United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/SR-93/044 May 1993
&EPA Project Summary
Characterization of
Emissions from the Simulated
Open-Burning of Non-Metallic
Automobile Shredder Residue
Jeffrey V. Ryan and Christopher C. Lutes
The reclamation process for recy-
clable 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 prod-
uct is often stockpiled or landfilled. A
number of these stockpiles have caught
fire, resulting in the emission of nu-
merous 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 emis-
sions collected and characterized.
Samples were collected and analyzed
for volatile and semivolatile organics,
particulate, and metal aerosols. Typical
combustion process gases—carbon di-
oxide, carbon monoxide, nitric oxide,
oxygen, and unburned hydrocarbons—
were monitored continuously. The
samples were analyzed using gas chro-
matography (GC)/mass spectroscopy,
GC/flame ionization detection, gravimet-
ric, and atomic emission methodolo-
gies to identify and quantify the types
of compounds present in the open com-
bustion process emissions. The result-
ing mass/volume concentrations were
related to the measured net mass of
material consumed through combus-
tion and known dilution air volume to
derive an estimate of overall emissions.
Volatile and semivolatile organics char-
acterized included mono- and polyaro-
matic hydrocarbons, substituted
alkanes and alkenes, aldehydes, nitrites,
phenols, chlorinated aromatics, hetero-
cycles, and polychlorinated dibenzodi-
oxins and furans. Of the 11 metal aero-
sols characterized, cadmium, copper,
lead, and zinc were found in significant
quantities. The emission characteriza-
tions performed indicated that substan-
tial quantities of air pollutants are
emitted. For the organic pollutants
alone, the emission of more than 200
g/kg of fluff combusted was observed.
This Project Summary was devel-
oped by EPA's Air and Energy Engi-
neering Research Laboratory, Research
Triangle Park, NC, to announce key find-
ings of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
The reclamation process for ferrous and
non-ferrous metals from scrap automo-
biles generates a non-metallic waste prod-
uct called "fluff." For the most part, fluff
consists of a combination of plastics such
as polyethylene, polypropylene, acryloni-
trile-butadiene-styrene, polyurethane foam,
polyvinylchloride, rubber, glass, wood prod-
ucts, cloth, paper, dirt, and electrical wir-
ing. Conservatively, it can be estimated
that 9.1 x 108 kg (2 x 109 Ib) of fluff is
produced annually.
The automobile fluff is discarded at land-
fills or, more commonly, stockpiled on site.
At several automobile reclamation facili-
ties, these stockpiles have caught fire.
One such stockpile fire, in Montvale, VA,
burned for 38 days emitting unknown quan-
tities of potentially harmful air pollutants. It
was estimated that 13,000-16,000 bales
Printed on Recycled Paper
-------
of fluff, weighing approximately 1,360 kg
(3,000 Ib) each, were burned in the fire.
During the fire, several attempts were
made both to extinguish it and to acceler-
ate combustion. The Commonwealth of
Virginia's Department of Air Pollution Con-
trol contacted EPA's Control Technology
Center (CTC) requesting emissions data
on combustion of this material. Unfortu-
nately, data pertaining to the open burn-
ing 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 Environmen-
tal performed a study which identified and
quantified organic and inorganic emission
products produced during the simulated
open combustion of fluff. The study was
designed to roughly estimate emission
rates (within an order of magnitude) for
volatile arid semivolatile organics, particu-
late, and selected metal aerosols identi-
fied in combustion emissions. Emphasis
was placed on gaining a broad overview
of the pollutants produced.
Experimental
The project consisted of replicate tests
to collect and qualitatively and quantita-
tively characterize organic and inorganic
emissions resulting from the simulated
open combustion of actual automobile fluff
waste. Small quantities (9-11 kg, 20-25 Ib)
of fluff, obtained from an automobile rec-
lamation facility, were combusted in a test
facility designed for simulation of open
combustion conditions (see Figure 1). A
representative air sample from the burn
hut environment is delivered to a sam-
pling facility, adjacent to the burn hut.
