United States
Environmental Protection
Agency
National Exposure
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/SR-98/030 May 1998
of
EPA
James F. McGaughey, Joan T. Bursey, and Raymond G. Merrill
A field test was completed to evalu-
ate EPA Method 0040 (Sampling of
Principal Organic Hazardous Con-
stituents from Combustion Sources
Using Tedlar® Bags), a method de-
signed to collect volatile organic com-
pounds at concentrations that are
above the range of EPA Method 0030
(VOST). Method 0040 is based on the
successful results of laboratory stud-
ies to develop and refine a sampling
train and methodology to collect and
analyze volatile organic compounds
existing in source emissions between
100 and 1000 |ig/m3. After demonstrat-
ing satisfactory recovery of target
compounds in the laboratory, a field
evaluation of this method was per-
formed to determine and document
the systematic error (bias) and ran-
dom error (precision) of the method
under stationary source sampling
conditions. Four similar trains were
operated simultaneously using a qua-
druplicate sampling probe (quad
probe). The field evaluation was ac-
complished by dynamically spiking
the sampling trains with specific com-
pounds while simultaneously sam-
pling emissions from a coal
combustion source. Analytical data
were statistically evaluated accord-
ing to the procedures of EPA Method
301 (Field Validation of Pollutant Mea-
surement Methods from Various
Waste Media). Fifteen of eighteen
volatile organic test compounds met
Method 301 acceptance criteria for
performance of Method 0040.
This Project Summary was developed
by EPA's National Exposure Research
Laboratory, Research Triangle Park, NC,
to announce key findings of the research
project that is fully documented in a
separate report of the same title
Project Report ordering information at
back),
Introduction
A field test was performed to evaluate a
method for sampling and analyzing vola-
tile organic compounds from stationary
source emissions using EPA Method 0040
(Sampling of Principal Organic Hazardous
Constituents from Combustion Sources
Using Tedlar® Bags). Method 00401 was
designed to collect volatile organic com-
pounds at concentrations between 100 and
1000 (ig/m3 (above the range of EPA
Method 0030 [VOST]). The field study was
performed using an experimental design
for the sampling strategy that follows the
statistical approach outlined in EPA
Method 301, Field Validation of Pollutant
Measurement Methods from Various
Waste Media.2
The original sampling train used in this
study was developed by Radian Corpora-
tion, under contract to the U.S. Environ-
mental Protection Agency (Contract No.
68-D1-0010, Work Assignment No. 57).
The laboratory development work involved
the design of the sampling train, labora-
tory evaluation of the train using dynamic
spiking, and preparation of a standard
sampling method in the SW-846 format.
The analytical portion of the method was
based on gas chromatography/mass spec-
trometry (GC/MS) to provide qualitative
identification of specific compounds and
quantitative emissions results based on
calibration for individual compounds. A let-
-------
ter report submitted to EPA on June 30,
1994, discussed the results of the labora-
tory study and provided suggestions for
modification of the train for improved op-
eration.
The most important issue raised by the
subsequent laboratory studies of the Pro-
posed Method was the correct measure-
ment of the volume collected in the
Tedlar® bag and confirmation of this value.
In the final laboratory study (Contract 68-
04-0022, Work Assignment 10), the vol-
ume of air collected by evacuating a
nominal 20 liters of air from the rigid con-
tainer was determined before and after
the removal of the Teflon© check valve
assembly located just prior to the bag
inlet valve. Significant deviation between
theoretical and observed volume was en-
countered with the check valve in place.
Removal of the check valve resulted in
experimentally determined volumes rang-
ing from -1.0 to 3.5% relative standard
deviation. Based on these results, the de-
sign of the sampling train was changed
before the field test, and the Teflon® check
valve was replaced with a glass/Teflon®
stopcock. An additional thermocouple was
also added in the rigid container to moni-
tor the temperature of the sampled gas.
