sensitivities of die
analytical methods applied, such as differences in sample size. Reasonable agreement between
simulations in our facility and actual open burning results has also been observed in a previous study
of tire combustion.31
In the absence of a general or site specific exposure and risk assessment, the threat posed from
the emissions measured here can not be quantitatively predicted. However, the combination of the
60
-------
emissions measured here for compounds of human health concern and Ihe substantial quantities of
combustible materials present in fluff landfills certainly merits a further evaluation of the risk posed by
open fluff combustion. The estimated emissions presented here could serve as an important data
source for such an assessment, (The estimated emissions are presented in terms of mass of pollutant
per mass of fluff consumed by combustion, not per mass exposed to combustion conditions.) In the
interim, this document should help provide some further basis for informed decision making by
personnel faced with controlling fluff fires. In particular, our data on volatiles emissions seems to
confirm the suggestion previously made that the partial extinguishment of open fluff combustion
s
processes may actually increase the total emission of pollutants.
61
-------
SECTION 5
REFERENCES
1. "Recycling Scrap Iron and Steel," Pamphlet published by the Institute of Scrap Recycling,
1990, Industries, Inc., Washington DC, 1990.
2. Valdez, E.G. et al., "Recovering Folyurethane Foam and Other Plastics from Auto-shredder
Reject," Report 8091, U.S. Bureau of Mines, Washington, DC, 1975.
3. Dean, K.C. et ai., "Bureau of Mines Research on Recycling Scrapped Automobiles,'' Bulletin
684, U.S. Bureau of Mines, Washington, DC, 1985.
4. Rousseau, M. and A. Melin, "The Processing of Non-magnetic Fractions from Shredded
Automobile Scrap: A Review," Resources. Conservation and Recycling. 2,139-159, 1989.
5. Mahoney, L.R. et al., "Hydrolysis of Polyurethane Foam Waste," Environmental Science and
Technology. 8(2):135-139, 1974.
6. Wrigley, A, "Automotive Use of Plastics Expands," American Metal Market 94(163),1, 1986.
7. Henderson, T.L., Commonwealth of Virginia, Department of Air Pollution Control, Lynchburg,
VA, personal communications, 199Z
8. Personal communication from Central Virginia Regional Office, Department of Air Pollution
Control, Lynchburg, VA, of records concerning the 1989 Montvale, VA, fluff fire to J. Ryan,
1992.
9. Personal communication from Chris Ralph, Washoe County, NV, District Health Department
to P.M. Lemieux, June 18, 1990.
10. Method 0030 in Test Methods for Evaluating Solid Waste. Volume II: Field Manual
Phvsical/Chemical Methods fThird Edition), EPA-SW-846 (NTIS PB88-239223), U.S.
Environmental Protection Agency, Washington, DC, September 1986.
62
-------
11. Method 5040 in Test Methods for Evaluating Solid Waste, Volume IB: Laboratory Manual
Physical/Chemical Methods (Third Edition), EPA-SW-846 (NTIS PB88-239223), U.S.
Environmental Protection Agency, Washington, DC, September 1986.
12. Lentzen, D.E. et al., IERL-RTP Procedures Manual: Level 1 Environmental Assessment
(Second Edition), EPA-600/7-78-201 (NTIS PB293795), U.S. Environmental Protection
Agency, Research Triangle Park, NC, October 1978.
13. Wasson, S.J. and J.T. Keever, "Validation Analysis of Polynuclear Aromatic Hydrocarbons";
Technical Report No. 3967/56B-21F, May 20, 1992; Research Triangle Institute, Research
Triangle Park, North Carolina.
14. Method 8280 in Test Methods for Evaluating Solid Waste. Volume IB: Laboratory Manual
Physical/Chemical Methods (Third Edition), EPA-SW-846 (NTIS PB88-239223), U.S.
Environmental Protection Agency, Washington, DC, September 1986.
15. Method 23 in Title 40 Code of Federal Regulations Part 60. Appendix A, U.S. Government
Printing Office, Washington, DC, 1991.
16. Method 200.7 in Methods for the Determination of Metals in Environmental Samples. EPA-
600/4-91/010 (NTIS PB91-231495), U.S. Environmental Protection Agency, Environmental
Monitoring Systems Laboratory, Cincinnati, OH, June 1991.
