EPA/600/A-96/068
Method Evaluation Study: The Application of SemiVOST to the
Nonhalogenated Semivolatile Organic Compounds from the Clean Air
Act Amendments
Merrill D. Jackson
National Exposure Research Laboratory, U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
James F. McGaughey, Raymond G. Merrill, and Joan T. Bursey
Radian Corporation, P. O. Box 13000, Research Triangle Park, North Carolina 27709
A laboratory study and three field studies were performed to evaluate the application of the SemiVOST (EPA
sampling Method 0010, EPA sample preparation Draft Method 3542, and EPA analytical Method 8270) to the
semivolatile halogenated organic compounds (approximately 80 compounds) listed in the Clean Air Act
Amendments (CAAA) of 1990. In these initial studies, PCBs, dioxins, and pesticides were excluded.
Subsequently, a laboratory study was performed to assess the feasibility of the application of the SemiVOST
methodology to the semivolatile nonhalogenated organic compounds listed in the CAAA (approximately 70
compounds). Several of the semivolatile organic compounds were eliminated from consideration as SemiVOST
analytes because they could not be analyzed successfully by gas chromatography/mass spectrometry (GC/MS),
reacted with other compounds in solution, or were insoluble in methylene chloride at the levels required to
perform dynamic spiking in the field. The remaining CAAA semivolatile organic analytes were grouped as
acid/neutrals and base/neutrals and evaluated in the field using the guidance of EPA Method 301 for
experimental design and statistical evaluation of the data. Quadruple SemiVOST trains were run in the field,
with dynamic spiking of the semivolatile analytes from a methylene chloride solution of either acid/neutral or
base/neutral compounds into two of the sampling trains. The bias and precision of the overall SemiVOST
methodology (sampling, sample preparation, and analysis) applied to each of the semivolatile organic analytes
were evaluated.
INTRODUCTION
A wide range of semivolatile organic analytes is encompassed in the SemiVOST sampling and analytical
methodology, which consists of the following components:
• Sampling Method: SW-846 Method 0010';
• Sample Preparation Method: SW-846 Proposed Method 35422: and
• Analytical Method: SW-846 Method 8270' (since the analytical procedure rather than the analyte list will
be used from Method 8270, the exact designation (i.e., 8270, 8270A, or 8270B) is irrelevant).
The SemiVOST is broadest in scope and applicability to semivolatile organic compounds, but the performance
of the methodology for a specific analyte is not known until the performance is established experimentally.
The parameters of concern are bias (accuracy) and precision (variability), since most criteria for acceptable
performance of a method are based on the bias and precision demonstrated for the entire method (sampling
through analysis) under field conditions.
Method evaluation studies for the halogenated semivolatile organic CAAA analytes have been performed at
three field sites in two source categories.3,4 The comparable studies have been initiated for the nonhalogenated
CAAA analytes. A preliminary laboratory study5 was performed to evaluate the ability of the nonhalogenated
1
-------
semivolatile organic compounds to be analyzed successfully, their stability and compatibility in methylene
chloride solution, and their ability to survive dynamic spiking into a SemiVOST train.
Because the following compounds exhibited reactivities, insolubility in methylene chloride, and chemical
interactions, they were excluded from further study in this program: hydroquinone, 1,4-phenylenediamine,
captan, 1,3-propane sultone, ethylene thiourea, hexamethylene-l,6-diisocyanate, catechol, 2,4-toluenediamine,
and maleic anhydride. The laboratory study indicated that the optimum grouping of analytes in order to avoid
chemical interactions in the spiking solutions was acid and neutral compounds in one solution, and basic and
neutral compounds in a second solution. The selection of acid, base and neutral groupings was somewhat
subjective. The phenols, cresols and phthalates were placed in the acidic group. The nitrogen-containing
(excluding nitro-compounds and quinone (mistaken for quinoline) made up the basic group. The neutral group
included hydrocarbons, oxygenated compounds, the chlorinated pesticides and quinoline.
EXPERIMENTAL APPROACH
The field test site selected was a coal-fired power plant owned and operated by a public utility. A presurvey
showed that none of the designated semivolatile organic analytes was present in the gaseous emissions at a level
above the method detection limits. A dynamic spiking level of 500 (ig was therefore planned. With 500 /*g
spiked into the train, virtually complete retention of the analyte on XAD-2®, and quantitative recovery should
be observed. This value is at approximately the middle of a standard Method 8270 calibration curve, and
should be readily analyzed even if only 50% recovery is obtained.