Ater testing the modified sampling train
and demonstrating satisfactory recovery
of target compounds in the laboratory, a
field evaluation of this revised method was
performed to determine and document the
systematic error (bias) and random error
(precision) of the method under stationary
source sampling conditions. Four (4) simi-
lar trains were operated simultaneously
using a quadruplicate sampling probe
(quad probe). The field evaluation was
accomplished by dynamically spiking the
sampling trains with specific compounds
from a certified compressed gas cylinder
while simultaneously sampling emissions
from a coal combustion source.
Experimental Approach
The primary objective of this program
was to evaluate EPA Method 0040, incor-
porating the modifications approved by the
Work Assignment Manager, under actual
field conditions following an experimental
design based on EPA's Method 301.
Method 0040 was evaluated for sampling
and analysis of 18 volatile organic com-
pounds selected from those listed in the
Clean Air Act Amendments of 1990 that
had previously been tested in the field
using the VOST methodology. Additional
volatile organic compounds that did not
meet Method 301 acceptance criteria in
the VOST methodology were also included
in the field test of Method 0040. EPA
Method 301 was used for the statistical
design of the sampling strategy and for
the statistical evaluation of the results ob-
tained from dynamic spiking of two of four
collocated trains. This section describes
the experimental approach to accomplish
this field test and includes the sampling
site description, the sampling approach,
the analyte spiking techniques, and ana-
lytical procedures.
Test Description
The field evaluation was conducted at a
coal-fired power plant with four boiler units
equipped with electrostatic precipitators but
no caustic scrubbers. This site was se-
lected because the source produced the
primary components of a combustion ma-
trix (moisture, sulfur dioxide, oxides of ni-
trogen and particulates) without any of the
targeted volatile organic compounds to in-
terfere with the interpretation of spiked
compound concentrations. The stack tem-
perature was measured during a previous
presurvey and found to be approximately
132°C (270°F). Sampling was performed
at Unit 2 because prior sampling efforts
performed by ERG had already character-
ized Unit 2. Unit 2 also has a large sam-
pling platform and easy access to sampling
ports, and the plant personnel provided
support and cooperation for the test se-
ries.
Sampling
Sampling procedures were consistent
with EPA Method 0040. The sampling train
(Figure 1) was modified by employing a
quad probe system which contained four
similar heated sampling probes that could
be inserted into the stack as one unit.
This multiple probe configuration allows
the simultaneous collection of stack gas
in four similar trains with gaseous dynamic
spiking in two of the trains.
Prior to shipment to the sampling site,
all glass components of the sampling train
were cleaned, wrapped in aluminum foil,
and segregated to prevent contamination.
Tedlar® bags were cleaned and blanked
according to the procedures of EPA
Method 0040 with 10% of the bags ana-
lyzed by GC/MS to verify that appropriate
cleaning criteria were met. All Tedlar®
bags and rigid containers were leak
checked prior to transport to the test site.
The rigid containers for the Tedlar® bags
were evacuated to approximately 25 inches
of mercury (in. Hg) and allowed to stand
for 30 minutes while monitoring a vacuum
gauge. All containers were found to be
leak free. Each container with correspond-
ing lid was numbered so these two com-
ponents were always used as a unit. In
addition each bag was filled almost to
capacity with clean nitrogen, sealed and
allowed to stand overnight. A visual in-
spection of each bag indicated that they
were leak free. Each bag was then num-
bered, attached to the inlet quick-connect
fitting on a rigid container lid, re-evacu-
ated and sealed in the rigid container.
The containers and bags were then ready
to be transported to the test site.
Dry gas meters were leak checked and
calibrated prior to transport to the field
site and the results of the calibrations
kept in dedicated laboratory notebooks.
