Emissions of Fluorinated Compounds From the Combustion of Carpeting

IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

EMISSIONS OF FLUORINATED COMPOUNDS FROM THE
COMBUSTION OF CARPETING

Paul Lemieux, Mark Strynar, Dennis Tabor, Joe Wood
U.S. Environmental Protection Agency, Office of Research and Development

Marcus Cooke, Barry Rayfield
CCI, Inc

Peter Kariher
ARCADIS G&M

ABSTRACT

One of the emerging waste streams that will likely be disposed of in combustors is carpet, due to its high
heating value and combustibility. Some of the stain-resistant coatings that carpeting is treated with
contain perfluorinated compounds (PFCs) such as perfluorooctanoic acid (PFOA) and their corresponding
homologues (C6 - Q4 acids) as well as fluorotelomer alcohols and fluoropolymers. PFOA has recently
been implicated as a chemical of concern due to its toxicity. It is unknown as to whether PFCs can be
released from combustion, or formed as by-products in combustors. This paper reports on a study in a
0.73 kW pilot-scale rotary kiln incinerator simulator to qualitatively and, where applicable; quantitatively
assess the potential for emissions of fluorinated compounds from combustion devices. In this study, a
limited number of PFCs were found in trace levels in the stack, and the concentrations were relatively
independent of kiln feed, suggesting that PFCs are effectively destroyed even under mild combustion
conditions, and the trace levels that were found were due to either trace contamination of the sampling
duct with fluorinated compounds due to historical use of Teflon and other fluorpolymers, or sampling
artifacts.

INTRODUCTION

When a building or outdoor area that was contaminated with biological/chemical agents or toxic industrial
chemicals, and is decontaminated and restored, a significant amount of the resulting residues may be
disposed of through thermal incineration. Common building materials have typically never been assessed
as combustion fuels; the materials removed from decontamination activities may not normally be fed in
large quantities to incinerators, and so research must be performed to assess the combustion of common
building materials when used as fuel in a combustion system. One of the main waste streams that will be
disposed of in combustors includes carpet, due to its high heating value, and combustibility. Cement
kilns, in particular, provide a good potential application for carpeting as a supplemental fuel (1).

Carpeting can be treated with stain resistant coatings. Some of these stain-resistant coatings contain
perfluorinated compounds (PFCs) such as perfluorooctanoic acid (PFOA) and their corresponding
homologues (C6 - Q4 acids) as well as fluorotelomer alcohols and fluoropolymers. PFOA has recently
been implicated as a chemical of concern due to its toxicity (2, 3, 4, 5, 6, 7). It is unknown as to whether
PFCs can be released from combustion, or formed as by-products in combustors. Ellis et al. 2002 (8)
suggest that thermolysis of fluoropolymers may be a source of halogenated organic acids in the
environment.

The primary goal of this study was to assess whether PFCs can be released from combustion facilities
burning carpeting.

-------
1X3"07 Conference, May 14-18, 2007, Phoenix, AZ

EXPERIMENTAL

The experiments were performed on the EPA's Rotary Kiln Incinerator Simulator (RKIS) facility located
in Research Triangle Park, NC. The RKIS has been described in detail elsewhere (9) and consists of a
0.73 kW (250,000 Btu/hr) primary combustion chamber followed by a 0.73 kW (250,000 Btu/hr)
secondary combustion chamber. Both the primary and secondary combustion chambers are fired by
natural gas. For these tests, the burner in the secondary combustion chamber was off. Fig. 1 shows the
RKIS.

Duel 2
TC

CEM Duct 3
(0) TC

Duct 4
TC

Secondary Combustion Chamber

SCO Mid
TC 0

SCC Mix
TC °

y/////////////////////^

Main
Burner

Afterburner

CEM (1),
Duct
TC

-2nd Floor

Surrogate
Waste
Injection

CEM (2),

Ramrod

-Ground Floor-

Kiln Section Transition Section

Fig. 1. EPA Rotary Kiln Incinerator Simulator

The approach to this study was to select a set of nominal operating conditions for the RKIS facility,
representing a relatively mild (i.e., temperatures < 1000 °C and the secondary combustion chamber off)
combustor operation, and manually feed a fixed size charge of carpeting into the RKIS over a period of
time, where Continuous Emissions Monitors (CEMs) were used to measure the concentrations of fixed
combustion gases and extractive organic sampling methods were used to assess the concentration of PFCs
and other fluorinated compounds in the stack.

Testing was performed using two different types of carpeting: one type that contained no stain-resistant
treatment, and one type treated with a stain resistant material that contains PFCs.

The test conditions are shown in Table I. The RKIS (with the main burner on and the secondary burner
off) was manually charged with 0.454 kg (1 lb) bundles, 1 bundle every 10 minutes, for a 3-hour run
duration. This translates to a 2.7 kg/hr (6 lb/hr) feed rate. Since the PFC emission levels were unknown,
it was desired that destruction efficiencies (DEs) of at least 99.99% be quantified based on concentrations
rather than based on the detection limits. It is estimated that there may be between 0.2 to 2 mg of PFC
present per kg of carpet (10). If it is assumed that there will be 99.99% DE of the PFC in the RKIS, 750
pg will be present in the sample train from a 1 hour run (feeding approximately 2.7 kg [6 lb] of carpeting).
This is approximately the same as the instrument detection limit. Therefore, based on reasonable
assumptions of DE and likely detection limits it was hypothesized that no PFC emissions might be

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

detected. In case of this occurrence, an additional test was performed where a commercially available
PFC-containing material (Zonyl®) was directly doped onto the carpet in the hopes that sufficient material
would be introduced into the RKIS to allow >99.99 % DE be quantified based on concentrations greater
than the detection limit. Carpet bundles were spiked by diluting 8.82 g of Zonyl into a 100 ml volumetric
flask and diluting with methanol. The 18 un-spiked bundles used in Run 6 were placed into a fume hood
for spiking and drying. Each bundle was spiked with 5 ml of the 88 mg/ml spiking solution for a total of
440 mg per bundle. Spiking was performed by pipetting the solution across the top face of the carpet
bundle. The spiking solution was observed to wick into the carpet bundles for uniform distribution. The
carpet samples were allowed to dry for 30 minutes to allow the methanol to evaporate. Two replicate runs
were performed on both treated carpet and untreated carpet, and a single combustion blank and a single
Zonyl-doped carpet run were performed.

Table I. Test Conditions

Run

1

2

3

4

5

6



NA
(Comb.

Untreated

Treated

Treated

Untreated

Zonyl
Doped

Feed

Blank)

Carpet

Carpet

Carpet

Carpet

Carpet

Run Time (min)

180

145

180

210

140

180

Carpet Fed (kg)

0.0

6.8

7.9

8.8

6.4

8.7

Avg. Carpet Feed Rate (kg/hr)

0.0

2.8

2.6

2.5

2.7

2.9

Avg. 02 f%l

14.0

12.6

12.2

12.1

11.7

12.1

Avg. C02 f%l

3.5

4.6

4.7

4.7

4.8

4.8

Avg. CO [ppml

23

60

83

85

113

82

Avg. NOx fppml

25

53

52

43

38

61

Avg. THC [ppml

2

3

3

3

6

1

Avg. Kiln T f°Cl

869

952

980

976

998

986

Avg. SCC Mix T f°Cl

472

521

536

532

542

534

Avg. SCC Mid T f°Cl

524

575

592

593

607

601

Avg. SCC Exit T f°Cl

490

542

558

562

575

571

Avg. Duct 1 T f°Cl

396

434

445

447

456

457

Avg. Duct 3 T f°Cl

311

339

347

348

354

357

Avg. Duct 4 T f°Cl

300

327

334

334

339

342

Avg. Duct 5 T f°Cl

282

307

313

314

319

322

Measurement of PFCs from combustors has never been done before. As such, there are no validated EPA
methods. Therefore, several modifications to standard EPA sampling methods were used, as well as an
innovative emerging sampling method (AMstack, with a version of Optizorb absorbent resin designed
specifically for PFOA/PFOS collection) (11) to attempt a qualitative assessment of PFCs in stack gases.
The sample probe used was a standard Modified Method 5 (MM5) (12) sample probe. The sampling was
not performed isokinetically, since the analytes of interest were generally gas-phase species only. The
collection media downstream of the probe varied. The samples that were simultaneously collected and
analyzed included:

1)	MM5 Probe/methanol impinger samples (n=6)

2)	MM5 Probe/XAD extract samples (n=6)

3)	MM5 Probe/Water impinger samples (n=19, combined into 6 runs) [downstream of XAD
cartridge)

4)	MM5 Probe/Tenax extracts (n=6)

3

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

5) AMstack Samples (n=6)

Methanolic extracts and impinger samples

Methanolic extracts of XAD and methanolic impinger samples were reduced in volume to 10 mL prior to
analysis. All samples were then centrifuged for 5 minutes at 5000 rpm to pelletize particulate matter. An
aliquot of methanolic samples (1.0 mL) was passed through a 0.45 (.un nylon centrifuge filter to remove
particulate matter. The sample methanol passing the filter was combined with 2mM ammonium acetate
(50:50) to give a concentration of the internal standard (IS) (13C2-PFOA) at 20 ng/ml. A standard curve
including amethod blank and 0.5, 1.3.5, 10 and 25 ng/ml of all analyzed PFCs was constructed as well.
Standards were treated as unknowns in all aspects. Standard curves were constructed for each individual
PFC by plotting the analyte/IS area on the y axis vs. analyte concentration/IS concentration on the x-axis
with 1/x weighting. All standards curves had r > 0.999. The Lower Limit of Quantitation (LLOQ) for
each compound was 0.5 ng/mL. Since only 1 or 10 mL was analyzed for the these samples, final
numbers will need to be multiplied by 10 to get the total PFC content of the whole methanolic solution.
The results section below gives the value in one ml and the total ng in the entire methanolic impinger
sample or XAD extract.

Water impinger samples

Solid Phase Extraction (SPE) HLB (60 mg) columns from Waters were preconditioned with 5 ml of
methanol and 5 ml of DI water. Each column then received 20 ng of 13C2-PFOA and was loaded with 3
ml of DI water. Next each vessel in its entirely was added to the SPE column via vacuum manifold and
funnel adapters. Some samples contained large amounts of particulate matter that clogged the SPE
column and not all was able to pass through, making quantitation difficult for these samples. Next the
SPE columns were centrifuged to remove residual water and eluted with 4.0 ml of methanol (2 ml; 2x).
Next 0.5 ml of elution methanol was centrifuged at 12,000 rpm for 10 minutes to pelletize particulate
matter. An aliquot of the supernatant was combined with 2 mM ammonium acetate (50:50) for LC/MS-
MS analysis. A standard curve was constructed by adding 0.5, 1, 2, and 5 ng of all PFC to a tube with 20
ng 13C-PFOA and 180-PFOS and 4 ml of methanol. Standards were otherwise treated as unknowns in
preparation.

Tenax extracts

Prior to analysis 2 Tenax tubes were utilized for some methods development. One tube was spiked with
internal standards, and one with internal standards plus the PFCs to be analyzed for. Replicate quantities
of IS and PFC of interests were spiked into polypropylene receiving vessels. Tenax tubes were eluted
with methanol, first with 10 ml, then 5 ml, followed by forced aeration. Quantities based on area counts
for IS and PFCs were compared between Tenax spiked samples and spiked samples for calculations of
recovery. Recoveries ranged from 86.5 % (PFHS) to 104.2 % (180-PFOS), and were deemed adequate
for analysis of Tenax for PFCs. Each Tenax tube was spiked with 50 ng 13C-PF0A/180-PF0S at the top
of the Tenax column. Tubes were then gravimetrically eluted with 10.0 ml methanol, followed by 5.0 ml
methanol. After gravimetric dripping had ceased, forced aeration was applied until no methanol was
recovered. Of the 15 mL applied 10.5 to 12.0 ml was recovered. A sample aliquot of the methanolic
extract (0.2 ml) was combined with 2mM ammonium acetate (50:50) for LC-MS/MS analysis. A standard
curve was constructed by adding 10, 20, 50 and 100 ng of all PFCs and 50 ng 13C-PF0A/180-PF0S to a
tube then adding 15 mL of methanol. Standards were treated as unknowns otherwise.

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

AMstack Samples

The AMstack sampling system utilized an unheated, borosilicate glass sampling probe, glass absorbent
module, with 316L stainless mesh support frits to contain the absorbent resin. This is followed by a series
of chilled impingers to remove moisture prior to the metering console. Two sections of resin were used,
one with 15 to 20 grams of resin, the second with 5 grams of resin for breakthrough verifications. The
second section of resin is also treated with a proprietary, chemically bonded color indicator determined to
be non-reactive to the target compounds. This indicator also is permanently bonded into the Optizorb
resin, so it is not extractable during sample preparations. The color indicating resin was incorporated into
this test series to verify presence/absence of two conditions that may affect PFC collection efficiencies,
namely overheating at the resin cartridge, and pH conditions that could affect the capture of the target
compounds. During this test series, excessive absorbent heating that could cause resin degradation, and
neutral to basic stack gas conditions that could lower the capture efficiency, were not found. This
indicates the sample integrity was likely not compromised by any adverse sampling conditions.

HPLC MS/MS analysis

All samples with the exception of the AMstack were analyzed by EPA's National Exposure Research
Laboratory (NERL) via HPLC/MS/MS. The AMstack samples were analyzed by AXYS Laboratories
(Sydney, B.C., Canada). The NERL-analyzed samples were prepared so as the final solution to be
analyzed consisted of 50:50 2 mM ammonium acetate: methanol. A sample volume (10 |J) was injected
into a mobile phase or 23:77 2mM ammonium acetate: Methanol flowing at 200 inutes. Analytes
were resolved chromatographically using a Waters Sunfire C8 column (50 x 3 mm) 5 um particle size.
Multiple Reaction Monitoring (MRM) was done for each individual PFC of interest. The area ratio of the
analyte to the corresponding IS was taken for quantitation. Standard curves were constructed where
necessary, and analyte concentrations were based on generated standard curves for that day. Double
blanks and method blanks were run with each assay to assess if systematic contamination was occurring.

The AMstack samples were spiked with isotopically-labeled surrogate standards, diluted with reagent
water, cleaned up on SPE cartridges and analyzed by liquid chromatography/mass spectrometry (LC-
MS/MS). Up to 100 mL of methanol were diluted to 1 L with reagent water. The sample was applied to a
Waters Oasis WAX SPE cartridge and then eluted with 2 mL of basic methanol. An aliquot of recovery
standard was added and the volume adjusted to 4 mL in preparation for LC/MS/MS analysis. Final
sample concentrations were determined by isotope dilution/internal standard quantification against matrix
matched calibration standards carried through the analysis procedure alongside the samples. Detection
limits were in the 0.5 - 0.8 ng range.

Target Analytes

Since PFCs had not been measured in combustion samples before, selection of a list of target analytes
was constrained by availability of standards. Table II shows the list of target analytes used in this study.
The samples analyzed by AXYS had a few additional analytes.

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

Table II. List of Target Analytes

Compound

Acronym

Structure

Perfluorobutanoic acid

PFBA

-K

214

Perfluoropentanoic acid

PFPeA

H—

264

Perfluorohexanoic acid

PFHxA

I——

314

Perfluoroheptanoic acid

PFHpA

j—]—

364

Perfluorooctanoic acid

PFOA

-++K

414

Perfluorononanoic acid

PFNA

'-j-j-j—

tt+K

464

Perfluorodecanoic acid

PFDA

~h+K

514

Perfluoroundecanoic acid

PFUnA

' —

| i i 11 (

564

Perfluorododecanoic acid

PFDoA

-hH+K

614

Perfluorobutane sulfonate

PFBS

H—

IU-
i

299

Perfluorohexane sulfonate

PBHxS

--j—j—

II .

1°

399

Perfluorooctane sulfonate

PFOS

499

Perfluorodecane sulfonate

PFDS

599

Perfluorooctane sulfonamide

PFOSA

~H+H

499

RESULTS

Analysis of Carpet by X-Ray Fluorescence

The carpet samples were subjected to X-Ray fluorescence (results shown in Table III). The carpet fiber
and backing were analyzed separately (3 samples each) and then the results were averaged into a
composite result. The fiber samples were evaluated as oxides, with the unaccounted-for mass assumed to
be Nylon-6. The backing samples were evaluated as oxides and C02, and the unaccounted-for mass
assumed to be cellulose. Fluorine was found in all the carpet samples, including those that were
supposedly not treated with stain resistant coatings. It is unknown why this was observed, but since XRF
is an elemental analysis, the F may not have been from stain-resistant coatings.

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

Table III. XRF Analytical Results (mass %)

Carpet Fiber 		 Carpet Backing

Element

Treated
Composite

Untreated
Composite

Element

Treated
Composite

Untreated
Composite

F

0.289

0.175

Na

0.486

0.331

Na

0.024

0.012

Mg

2.093

0.312

m2

0.013

0.005

A1

0.239

0.085

Si

0.004

0.006

Si

0.376

0.199

P



0.007

P

0.114

0.037

S

0.053

0.071

S

0.089

0.092

CI

0.009

0.010

CI

0.005

0.004

K

0.078

0.036

K



0.009

Ca

8.12

9.31

Ca

0.032

0.046

Ti

0.018

0.004

Ti

0.004

0.077

Mn

0.012

0.002

Mn



0.002

Fe

0.172

0.035

Fe

0.005

0.004

Ni

0.004

0.004

Ni

0.004

0.004

Sr

0.007

0.018

Sr

0.002

0.011

O

5.81

4.55

Ba



0.019

C02

12.69

10.77

O

0.122

0.190

Cellulose

69.68

74.20

Nylon 6

99.43

99.45



Methanolic XAD extracts and Methanolic impinger samples

Table IV lists the PFCs found in the methanol impingers. Most compounds were either not detected or
below the limits of quantitation. The only PFC that was consistently found in the samples was PFHxA,
and there was not a statistically significant difference between the results from all the runs, including Run
6 which had doped large quantities of Zonyl® into the carpet. These results suggest that any PFCs that
we are measuring are not likely due to any PFCs that were present in the carpet that was fed into the
RKIS. There is also the possibility that the PFCs aren't getting trapped by the sampling media; although
a wide range of potential methods were used to acquire the sample, the entire universe of options was not
explored.

One possible explanation is that the nearly ubiquitous use of Teflon and other fluoropolymers on the
RKIS ducting (in the form of Teflon tape and sampling line) has resulted in trace contamination of the
sampling duct with low levels of PFCs. Another possible explanation is that in spite of best efforts to
minimize use of Teflon in all aspects of sampling for these tests, some residual Teflon may have
contacted the sampling equipment or may have been used in the manufacturing process of the solvents,
sampling, or analytical equipment, resulting in artifacts.

7

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

Table IV. PFCs in Methanol Impingers

Run

1 (Comb.
Blank)

2 (Untreated
Carpet)

3 (Treated
Carpet)

4 (Treated
Carpet)

5 (Untreated
Carpet)

6 (Zonyl-
doped Carpet)

Compound

ng/m3

ng/m3

ng/m3

ng/m3

ng/m3

ng/m3

PFDoA

ND

ND

ND

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ



1 (Comb.
Blank)

2

(Untreated
Carpet)

3 (Treated
Carpet)

4 (Treated
Carpet)

5

(Untreated
Carpet)

6 (Zonyl-
doped
Carpet)

Compound

ng/m3

ng/m3

ng/m3

ng/m3

ng/m3

ng/m3

PFOA

0.4

0.5

0.3

0.3

-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

Table V

II. PFCs from AMstack Sampling Train

Run

1 (Comb.
Blank)

2

(Untreated
Carpet)

3 (Treated
Carpet)

4 (Treated
Carpet)

5

(Untreated
Carpet)

6 (Zonyl-
doped
Carpet)

Compound

ng/m3

ng/m3

ng/m3

ng/m3

ng/m3

ng/m3

PFBA

ND

ND

ND

ND

ND

ND

PFPeA

ND

ND

ND

ND

ND

ND

PFHxA

0.2

0.3

0.3

0.2

0.5

0.8

PFHpA

ND

ND

0.2

ND

ND

0.3

PFOA

0.5

1.0

1.3

1.3

2.2

2.0

PFNA

ND

ND

ND

ND

ND

ND

PFDA

ND

ND

ND

0.2

ND

0.5

PFUA

ND

ND

ND

ND

ND

ND

PFDoA

ND

ND

ND

ND

ND

0.3

PFBS

ND

ND

ND

ND

ND

ND

PFHS

ND

ND

ND

ND

ND

ND

PFOS

0.6

1.6

1.5

1.6

3.3

3.4

PFOSA

0.3

0.3

0.6

0.4

0.9

0.4

PFDS

ND

ND

ND

3.5

3.0

3.8

< LOQ = detected but below limit of quantitation
ND = not detected

Summary of Sampling Method Effectiveness

Fig. 2 shows a grid of which sample techniques caught which species. The Tenax clearly did not catch
any of the target analytes. The methanol impingers seemed to catch the greatest diversity of the acids, but
did not catch any of the sulfur-containing compounds. The AMstack was able to catch the sulfur-
containing target analytes. It must be noted that the water impingers were downstream of the XAD trap,
so only the materials that broke through the XAD were caught in the water impingers.

Fig. 3 shows a subset of the data, showing the concentrations of the Hexa- and Octa- substituted PFC
concentrations as a function of sampling technique. Again, clearly the Tenax did not catch any of these
particular targets. The combined total for the XAD and the water impingers appears to be approximately
the same as the results for the methanol impingers, suggesting that similar amounts of those compounds
were caught in the XAD and water impingers compared to the methanol impingers. However, looking at
Fig. 2, some compounds that were caught in the methanol impingers were not caught at all in the
XAD/water impinger train.

10

-------
1X3"07 Conference, May 14-18, 2007, Phoenix, AZ

Quantified in at least half the samples

T

Detected at LOQ in at least half the samples



Not Detected in at least half the samples

i

Not Analyzed for

9

PFBA

9

9

9

9

1

PFPeA

9

9

9

9



PFHxA

T

t

T

i

t

PFHpA



t

T

i



PFOA

T

t

t

i

t

PFNA





I

1

I

PFDA

->

i

I

1

I

PFUA

->

I

I

1

I

PFDoA

i

I

I

1

I

PFBS

i

I

I

1

I

PFHS

i

I

I

1

I

PFOS

i



I

1

t

PFOSA

9

9

9

9

t

PFDS

9

9

9

9

t



MM5
Probe/
methanol
impinger

MM5
Probe/
XAD
extract

MM5
Probe/
Water
impinger

MM5
Probe/
Tenax
extracts

AMstack

Fig. 2. IV

atrix of Ability of Sampling to Capture Different Species

11

-------
IT3'07 Conference. May 14-18, 2007, Phoenix, AZ

prj 1000

E

o
*3
(0

c

QJ
-------
IT3'07 Conference, May 14-18, 2007, Phoenix, AZ

REFERENCES

1.	Stephenson, G., Source Testing Final Report: Fuel Comparison - Coal only and Carpet Co-Fire,
EPA/600/R-06/039, 2006.

2.	Lau, C., Butenhoff, J.L., and Rogers, J.M., "The developmental toxicity of perfluoroalkyl acids
and their derivatives," Toxicol ApplPharmacol. 2004 Jul 15; 198 (2): 231-41.

3.	Lau, C., Thibodeaux, J.R., Hanson, R.G., Narotsky, M.G., Rogers, J.M., Lindstrom, A.B.,

Strynar, M.J., "Effects of perfluorooctanoic acid exposure during pregnancy in the mouse,"
Toxicol Sci. 2006 Apr; 90(2): 510-8.

4.	Wolf, C.J., Fenton, S.E., Schmid, J.E., Calafat, A.M., Kuklenyik, Z., Bryant, X.A., Thibodeaux
J., Das, K.P., White, S.S., Lau, C.S., Abbott, B.D., "Developmental toxicity of perfluorooctanoic
acid in the CD-I mouse after cross-foster and restricted gestational exposures," Toxicol Sci. 2007
Feb; 95(2): 462-73.

5.	EPA Science Advisory Board, Perfluorooctanoic Acid Human Health Risk Assessment Review
Panel (PFOA Review Panel, http://www.epa.gov/sab/panels/pfoa rev panel.htm

6.	Mortimer D. N., et al., "Perfluorinated Compounds In The UK 2004 Total Diet," 26th
International Symposium on Halogenated Persistent Organic Pollutants, August 21-25, 2006,
Oslo, Norway.

7.	D Brooke, A Footitt, T A Nwaogu, Environmental Risk Evaluation Report:
Perfluorooctanesulphonate (PFOS), Environment Agency, Chemicals Assessment Section, Isis
House, Howbery Park, Wallingford OX 10 8BD, UK

8.	David A. Ellis, Scott A. Mabury, Jonathan W. Martin & Derek C. G. Muir. (2002). "Thermolysis
of Fluoropolymers as a potential source of halogenated organic acids in the environment," Nature
(412) 321-324.

9.	P. Lemieux, E. Stewart, M. Realff and J. A. Mulholland, 2004, "Emissions Study of Co-firing
Waste Carpet in a Rotary Kiln," Journal of Environmental Management, 70, 27-33

10.	Washburn, Stephen T., T. S. Bingman, S. K. Braithwaite, R. C. Buck, L. W. Buxton, H. J.
Clewell, L. A. Haroun, J. E. Kester, R. W. Rickard, and A. M. Shipp, "Exposure Assessment and
Risk Characterization for Perfluorooctanoate in Selected Consumer Articles," Environ. Sci.
Technol., 2005, 3904-3910.

11.	Clark, G., Cooke, M., Touati, D., Rayfield, B., Chu, M. (1998) "Industrial Combustor Validation
Test Measuring Air Emissions of Dioxin and Furans on a TEQ basis using EPA Method 23 and
HRGC/HRMS versus AMBStack Sampler and Calux Bioassay,"

12.	EPA Test Method 0010 "Modified Method 5 Sampling Train" in Test Methods for Evaluating
Solid Waste, Volume II, SW-846 (NTIS PB88-239223). Environmental Protection Agency,
Office of Solid Waste, Washington, DC, September 1986.

13

-------