17. McFarland, AR. and CA Cortiz, "A 10 /an Cutpoint Ambient Aerosol Sampling Inlet,"
Atmospheric Environment 16: 2959-2965, 1982.
18. Levin, B.C., "A Summary of the NBS Literature Reviews on the Chemical Nature and
Toxicity of the Pyrolysis and Combustion Products from Seven Plastics: Acrylonitrile-
Butadiene-Styrenes (ABS), Nylons, Polyesters, Polyethylenes, Polystyrenes, Polyvinyl
Chlorides, and Rigid Polyurethane Foams," Fire and Materials. 11,143-157, 1987.
19. Paabo, M. and B.C. Levin, "A Literature Review of the Chemical Nature and Toxicity of the
Decomposition Products of Polyethylenes," Fire and Materials, 11, 55-70, 1987.
20. Huggett, C. and B.C. Levin, "Toxicity of the Pyrolysis and Combustion Products of
Polyvinylchlorides: A Literature Assessment," Fire and Materials, 11, 131-142, 1987.
21. Mitera, J. and J. Michal, "The Combustion Products of Polymeric Materials," Fire and
Materials. 9(3),111-116, 1985.
22. van Wijnen, J.H., et al., "Soil Contamination with PCDDs and PCDFs of Small (Illegal) Scrap
Wire and Scrap Car Incineration Sites," submitted for publication in Chemosphere.
23. Ballschmiter, K., et al., "Automobile Exhausts Versus Municipal Waste Incineration as Sources
of the Polychloro-dibenzodioxins (PCDD) and -furans (PCDF) Found in the Environment,"
Chemosphere. 15(7), 901-915, 1986.
63
-------
24. Smith, R.M., ct al., Chlorinated Dioxins and Dibenzofurans in Perspective; C. Rappe et al.,
eds. Lewis Publishers Inc., Chelsea, MI, pp 93-108, 1986.
25. Czuczwa, J.M. and R.A. Hites, "Environmental Fate of Combustion-Generated Folychlorinated
Dioxins and Furans," Environmental Science and Technology. 18, 444-450, 1984.
26. Bumb, R.R., et al., "Trace Chemistries of Fire: A Source of Chlorinated Dioxins," Science,
210(4468), 385-390, 1980.
27. Gullett, B.K., et al., "The Effect of Metal Catalysis on the Formation of Polychlorinated
Dibenzo-p-dioxin and Polychlorinated Dibenzofuran Precursors," Chemosphere, 20 (10-12),
1945-1952, 1990.
28. Hans-Rudolf Buser: Polybrominated Dibenzofurans and Dibenzo-p-dioxins: Thermal Reaction
Products of Polybrominated Diphenyl Ether Flame Retardants," Environmental Science and
Technology. 20, 404-8, 1986.
29. Dumler, R., H. Thoma, D. Lenoir, and O. Hutzinger, "Thermal Formation for Polybrominated
Dibcnzodioxins (PBDD) and Dibenzofurans (PBDF) From Bromine Containing Flame
Retarded Polymers: A Survey," Chemosphere. 12, 2023-31, 1989.
30. Holmes, C.E., Commonwealth of Virginia, Department of Air Pollution Control personal
communication to C.H. Darvin (U.S. Environmental Protection Agency/Air and Energy
Engineering Research Laboratory), 1989.
31. Ryan, J.V., "Characterization of Emissions from the Simulated Open Burning of Scrap Tires,"
EPA-600/2-89-054 (NTIS FB90-126004), October 1989.
64
-------
APPENDIX A
QUALITY CONTROL EVALUATION REPORT
This task was conducted under the guidance of an EPA-approved Quality Assurance Project
Plan. This plan was used to establish data quality objectives suitable for this study. The quality
control measures employed during this study were performed to ensure that the data collected would
be suitable to collect, identify, and semi-quantitate air emissions resulting from the simulated open
burning of automobile fluff. The primary project goal was to obtain qualitative information as to the
types and identities of both inorganic and organic combustion by-products. The secondary goal was to
provide estimate emissions accurate within an order of magnitude (factor of ten) for selected
compounds and compound types identified.
Table A-l presents the data quality indicator (DQI) summaries for accuracy, precision, and
completeness achieved during testing along with the planned DQI goals for each respective
measurement or analysis performed. In general, the intended DQI goals were achieved. In several
instances, however, targeted DQI goals were not achieved.