The Method 0010 sampling train was modified to allow simultaneous collection of samples in quadruplicate as
per EPA Method 3016, with liquid dynamic spiking of a solution of semivolatile organic compounds occurring
in the sampling train between the probe and heated filter. A quad probe, consisting of four heated borosilicate
glass probe liners mounted in one probe assembly, was used with four similar SemiVOST trains. Each of the
probe liners and spiking injection ports was individually heated and the temperatures were maintained at 130 ±
15°C. Flue gas temperatures and velocity measurements were monitored and the sample was collected as
closely to isokinetic conditions as possible. Because accurate determination of the background levels of organic
compounds in the gaseous emissions was of no interest to this program, the stack was not traversed.
A heated glass elbow equipped with a spiking injection port was used to connect the probes of the spiked trains
to the heated filter. The mass of solution delivered was used in the calculation of theoretical amount spiked for
the purpose of determining recovery. The liquid spike was maintained as a droplet at the tip of the glass-lined
stainless steel tubing so that the liquid spike could volatilize as it entered the gas stream and become a gaseous
spike at that point. The spiking liquid was never allowed to drip into the sampling line.
Clean sampling train components were assembled in the on-site mobile laboratory, with final assembly of the
trains after they were moved to the actual sampling location. Once they were assembled, the sampling trains
were leak-checked. Upon completion of the pre-test leak checks, heaters for spiking glassware, filter holders,
and probes were turned on. Sampling at a rate of approximately 0.50 ft3/min was initiated, and the syringe
pump for dynamic spiking was started after the gas flow had been established. At the end of the one-hour
sampling period, the syringe pumps were turned off and then the meter boxes were turned off. The sampling
trains were then leak-checked in the same manner used to perform the pre-test leak checks. Sampling trains
were disassembled into three sections: the spiking glassware/filter holder, XAD-2® module, and the impinger
train. Since the spiking point was after the probe, the probes could remain in place for the next quad sampling
run.
RESULTS AND DISCUSSION
Two sets of ten dynamic spiking SemiVOST sampling runs were performed in the field, using an Acid/Neutral
spiking solution to generate one set of samples and a Base/Neutral spiking solution for the second set. Results
for the nonhalogenated semivolatile compounds are shown in Table 1, using Method 301 statistical calculations
2
-------
and EPA QA/QC Handbook7 statistical calculations. The two modes of statistical calculations generally produce
comparable results when acceptance criteria are considered for a given analyte. Although there are more
factors to be considered in the overall Method 301 statistical calculation, in general analytes with a recovery of
70-130% and precision ^ 50% relative standard deviation will meet Method 301 acceptance criteria. Method
301 also requires a calculation of a correction factor, this factor must be between 0.70 and 1.30. The EPA
QA/QC Handbook acceptance criteria are slightly broader, with recovery of 50-150% and precision <;50%
relative standard deviation required to meet acceptance criteria.
For all analytes and both methods tested, Table 1 summarizes the experimental results for the field testing
at a coal-fired power plant. Of the 55 semivolatile organic compounds tested by the SemiVOST method, 36 met
Method 301 acceptance criteria and 45 met EPA QA/QC Handbook acceptance criteria. Analyte distribution
through the SemiVOST train generally follows a predicted path, with the most polar and water-soluble analytes
showing a significant component in the condensate and the least volatile compounds showing a significant
component on the filter. The major quantity of most of the analytes is collected on the XAD-2®.
The effect of an emissions matrix is not straightforward to predict, but for most semivolatile compounds,
laboratory performance is a reasonably reliable guide to field performance.
REFERENCES
1. "Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846 Manual, 3rd ed."
Document No. 955-001-0000001. Available from Superintendent of Documents, U. S. Government
Printing Office, Washington, D C. November, 1986.
2. Proposed Third Update to SW-846 Manual, Published for Public Comment Federal Register July, 1995.
3. Field Test of a Generic Method for Halogenated Hydrocarbons. U. S. Environmental Protection Agency.
EPA 600/R-93/101. NTIS PB93-212181.