The remaining preparation included leak
checking the sampling umbilicals, tempera-
ture readout calibration, and functional
checks of other associated sampling equip-
ment
Sampling trains were assembled in the
field laboratory, without Tedlar® bags or
their rigid containers. Tedlar® bags in rigid
containers were transported separately to
the sampling location and positioned in
the sampling train after the probe was
inserted in the stack port. Sampling was
started after all sampling trains were com-
pletely assembled and leak checked.
Each heated train component was
brought to 130°C and each train was leak
checked in the field by following the pro-
cedures outlined in Method 0040. Each
probe of the quad assembly was capped,
and the train (excluding the bag and rigid
container) was evacuated to 10 in. Hg.
Any leaks were located and repaired until
all trains were found to be leak free. Each
rigid container was then inspected to verily
that the previously evacuated bags were
still leak free.
Sampling was performed by evacuating
the rigid container containing the Tedlar®
bag. Stack gas was withdrawn from a
single port in the stack through the four
(4) probe quad assembly and directed to
four similar sampling trains as shown in
Figure 1. The front end of the quad probe
was positioned in the center of the stack
and remained in that location during each
day of testing. The true concentration of
the components of the stack was of no
interest to this program and, therefore,
traversing was not required. The sampling
flow rate always exceeded the spiking rate
by at least a factor of 2 to ensure that all
of the spiked analytes were carried totally
into the sampling train. Two of the trains
of each quad run were spiked and two
were unspiked. Eleven (11) quad runs
were performed. On three of the four days
of sampling, the rigid containers were
heated to maintain the temperature above
0°C.
The sampling flow rate was set to a
nominal 0.33 L/min and sample was col-
lected for one hour simultaneously in each
-------
Duct
Temperature
sensor
Probe
Probe isolation valve
Heated filter
holder
(with filter)
,= — Charcoal trap
, Purge line
Condenser Quick'
^ connectors
, Glass condensate trap
fc. I
-ffi
Ice bath I I"
ra
Bag isolation
valve
,®
' " T>
Condenser temp (4)^11
! f 13 1 r-;n-. » J.A »».» . I—I —
Teflon
Bulkhead
Union
Stack temp fj
j Filter temp ^ • ' L_
©Probe temp
Airtight container'
Tedlar^
Injection port
for spiking
Figure 1. Schematic of the Method 0040 sampling train.
, Silica gel
, Charcoal trap
Quick connector
A- .1
To VOST control
console
- Quick connectors
Teflon®Union
train. At the end of each one-hour sam-
pling run, the rigid container was isolated
from the rest of the train and then the
trains were leak checked. Each rigid con-
tainer was opened to visually inspect the
Tedlar® bag to determine if it had been
filled to approximately 80% of capacity
indicating that sample had been collected.
No condensate was observed in any of
the trains during any of the sampling runs,
so no condensate samples were collected.
However, each condensate flask was
rinsed with organic-free water after each
run. Sample transport to the laboratory
was scheduled for every day of sample
collection to ensure that the Tedlar® bag
samples remained above 0% C. After the
completion of sample recovery, the next
quad run was performed as discussed
above.
In addition to the eleven quad samples
taken for method evaluation, four field
blanks, one for each of the trains, were
taken following the procedures of Method
0040. These field blank samples were
transported and analyzed with the stack
gas samples. Two Tedlar© bags were filled
with high purity nitrogen, labeled as labo-
ratory blanks, and left in the laboratory
under the same storage conditions as the
field samples. Samples were transported
daily from the test site to the laboratory.
Dynamic Spiking
To evaluate the complete sampling and
analytical methodology for Method 0040,
it was necessary to perform dynamic spik-
ing in the field. To ensure the performance
of the spiking equipment and the sam-
pling techniques under controlled condi-
tions, Method 0040 trains were set up in
the laboratory and dynamically spiked us-
ing the proposed apparatus prior to field
deployment. In the field evaluation test,
the same certified standard cylinder and
the same dynamic spiking procedures
were used.
The compounds dynamically spiked into
two of the Method 0040 sampling trains
were contained in a compressed gas cyl-
inder. The cylinder was commercially pre-
pared at a concentration of 100 ppm for
each analyte, as verified by GC/MS. A
second similar compressed gas cylinder
was purchased and verified to be avail-
able as a backup. During each quad sam-
pling run, spiking gas was continuously
introduced into two of the four Method
0040 trains through two fine-metering
valves. The flow rate of spiking gas into
each train was nominally 80 mL/min, re-
sulting in the introduction of approximately
20 ppm of each compound into the sam-
pling train over a sampling period of one
hour while stack gas was collected at 0.33
liter/min (i.e., nominal 5 liters of spiking
gas with 20 liters of stack gas in a 30-liter
Tedlar® bag). Each gas metering system
was equilibrated for approximately 30 min-
utes before the start of sampling. The
gaseous spike was introduced into each
train at a point immediately after the probe
and before the filter and condenser. The
regulator and tubing leading to each train
were maintained at a temperature of 130 -
140°C(266-284°F).
A three-way glass and Teflon® valve
(glass tee) was used to introduce the spik-
ing gas into the trains. During the condi-
tioning period and between each of the
runs, the valve was used to by-pass the
train and allow gas to exit through a char-
coal trap to the ambient air. The valve
position was then changed to deliver the
spiking gas into the train for a specified
period of time. The sequence of events
for each spiking run was as follows:
Leak check each train;
-------
Adjust spiking gas flow rate to a
nominal 80 mL/min with the fine
metering valve;
Measure the flow rate of the spik-
ing gas at each of the two trains
designated for spiking using a
bubble flow meter. Make the mea-
surement in triplicate and average
the values;
Record the initial meter reading,
start all four trains simultaneously,
connect the rigid containers to the
meter box and the sampling train,
adjust the sampling flow to a flow
rate of 0.333 L/min, and sample for
one hour;
Immediately after the start of the
sampling run, divert the spiking gas
into the two spiked trains by mov-
ing the valve and activate a stop
watch;
Spike the deuterated analogs if ap-
propriate;
After a one-hour sampling period,
move valve back, record the
elapsed time from the stop watch,
stop the sampling pump and iso-
late the rigid container/Tedlar® bag
from the train, and measure the
spiking flow in triplicate and aver-
age;
Perform a post-leak check on all
trains; and
Recover train: Collect any conden-
sate into a 40 ml vial and label all
recovered train components.
Prior to the field test, the spiking gas
flow rate to be delivered to each train was
calculated to be a nominal 80 mL/min to
provide a nominal concentration of 20 ppm
of each compound in the bag (approxi-
mately a 1:5 dilution of a 100 ppm cylin-
der spike gas). The analysis of a 5 ml
aliquot of each sample with the analytes
at a concentration of 20 ppm provided
results in the range of the GC/MS calibra-
tion curve.
Selected quad trains were also spiked
with isotopically labeled (deuterated)
compounds as field spike samples. Ben-
zene-d6, hexane-d14, and 2,2,4-trimethyl-
pentane-d18 were spiked into the Tedlar®
bags. A 10|xL amount of a methanol
solution containing these compounds
was spiked through an injection tee at
the point where the stack gas enters the
Tedlar® bag. These isotopically labeled
field spikes were used to verify the vol-
ume of gas collected in the Tedlar®
bags.
Analysis
Gaseous samples in Tedlar® bags were
analyzed by GC/MS, using an injection
loop to introduce a constant volume of
sample into the GC. Analytes were
cryofocused and then introduced onto the
head of the analytical column. Analysis
was performed using a fused silica capil-
lary column, DB-1, under analytical condi-
tions following the guidelines of EPA
Method 8260.3 Electron ionization (El)
mass spectrometry was used.
Calculations of compound concentra-
tions were based on the injection of a 5
ml sample from the Tedlar® bag, a nomi-
nal concentration range of 200-600 ng on
column on the basis of field-spiked val-
ues. Appropriate dilutions from the spiking
cylinder were used to prepare calibration
standards. The initial five-point calibration
for the GC/MS analysis was performed
over the range of approximately 20-1000
ng for each compound of interest. All stan-
dards were prepared in Tedlar® bags and
stored at ambient temperature. Daily veri-
fication of the response factors was per-
formed following the guidelines in Draft
Method 5041.4
Tedlar® bag samples were stored at
laboratory ambient temperature and ana-
lyzed within 72 hours of sampling to meet
the method hold time requirements. A 5
ml gaseous aliquot of the internal stan-
dards (bromochloromethane, 1-bromo-4-
fluorobenzene, chlorobenzene-d5, and 1,4-
difluorobenzene) was cryofocused along
with each sample prior to introduction onto
the GC column.
Six complete and valid quad runs (24
Tedlar® bags) are required to satisfy the
requirements of EPA Method 301, the field
validation protocol. Eleven quad sampling
runs were collected to provide additional
samples in case of loss or breakage. All
samples were analyzed and results are
reported and used in the statistical calcu-
lations.
Discussion
All samples were analyzed within the
72-hour hold time specified in Method
0040. The GC/MS system was calibrated
using the same compressed gas standard
used for spiking in the field. The concen-
tration of each of the compounds in the
cylinder was a nominal 100 ppm. Dilu-
tions of this 100 ppm standard were made
in Tedlar® bags covering the range of 5
ppm to 100 ppm (or approximately 20 ng
to 1000 ng on column). The expected
concentration of each analyte in the bag
samples was a nominal 20 ppm. There-
fore, a five ml sample aliquot from each
bag sample used for analysis was ex-
pected to provide 200-600 ng on column
depending on the specific compound. All
analytical results were presented as ppm.
The theoretical concentration of each
analyte in the spiked trains was calcu-
lated by determining the amount of dilu-
tion of the volume of gas spiked into each
Tedlar® bag. The volume of spiked gas
was calculated by multiplying the average
of the spiking gas flow rate (values from
pre- and post-sample collection) in ml/
min by the length of the spiking period
(nominally 60 min). This value was then
divided into the sum of the dry gas meter
volume and the spiking volume to get a
dilution factor. This dilution factor (nomi-
nally a value of 5) was then divided into
the concentration of each analyte con-
tained in the spiking gas cylinder (nomi-
nally 100 ppm). No target analytes were
detected in the unspiked trains.
The percent recovery of each compound
for each spiked train was calculated by
dividing the analyzed value by the theo-
retical value and multiplying by 100. Us-
ing the criterion of 50 to 150% as
acceptable recovery, all compounds ex-
cept bromomethane fell within this range.
Using the precision criterion (%RSD less
than 50), all compounds except dichlo-
rodifluoromethane performed successfully.
Dichlorodifluoromethane and 1,3-butadiene
exhibited recoveries that were generally
low and variable while bromomethane ex-
hibited a high recovery with an acceptable
precision. The recoveries of bromometh-
ane for the initial runs were in the range
of 200-250% showing a definite decreas-
ing trend down to approximately 90% by
Runs 10 and 11.
Using EPA Method 301 acceptance cri-
teria of precision <50% relative standard
deviation with a correction factor between
0.70 and 1.30, the following compounds
met the criteria for acceptability: 1,1,1-
trichloroethane, 1,1,2-trichloroethane, 1,1-
dichloroethane, 1,1-dichloroethene,
2,2,4-trimethylpentane, allyl chloride, ben-
zene, carbon tetrachloride, chloromethane,
n~hexane, methylene chloride, toluene, tri-
chlorofluoromethane, vinyl chloride, and
vinyl bromide. The three compounds that
did not meet Method 301 acceptance cri-
teria were bromomethane, 1,3-butadiene,
and dichlorodifluoromethane.