Included in Table A-l are the DQI summaries for the continuous emission monitoring systems.
The performance indicators demonstrate that the systems performed well within project goals.
However, operational difficulties were encountered with the total hydrocarbon analyzer and the nitric
oxide analyzer. On two test days, the post-test span check for the THC analyzer exceeded established
accuracy limits (Day 1: 34 percent bias, Day 2: 23 percent bias). Similarly, for the NO analyzer, the
-------
Day 1 post-test span check exceeded established accuracy limits (40 percent bias). Although the data
collected for those test days were not validated by the post-test QC checks, the data are, however,
presented and considered usable for characterization purposes.
TABLE A-l. DATA QUALITY SUMMARY
Accuracy (c.
% Bias)
Precision (% RSD)
Completeness
(% Valid data)
Measurement
Goal
No. of
points
Achieved
(Avg.)
Goal
Achieved
Goal
Achieved
o2
15
4
5.15
10
3.21
> 75
100
CO
15
4
2.39
10
1.57
> 75
100
co2
15
4
1.96
10
1.28
> 75
100
THC
15
4
15.20
10
14.21
> 75
50
NO
15
4
13.09
10
14.07
> 75
75
Volatile organics
50
16
72.7 f
25
15.9
> 75
82
Semivolatile organics
50
* *
* *
25
21.5
> 75
100
TCO
20
1
3.2
15
3.3
100
100
GRAV
20
1
6.2
15
0.4
100
100
Legend:
* Relative percent difference
** See Table A-2
t Expressed as percent recovery
Table A-l includes DQI summaries for volatile organic compound characterizations. The
VOST tube pairs used to collect volatile organics were spiked with deuterated benzene before
sampling to assess method performance. The recovery of the Dfi-benzene is used to assess method
accuracy. The average Dg-bcnzene recovery values are presented in Table A. With the exception of
several VOST samples (Test 2 Field Blank: 10 percent, Laboratory Blank: 38 percent, Hut Blank No.
2: 0 percent recovery), recovery values were 40-120 percent.
Once the GC/MS/FID analytical system was calibrated and the system's relative response
factors determined, continuing calibration QA/QC standard solutions were analyzed at the beginning
66
-------
and end of each analytical period. These QC checks were used to determine analytical method
precision. Results of these QC checks are also included in Table A-l.
Table A-l includes DQI summaries for semivolatile organic compound characterizations. The
QA/QC approach to the semivolatile organic analyses was similar in nature to that of the volatile
organic analyses. Once the GC/FID (TCO) and GC/MS systems were calibrated and system responses
established, continuing calibration check solutions were analyzed at the beginning and end of each
analytical period to evaluate system performance. The results of these QA/QC checks are also
included in Table A-l.
Additional standards, containing compounds identified in test samples, were also prepared and
analyzed. These standards were used to confirm tentatively identified compounds as well as to
evaluate the generalized quantitative approach.
A PAH Performance Evaluation Audit (PEA) was provided by an independent QA laboratory.
The sample was analyzed using the general TCO GC/FID semivolatile organic analytical method. The
results of this analysis are presented in Table A-2. The results indicate that, for the PAHs identified,
accuracy, expressed as percent recovery, ranged from 55-105 percent for all compounds quantified.
Three of the compounds present in the performance evaluation sample were not detected.
These 3 PAHs were not detected as a result of chromatographic (temperature) limitations as opposed
to detector sensitivity limitations.
Although individual analytical accuracy values for each compound identified were not
determined, it is possible to estimate the accuracy of these measurements. Many of the compounds
identified were quantified using both GC/MS and GC/FID system responses. The quantitative
agreement between these analytical approaches were generally within a factor of two. The analytical
accuracy is also supported both by the PAH PEA sample as well the TCO alkane mix accuracy
performance checks which were found to exhibit less than 15 percent analytical bias. The estimated
67
-------
analytical accuracy of a factor of two is deemed more than the adequate to determine emission factors
for these identified compounds that are accurate to within project goals of an order of magnitude.
The recovery of selected semivolatile organic compounds extracted from the XAD-2 sorbent
resin was also evaluated. A FAH-containing solution, a TCO/GRAV alkane-containing solution, as
well as a base/neutral/acid solution were spiked onto separate XAD-2 sampling canisters and then
soxhlet extracted. The recovery values for these performance evaluation checks are presented in Table
A-3.