4. Jackson, Merrill D., Joan T. Bursey, James F. McGaughey, Raymond G. Merrill, An Evaluation of the
SemiVOST Method for Halogenated Compounds at a Chemical Manufacturing Facility, Proceedings of
1995 EPA/AWMA Symposium of Toxic and Related Air Pollutants, , pp. 227-232, Research Triangle
Park, NC, May 16 -18, 1995.
5. Jackson, Merrill D., Joan T. Bursey, James F. McGaughey, Raymond G. Merrill, Application of VOST
and SemiVOST to nonHalogenated CAAA Compounds, Proceedings of 1995 EPA/AWMA Symposium of
Toxic and Related Air Pollutants, pp.233-240, Research Triangle Park, NC, May 16 -18, 1995.(NTIS
PB96-116884)
6. EPA Method 301. Protocol for the Field Validation of Emission Concentrations from Stationary Sources.
U. S. Environmental Protection Agency. EPA 450/4-90-0015. April, 1991.
7. Handbook. Quality Assurance/Quality Control (QA/QC) Procedures for Hazardous Waste Incineration.
EPA/625/6-89/023. January, 1990.
DISCLAIMER
The information in this document has been funded wholly by the United States Environmental Protection
Agency under contract 68-D4-0022 to Radian Corporation. It has been subjected to Agency review and
approved for publication. Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
3
-------
Table 1. Statistical Evaluation of Field Data Obtained from Ten Valid Quadruple SemiVOST
Sampling Runs with Dynamic Spiking of Title III Clean Air Act Analytes
(Ten Acid/Neutral Sampling Runs, Ten Base/Neutral Sampling Runs)
Semivolatile Organic
Compound
Mean Recovery
Meets Method 301
Acceptance Criteria?3
Meets EPA QA/QC
Handbook
Acceptance Criteria?4
Acids'
di-ji-butyl phthalate
46 ± 54%
No
No
bis(2-ethylhexyl)
phthalate
48 ± 23%
No
No
rQ-/g-cresol7
69 ± 14%
No
Yes
dimethyl phthalate
82 ± 17%
No
Yes
phenol
89 ± 9%
Yes
Yes
o-cresol
90 ± 15%
Yes
Yes
2,4-dinitrophenol
111 ±31%
Yes
Yes
4-nitrophenol
114 ± 31%
Yes
Yes
4,6-dinitro-o-cresol
122 ± 14%
Yes
Yes
Bases'
quinone6
2 ± 438%
No
No
hexamethylphos-
phoramide
14 ± 118%
No
No
trifluralin
27 ± 41%
No
No
dimethylaminoazo-
benzene
31 ± 51%
No
No
3,3'-
dimethoxybenzidine
37 ± 38%
No
No
o-anisidine
39 ± 39%
No
No
o-toluidine
56 ± 30%
No
Yes
benzidine
65 ± 119%
No
No
N, N ,-dimethy laniline
67 ± 24%
No
Yes
aniline
70 ± 24%
No
Yes
4,4'-methylene
bis(2-chloroaniline)
89 ± 36%
Yes
Yes
3,3' -dimethylbenzidine
92 ± 44%
Yes
Yes
4
-------
Table 1. Continued
Semivolatile Organic
Compound
Mean Recovery
Meets Method 301
Acceptance Criteria?3
Meets EPA QA/QC
Handbook
Acceptance Criteria?4
N,N,diethylamide
95 ± 19%
Yes
Yes
carbaryl
99 ± 19%
Yes
Yes
ethyl carbamate
103 ± 14%
Yes
Yes
caprolactam
114 ± 12%
Yes
Yes
N-nitrosomorpholine
116 ± 12%
Yes
Yes
N-nitrosod imethy 1-
amine
117 ± 13%
Yes
Yes
propoxur
123 ± 12%
Yes
Yes
2-acetylaminofluorene
147 ± 23%
No
Yes
Neutrals2
styrene oxide
0.5 ± 1481%
No
No
phthalic anhydride
5.3 ± 144%
No
No
methoxychlor
73 ± 19%
No7
Yes
toluene
76 ± 11%
No7
Yes
m-/g-xylene6
79 ± 12%
Yes
Yes
quinoline5
80 ± 19%
Yes
Yes
styrene
84 ± 10%
Yes
Yes
o-xylene
85 ± 11%
Yes
Yes
1,4-dioxane
87 ± 11%
Yes
Yes
cumene
88 ± 11 %
Yes
Yes
ethylbenzene
89 ± 12%
Yes
Yes
parathion
89 ± 28%
Yes
Yes
isophorone
93 ± 12%
Yes
Yes
acetophenone
96 ± 12%
Yes
Yes
naphthalene
96 ± 11%
Yes
Yes
dibenzofuran
100 ± 12%
Yes
Yes
dichlorvos
101 ±18%
Yes
Yes
DDE
102 ± 15%
Yes
Yes
-------
Table 1. Continued
Semivolatile Organic
Compound
Mean Recovery
Meets Method 301
Acceptance Criteria?3
Meets EPA QA/QC
Handbook
Acceptance Criteria?4
4-nitrobiphenyl
102 ± 14%
Yes
Yes
heptachlor
103 ± 12%
Yes
Yes
biphenyl
103 ± 12%
Yes
Yes
lindane
104 ± 12%
Yes
Yes
nitrobenzene
109 ± 12%
Yes
Yes
2,4-dinitrotoluene
109 ± 12%
Yes
Yes
methyl isobutyl ketone
112 ± 11%
Yes
Yes
chlordane
142 ± 16%
Yes7
Yes
'Values represent the mean from ten complete quad sampling runs with dynamic spiking, two spiked trains and
two unspiked trains.
2Values represent the mean from twenty complete quad sampling runs with dynamic spiking, two spiked trains
and two unspiked trains. Neutral compounds were spiked with both the Acid and the Bases, and all neutral
data are included in the composite values.
3EPA Method 301 acceptance criteria include recovery of 70 - 130%, with a precision s 50 % relative standard
deviation and a correction factor between 0.70 and 1.30.
4EPA QA/QC Handbook acceptance criteria include recovery of 50 - 150%, with a precision ^ 50 % relative
standard deviation.
5Quinoline was placed in Neutral solution rather than Basic solution because of confusion of name with
quinone.
'Listed together in the table because of chromatographic coelution.
7The correction factors (CF) for toluene and methoxychlor were 1.31 and 1.40; the CF for chlorodane was
0.71, however this was due to high variability of the unspiked trains.
6
-------
t
TECHNICAL REPORT DATA
1. REPORT NO.
EPA/600/A-96/068
2.
IIIIIUIIIIIIIIIIIIIIIIIIIII
PB97-122725
4. TITLE AND SUBTITLE
Method Evaluation Study: The Application of VOST to the
Nonhalogenated Semivolatile Organic Compounds from the Clean
Air Act Amendments
5.REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTHORIS)
Merrill D. Jackson , James F. McGaughey, Raymond G. Merrill, and Joan
T. Bursey
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
P. 0. Box 13000
Research Triangle Park, North Carolina 27709
10.PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68D40022
12. SPONSORING AGENCY NAME AND ADDRESS
National Exposure Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13.TYPE OF REPORT AND PERIOD COVERED
Symposium Paper, 9/94-6/96
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
-A laboratory study and two field studies were performed to evaluate the application of the Volatile Organic
Sampling Train (VOST, EPA Method 0030) to the volatile halogenated organic compounds (approximately
35 compounds) listed in the Clean Air Act Amendments (CAAA) of 1990. Subsequently, a laboratory
study was performed to assess the feasibility of the application of the VOST methodology to the volatile
nonhalogenated organic compounds listed in the CAAA (approximately 20 compounds).^Several of the
volatile organic compounds were eliminated from consideration as VOST analytes because they could not
be analyzed successfully by gas chromatography/mass spectrometry (GC/MS), and oth'ers were eliminated
because of poor analytical system response. The remaining CAAA volatile organic,analytes (benzene,
toluene, carbon disulfide, 2,2,4-trimethylpentane, and hexane) were tested in thefield using the guidance
of EPA Method 301 for experimental design and statistical evaluation of the data^Quadruple VOST trains
were run in the field, with dynamic spiking of the volatile analytes from a certified cylinder into two of the
sampling trains. The bias and precision of the VOST sampling and analytical methodology applied to each
of the volatile organic analytes were evaluated.^-
{ f
1 7. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED TERMS
c. COS AT I
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21.NO. OF PAGES
20. SECURITY CLASS (This Page)
Unclassified
22. PRICE
------- |