The sampling facility contains most of
the associated sampling equipment: the
volatile organic sampling train (VOST) sys-
tem, the semivolatile organics/particulate
sample collection systems, and the par-
ticulate removal system for the continuous
emission monitors (CEMs).
Fixed combustion gases—carbon mon-
oxide (CO), carbon dioxide (CO2), nitric
oxide (NO), oxygen (02), and unburned
hydrocarbons (THC)—were measured con-
tinuously using on-line process analyzers.
Volatile organics were collected using the
VOST system. Semivolatile organics and
particulate were collected using a sample
system modified for use in this study. Two
separate semivolatile organic/particulate
collection systems were operated simulta-
neously during the test period. One sample
system was used to collect samples for
general semivolatile organic and particu-
Sample Duct
Fluff Combustion
Container
Air Inlet
Air Inlet
Weighing Platform
Figure 1. Diagram of burn hut.
late characterization, while the remaining
system was used to collect samples for
polychlorinated dibenzodioxin (PCDD) and
polychlorinated dibenzofuran (PCDF)
analyses. A separate particulate sampling
system was used to collect metal aero-
sols. A medium volume (4 cfm, 0.00189
m3s) ambient particulate sampler was used
to collect particulate 10-um in diameter
and less.
The VOST samples were analyzed by
gas chromatography/mass spectroscopy/
flame ionization detection (GC/MS/FID) on
a purge-and-trap thermal desorption sys-
tem. Compounds were identified using
multi-component calibration standard com-
parisons, mass spectral library searches,
and investigator interpretation. Identified
analytes were quantified using a combi-
nation of GC/MS and GC/FID system re-
sponses based on the characteristics of
the identified compound.
The semivolatile organics from the gen-
eral organics samples were retrieved from
the collection media by soxhlet extraction
using dichloromethane. The XAD-2 was
extracted separately from the particulate
fraction. Both the particulate extracts and
the XAD-2 extracts were analyzed indi-
vidually for total chromatographable or-
ganics (TCO)—(organic compounds with
boiling points of 100-300 °C)—and total
gravimetric organics (GRAV)—(organic
compounds with boiling points > 300 °C).
Individual semivolatile organic com-
pounds were identified and quantified us-
ing an approach similar to that used for
the volatile organics. The XAD-2 and par-
ticulate extracts were analyzed separately
by GC/MS to obtain mass spectral infor-
mation. The mass spectra of acquired data
were compared to mass spectra of multi-
component standard mixes as well as the
mass spectral database to assign com-
pound identification. Identified analytes
were quantified using a combination of
GC/MS and GC/FID system responses
based on the characteristics of the com-
pound identified.
The PCDD/PCDF samples were ana-
lyzed by low resolution GC/MS. Isotopi-
cally labeled homologues for all congeners
were used for qualitative and quantitative
purposes.
Metals potentially present in fluff were
chosen for characterization. The samples
were analyzed by inductively coupled ar-
gon plasma-atomic emission (ICAP) for
aluminum (Al), arsenic (As), barium (Ba),
-------
cadmium (Cd), total chromium (Cr), cop-
per (Cu), lead (Pb), magnesium (Mg), se-
lenium (Se), and zinc (Zn).
Results and Discussion
Nominally, 11 kg (25 Ib) of fluff was
evaluated for each test. Not all of the
material tested was actually combusted
during the testing. Only about 45% of the
mass of fluff placed in the combustion
apparatus was actually combusted during
the 200 min test. The remaining ash and
incombustible material was not character-
ized.
Maximum burn rates were observed
within 20 min of material ignition. After
this time, burn rates gradually decreased
throughout the burn. Peak temperatures,
observed by a thermocouple placed di-
rectly over the combustion apparatus, cor-
relate well with peak burn rates. Similarly,
peak concentrations for CO, CO2, and NO
emissions, correlate reasonably well with
peak burn rates. During the burns, O2
concentrations remained greater than 19%.
GC/MS analysis of the collected VOST
samples identified more than 50 com-
pounds. However, for the range of volatile
compounds characterized, more than 100
peaks were evident in the GC/FID chro-
matograms. Most of the compounds iden-
tified were alkanes, alkenes, 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.