A portion of the general semivolatile organic extracts were analyzed specifically for 16 FAHs.
Originally, these analyses were to be performed by a contracted laboratory using high-performance
liquid chromatography (HFLC) methodologies. These methods are desirable because of their
specificity to PAHs and the reduced susceptibility to sample interferences. However, the intended
laboratory was unavailable for these analyses. A separate contracted laboratory was identified and this
laboratory used a PAH-specific GC/MS analytical method.
The identification of organic compounds collected during the combustion of fluff required a
generalized approach which was necessary for a numher of reasons, foremost of which was the
configuration of available analytical instrumentation and the emphasis of qualitative data over
quantitative data. Identification of unknowns relied primarily on mass spectral information. Acquired
spectra were compared to spectra contained in a computer database associated with the GC/MS system
using a probability-based spectral matching program. The quality of the resulting match was evaluated
manually, and coupled with investigator interpretation and other information such as likelihood of
presence and compound boiling point, and retention time, a tentative identification assigned. Where
possible, specially prepared qualitative standards, containing tentatively identified compounds, were
analyzed under identical analytical conditions. Low resolution mass spectrometry has inherent
limitations when applied to the identification of unknowns from spectral information. While it is often
68
-------
possible to determine the molecular formula of organic unknowns, determining functional substitution
groups and specific isomers often proves difficult Similarly, the relatively low ionization potential of
alkanes coupled with electron ionization (EI) makes determination of molecular ions difficult.
Therefore, many of the compounds tentatively identified in this study are unable to be presented
further than the molecular formula and organic class. An added emphasis was placed on using the
aforementioned qualitative standards. Project resources limited further confirmatory analyses.
The metals analyses were performed by a contracted commercial laboratory. The QA/QC
measures described in the respective referenced procedures were adhered to and achieved. Because
many of the targeted metals were found at less than detectable levels, emission factors were also
presented as less than levels based on method detection levels.
As stated earlier, PCDDs and PCDFs (tetra - octa congeners) were analyzed using a
combination of techniques found in SW-846 Method 8280 and 40 CFR Part 60, Appendix A Method
23. The samples were analyzed by high resolution gas ehromatography/low resolution mass
spectrometry (HRGC/LRMS). This procedure serves to confirm the presence or absence of PCDDs
and PCDFs as well as quantify the number of confirmed isomers found in each congener class. This
analytical method follows the QA/QC guidelines listed in the SW-846 method. In addition, it uses
isotopically labeled PCDD/PCDF homologues for each congener (with the exception of
octachlorodibenzofuran). The procedure differs in that specific isomers, including the 2,3,7,8
substituted isomers, are not confirmed. Method performance is evaluated by the recovery of the
isotopically labelled internal standards. The actual recovery values for each congener respective to
each sample are too numerous to be included here. Table A-4 provides a summary of recovery values
for each congener. In several particulate phase samples, the recovery values for several congeners
were less than the targeted 40 percent The low recoveries were not found to significantly affect data
quality and were, therefore, reported.
69
-------
Facility, field, and laboratory blanks were performed routinely for many of the measurements
performed during this study. The results of the field and laboratory blank analyses are described in
the respective data presentation section of this report The test data have not been corrected for the
blank values. Where possible, the blank values have been presented along with the actual test data.
In summary, the OA project objectives set forth have been adequately obtained and the data
collected from this study are sufficient to meet project objectives.