Conclusions
Recommendations
Based on the results of the sampling
and analysis of the samples collected us-
ing Method 0040 and the protocol pro-
vided in EPA Method 301, the following
conclusions are drawn:
Acceptable method precision (less
than 50% relative standard
deviation) was observed for all
compounds tested except di-
chlorodifluoromethane.
-------
The accuracy or bias of the method
as measured by the recovery of
the spiked compounds relative to
the theoretical amount spiked was
found not to be significantly differ-
ent for seven compounds (1,1,1 tri-
chloroethane, 1,1,2-trichloroethane,
2,2,4-trimethylpentane, benzene,
carbon tetrachloride, vinyl bromide
and vinyl chloride). These com-
pounds, therefore, do not require a
correction factor and the method is
acceptable for the determination of
these compounds from stationary
sources.
The bias of the method for eight
compounds was found to be sig-
nificantly different from zero (1,1-
dichloroethane, 1,1-dichloroethene,
allyl chloride, chloromethane, hex-
ane, methylene chloride, toluene
and trichlorofluoromethane). How-
ever, each bias could be corrected
by applying a correction factor to
the result. Correction factors be-
tween 0.70 and 1.30 are accept-
able following EPA 301 criteria.
Using the appropriate correction fac-
tor the method is acceptable for
these compounds.
The bias of the method for three
compounds (bromomethane, 1,3-
butadiene and dichlorodifluo-
romethane) was found to be signifi-
cantly different from zero and could
not be corrected through the use of
a correction factor that was within
the acceptable range of 0.70 to
1.30. Therefore, the method is not
acceptable for these compounds.
The total volume of sample col-
lected in the Tedlar® bags for the
dynamically spiked trains was de-
termined by summing the dry gas
meter volume and the volume of
spike gas added. This approach
was confirmed by calculating the
percent recovery of the deuterated
analogues of several of the target
compounds that were injected as a
liquid spike during the collection of
the samples.
The fundamentals of Method 0040
were acceptable.
The information in this document has
been funded wholly by the United States
Environmental Protection Agency under
EPA Contract Number 68-D4-0022 to East-
ern Research Group. It has been sub-
jected to Agency review and approved for
publication. Mention of trade names or
commercial products does not constitute
endorsement or recommendation for use.
References
1. Method 0040, in Test Methods for
Evaluating Solid Waste, Physical/
Chemical Methods, SW-848
Manual, 3rd ed. and updates, Docu-
ment No. 955-001-000001. Avail-
able from Superintendent of
Documents, U.S. Government Print-
ing Office, Washington, DC, No-
vember 1986.
2. Method 301, "Field Evaluation of
Pollutant Measurement Methods
from Various Waste Media," Fed-
era/ Register, Volume 57, Number
250, December 29, 1992, pp61998-
62002.
3. Method 8260, in Test Methods for
Evaluating Solid Waste, Physical/
Chemical Methods, SW-846
Manual, 3rd ed. and updates, Docu-
ment No. 955-001-000001. Avail-
able from Superintendent of
Documents, U.S. Government Print-
ing Office, Washington, DC, No-
vember 1986.
4. Method 5041, in Test Methods for
Evaluating Solid Waste, Physical/
Chemical Methods, SW-848
Manual, 3rd ed. and updates, Docu-
ment No. 955-001-000001. Avail-
able from Superintendent of
Documents, U.S. Government Print-
ing Office, Washington, DC, No-
vember 1986.
-------
James F. McGaughey, Joan T. Bursey, and Raymond G. Merrill are with Eastern
Research Group, Inc., Morrisville, NC 27560-2010.
Robert G. Fuerst is the EPA Project Officer (see below).
The complete report, entitled "Field Evaluation of EPA Method 0040 (Volatiles
Using Bags)," (Order No. PB98-133085, Cost: $28.00, subject to change) will
be available only from
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-605-6000
The EPA Project Officer can be contacted at
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use
$300
EPA/600/SR-98/030
------- |