Figure 2 depicts estimated emissions
for volatile organic compounds (VOCs)
common to Title III of the Clean Air Act
Amendments' (CAAA) Hazardous Air Pol-
lutants (HAP) list. Benzene represents the
largest single VOC emitted, generating
nearly 10 g for every kg of fluff consumed
in combustion.
The compounds identified in the XAD-2
and particulate fractions are similar to the
types of compounds identified in the VOST
samples. In addition, phenols, polycyclic
aromatic hydrocarbons (PAHs), phthalates,
and heterocycles were identified. Again,
the types of compounds identified were
consistent with the types of organics iden-
tified in various studies of the thermal
decomposition of plastics. Many of the
compounds identified common to the HAP
list were PAHs.
Table 1 presents estimated emissions
data for selected individual compounds
present in the XAD-2 and particulate frac-
tions. Because of the complexity of the
sample and its components, identifying all
Uj
40
35
30
25
20
S 15
.^
"w
Uj
10
Benzene
Toluene
Chlorobenzene
T77\ Ethyl Benzene
I 1 M/PXylene
1 ' I ' I
Test 1, Sample 1 ' Test 2, Sample 1 ' Test 3, Sample 2 '
Test 1, Sample 3 Test 2, Sample 2 Test 3, Sample 3
Figure 2. Estimated emissions for volatile HAPs.
compounds present in the organic frac-
tions was not within the scope of this
study.
Separate samples were also collected
for characterization of PCDD/PCDF emis-
sions. Separate analyses were performed
on the XAD-2 and particulate filter
samples. The total PCDD/PCDF emission
results are summarized in Figure 3,
Overall, the resulting emissions favored
the formation of the less-substituted chlo-
rinated dibenzofurans. The tetrachloro and
pentachloro dibenzofurans (TCDF/PeCDF)
were roughly an order of magnitude greater
in concentration than the dioxin homo-
logues. These profiles are similar to those
observed from soil samples collected from
scrap automobile incineration sites in the
Netherlands.
Of the 11 metals targeted, only Cd, Cu,
Pb, and Zn were detected in the samples
collected. Figure 4 presents the emission
rates for these metals. It is interesting that
Cu is present in relatively large concentra-
tions; Cu compounds have been investi-
gated as catalysts in the formation of
PCDDs/PCDFs in municipal waste incin-
eration processes.
Particulate matter was collected by sev-
eral different sampling systems: the semi-
volatile organics systems, the metal
aerosol system, and the PM10 (airborne
particles with diameters < 10 urn) ambient
sampler. A comparison of the PM10 to
total particulate, based on total averaged
values, indicates that PMin consists of
about 40% of the total particulate matter
collected.
To assess the overall organic emissions,
the volatile and semivolatile organics emis-
sion data were summarized. The total or-
ganics emitted (volatile, vapor-phase
semivolatile, and particulate-bound semi-
volatile) averaged more than 200 g/kg fluff
combusted. The actual mass contribution
from each fraction is summarized in Table
2.
As a measure of the quality of esti-
mated mass emissions, a total mass bal-
ance was performed. The diversity of the
measurements performed during testing
enabled the determination of total mass
emissions: they included most classes of
observed products of incomplete combus-
tion as well as several common gaseous
products of complete combustion. The ac-
tual mass balances, based on individual
and overall test average emission rate
values, are presented in Table 3.
Summary and Conclusions
The primary objective of this study was
to characterize, as completely as possible,
the emissions resulting from the simulated
open combustion of fluff. This necessi-
tated an approach where qualitative infor-
mation was given greater emphasis than
quantitative information. This approach pro-
vided the data and insight to direct subse-
quent, specialized, and more quantitatively
detailed investigations. An attempt was
made to characterize the diversity of the
emissions as efficiently as possible.
-------
Table 1. Estimated Emissions for Selected Pollutants (g/kg)
Compound
Ethyl Benzene
m- or p-Xylene
Ethynyl Benzene
Styrene
Benzaldehyde
Phenol
1,2 Dichlorobenzene
Naphthalene
Methylethylphenol
Biphenol
Acenaphthylene
Caprolactam
Phenanthrene
Fluoranthrene
Pyrene
Terphenyl
Bis(-2Ethylhexyl)-
phthalate
XAD
Test 1
2.26
1.03
0.38
6.27
1.20
1.39
ND
0.90
0.49
0.29
0.20
ND
0.211
ND
ND
ND
ND
XAD
Test 2
2.05
1.11
0.39
6.49
1.53
1.59
0.17
0.95
ND
0.30
0.18
ND
0.177
ND
ND
ND
ND
XAD
Average
2.16
1.07
0.38
6.38
1.36
1.49
0.09
0.92
0.24
0.29
0.19
0.000
0.194
0.000
0.000
0.000
0.000
Particulate Particulate
Test 1 Test 2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.068
ND
0.110
0.050
0.761
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.380
0.129
0.109
0.118
0.070
1.995
Particulate
Average
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.190
0.099
0.055
0.114
0.060
1.378
ND = Not Detected
The diversity of data produced from this
study is sufficiently comprehensive to pro-
vide a semiquantitative characterization of
the emissions resulting from the simulated
open combustion of automobile recycling
fluff. While the data may be adequate
from a physical and chemical character-
ization standpoint, data are lacking as to
the toxic effects from these emissions.
The relationship of this experiment to ac-
tual uncontrolled fluff combustion has not
been established.
0.0032
0.003
0.0028
0.0026
oi 0.0024
S 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
^H First Fluff Test
ES8SSI Second Fluff Test
Third Fluff Test
I I Average
HxCDDs
TCDFs
PeCDFs
HpCdds
OCDDs
HxCDFs HpCDFs
OCDFs
Figure 3. Total estimated PCDD/PCDF emissions.
-------
Only Detected Metals Shown
g
o
w
.w
1
1
CO
.w
LU
First Fluff Test
Second Fluff Test
Third Fluff Test
I I Average
Cadmium Copper
Figure 4. Estimated emissions for selected metals.
Zinc
Table 2. Estimated Emissions for Classes of Pollutants (g/kg)
Day 1
Day 2
Table 3. Mass Balance for Combustion Emissions (g/kg)
Day 1
Day 2
Day 3
Day 3
Average
Volatiles:
Early in test 5.93 43.06 25.73 24.91
Mid-test 17.28 69.16 70.87 52.44
Late in test 62.54 NA 58.44 60.49
Semivolatiles:
XAD-2 TCO 56.99 50.05 90.72 65.92
XAD-2GRAV 6.68 10.12 23.58 13.46
Paniculate TCO 0.61 1.27 0.85 0.91
Particulate GRAV 53.78 69.11 113.71 78.87
Particulates:
General Organic Train 91.25 116.17 183.44 130.29
Dioxin Train 85.82 115.68 188.63 130.04
Metals Train 81.55 89.93 174.49 115.32
Average (3 trains) 86.20 107.26 182.19 125.22
PM 66.03 NA 41.11 53.57
Average
Volatiles 28.58 56.11 51.68 45.46
Vapor-phase Semivolatiles 63.67 60.17 114.30 79.38
Particulate 86.20 107.26 182.19 125.22
(Average 3 trains)
COasC 67.93 71.54 72.09 70.52
CO asC 915.71 746.30 771.93 811.31
NO2asN NA 2.54 2.43 2.49
Sum 1162.10 1043.92 1194.62 1134.37
NA = Not Available
•U.S. Government Printing Office: 1993 — 750-071/60246
-------
-------
-------
J.V. Ryan and C.C. Lutes are with Acurex Environmental Corporation, Research
' Triangle Park, NC 27709.
Paul M. Lemieux is the EPA Project Officer (see below).
The complete report, entitled "Characterization of Emissions from the Simulated
Open-Burning of Non-Metallic Automobile Shredder Residue," (Order No. PB93-
172914; Cost: $19.50), will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC27711
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
EPA/600/SR-93/044
------- |