70
-------
TABLE A-2. RESULTS OF PAH PERFORMANCE EVALUATION AUDIT
COMPOUND
RECOVERY (%)
Naphthalene
92.2
Acenaphthylene
88.7
Acenaphthene
101
Fluorene
91
Phenanthrene
89.6
Anthracene
87.6
Fluoranthene
88.2
Pyrene
1026
Chrysene
77.1
Benzo[a]anthracene
70.1
Benzo[b]fluoranthene
62.9
Benzofk]Ouoranthene
55.2
Benzo[a]pyrene
55.8
Benzo[ghi]perylene
ND
Dibenzo[a,h] anthracene
ND
Indeno[l,2,3-cd]pyrene
ND
Note: ND = Not detected
71
-------
TABLE A-3. RECOVERIES FROM SPIKED XAD-2 SAMPLES
COMPOUND
RECOVERY (%)
Chlorophenol
Dichlorobenzene
N-Nitroso N-propylamine
2,4- Dinitrotolune
Nitrophenol
93.9
88.3
190.6
106.8
101.7
TCO (CIO, C12, C14, n-alkanes)
GRAV (C20 n-alkane)
103.6
96.3
Naphthalene
Acenaphthylene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo[a]anthracene
Chryscne
Bcnzo[b+k]fluoranthene
Benzo[a]pyrenc
Indeno [ 1,2,3-cd]pyrene
Dibenzo[a,h]anthraeene
Benzo [g,h, i ]pery lene
80.1
81.4
91.8
103.8
87.5
ND
209.3
125.0
127.3
93.6
87.1
150.0
170.4
96.5
NOTE: Data arranged so that compounds spiked into each individual XAD-2 sample are separated by
blanks.
72
-------
TABLE A-4. RECOVERIES OF ISOTOPICALLY LABELED PCDD/PCDF INTERNAL STANDARDS
Vapor phase samples (XAD), No, of samples and blanks in each recovery category
RECOVERY TCDD
TCDF
PCDD
PCDF
HxCDD
HxCDF
HpCDD
HpCDF
OCDD
OCDF
40-120% (ACCEPTABLE) 5
5
5
5
5
5
5
5
5
5
<40 0
0
0
0
0
0
0
0
0
0
> 120% 0
0
0
0
0
0
0
0
0
0
Particulate phase samples (filter), No. of samples and blanks in each recovery category
RECOVERY TCDD
TCDF
PCDD
PCDF
HxCDD
HxCDF
HpCDD
HpCDF
OCDD
OCDF
40-120% (ACCEPTABLE) 5
4
4
5
4
5
4
4
1
1
<40 0
1
0
0
1
0
1
1
4
4
> 120% 0
0
1
0
0
0
0
0
0
0
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before complet'
T, REPORT NO. 2.
EPA- 600 /R-93-044
3- PB93-172914
4, TITLE AND SUBTITLE
Characterization of Emissions from the Simulated
Open-burning of Non-metallic Automobile Shredder
Residue
5, REPORT DATE
March 1993
6, PERFORMING ORGANIZATION CODE
7. AUTHOH(S)
Jeffrey V. Ryan and Christopher C. Lutes
8. PERFORMING ORGANIZATION REPORT NO.
9, PERFORMING ORGANIZATION NAME AND ADDRESS
Acurex Environmental Corporation
P. O. Box 13109
Research Triangle Park, North Carolina 27709
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-DO-0141, Tasks 91-030
and 92-055
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 2/89 - 10/92
14. SPONSORING AGENCY CODE
EPA/600/13
is.supplementary notes j^eERL project officer is Paul M. Lemieux, Mail Drop 65, 919/
541-0962
is, abstract rep0rj- gives results of a study in which the open combustion of a non-
metallic waste product called "fluff" was simulated and the resulting emissions col-
lected and characterized to gain insight into the types and quantities of these air pol-
lutants. (NOTE: The reclamation process for recyclable ferrous and non-ferrous
metals from scrap automobiles generates fluff consisting of a combination of glass,
plastics, rubber, wood products, and electrical wiring. The waste product is often
stockpiled or landfilled. A number of the stockpiles have caught fire, resulting in the
emission of many air pollutants.) Samples were collected and analyzed for volatile
and semivolatile organics, particulate, and metal aerosols. Typical combustion pro-
cess gases--carbon dioxide, carbon monoxide, nitric oxide, oxygen, and unburned
hydrocarbons--were monitored continuously. The samples were analyzed using GC/
MS, GC./FID, gravimetric, and atomic emission methodologies to identify and quan-
tify the types of compounds present in the open combustion process emissions. The
resulting mass/volume concentrations were related to the measured net mass of ma-
terial consumed through combustion and known dilution air volume to derive an esti-
mate of overall emissions. Of 11 metal aerosols characterized, cadmium, copper,
lead, and zinc were found in significant quantities.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. cosati Field/Group
Pollution Nonmetallic Com-
Combustion pounds
Emission
Residues
Automobiles
Shredders
Pollution Control
Stationary Sources
Fluff
13B
2 IB 07.B
14 G
13 F
131, 07A
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21, NO. OF PAGES
79
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |