EPA/600/A-93/106
Using Method 301 to Validate Sampling and Analytical Methods
for Selected CAAA Compounds
Merrill D. Jackson, Source Methods Research Branch, AREAL/USEPA, Research Triangle Park, North
Carolina 27711, R. G. Merrill and J. T. Bursey, Radian Corporation, Research Triangle Park, North Carolina
27709
ABSTRACT
Stationary source sampling and analysis methods can be validated for a set of analytes and stack
conditions using EPA Method 301 before these methods are jsed to con- ply with the requirements of the Clean
Air Act Amendments of 1990. EPA Method 301 describes :i protocol designed to provide sufficient numbers
of samples to determine the precision and bias of the method. Most method validations require spiking of
analytes under field conditions by gaseous dynamic spiking int;i two of foi.r sampling trains on a quadruple probe
(used with the Volatile Organic Sampling Train, VOST), liquid spiking into two of four sampling trains using
a quadruple probe (Semivolatile Organic Sampling Train, SemiVOST), and spiking of various analytes into
impingers or onto filters (used with Method 29 and Method 0)11;. A quadruple sampling probe with four trains
is usually used to minimize the number of sampling runs needed to provide sufficient replicates for statistical
calculations. New method validations require initial laboratory lesting prior to field validation to demonstrate
the feasibility of the application of the proposed sampling and analytical method for a particular analyte.
Laboratory testing establishes experimental parameters such as stability, analytical method performance, sample
preparation procedures, spiking conditions, and precision and accuracy o: analysis. Successful laboratory testing
supports a full field validation to evaluate the applicability of a gi\en method to a particular analyte. Laboratory
preparation for a field validation of the VOST and Semi VOST for halouenated organic compounds from the
Clean Air Act has been performed using Method 301 techniques.
INTRODUCTION
The Clean Air Act Amendments (CAAA) of 199*i. Title III, present a need for stationary source
sampling and analytical methods for a list of 189 analytes. For volatile and semivolatile organic compounds, the
U. S. Environmental Protection Agency (EPA) has used the Volatile Organic Sampling Train (VOST, Method
0030)(1) and the Semivolatile Organic Sampling Train (SemiVONT, Method (X)10)(1) for sampling at stationary
sources. Corresponding analytical methods are Method 50-Mi or 5041 for VOST, and Method 8270 for
SemiVOST(l). The VOST and SemiVOST sampling and analytical methodology has been used extensively for
volatile and semivolatile organic compounds, but complete method validation data are available for only a few
compounds(2,3). Validation defines the performance of a method under a given set of conditions: that is,
validation determines the precision and bias of the method when the method is applied to a particular compound
at a given stationary source.
Several preliminary steps are necessary before validation of a sampling and analytical method can be
performed in the field. In this program, the ultimate goal was field validation of the VOST and SemiVOST for
the halogenated organic compounds listed in Title III of the CAAA. Neither 2,3,7,8-tetrachlorodibenzodioxin,
dibenzofurans, nor the polychlorinated biphenvls were included in this program, because EPA has specialized
methods for measuring these compounds in stationary source emissions. The VOST and SemiVOST methods
specify boiling point criteria for analytes: VOST analytes must have boiling points <100°C, SemiVOST analytes
have boiling points of 100°C or above. Table I shows the analytes that were selected for testing in this program,
with their boiling points. Halogenated organic compounds with boiling points in the range of 100 - 135°C were
designated for testing in both VOST and SemiVOST, to establish their performance in both methods.
In preparation for a full-scale field validation, the following laboratory tests were performed:
• Halogenated organic compounds from Title III of the Clean Air Act Amendments of 1990 were
subjected to analysis under the conditions used for VOST and SemiVOST. Chromatographic
retention times, reference mass spectra, and a quantisation scheme for the analytes were

-------
developed.
Recovery of analytes from sorbents (thermal desorption for VOST, liquid extraction for
SemiVOST) was determined.
Analytical method detection limits were determined for VOST and SemiVOST.
Dynamic spiking apparatus for spiking of liquid solutions into the SemiVOST and gases into
the VOST was designed, fabricated, and tested
Dynamic spiking apparatus and procedures wer.- statistically evaluated using quadruple VOST
and SemiVOST trains prior to use in the field at a test site.
EXPERIMENTAL PROCEDURES AND RESULTS
.For VOST analytical determinations, the chromatographic column wzis DB-624, 0.53 ID, 3 /i film thickness,
with a program of 0°C for 4 min, then 6°C/min to 200°C. The (iC/MS system was a Finnigan-MAT 4500
GC/MS. Other instrumental conditions followed the recommendations ol SW-846 Method 5041 (1). Reference
mass spectra were generated for the analytes. relative retention limes were determined, and primary and
secondary quantitation ions were assigned (secondary ions t;> he used for quantitation only in case of
chromatographic/mass spectrometric interference with the primary ion) The reference spectra, quantitation
scheme, and retention times are available in the complete EPA report(4) describing this program. All of the
candidate VOST targets except bis(chloromethvl) ether, chlorometlivl methyl ether, and epichlorohydrin were
analyzed successfully using the VOST procedure.
The VOST sorbent tubes consist of a front tube containing Tenax GC", with a back tube consisting of Tenax
GC® and petroleum-based charcoal. These tubes are used as a pair in the VOST train. To determine recovery
of the candidate analytes from the VOST tubes, pairs of clean VOST tubes were spiked with a methanolic
solution of the compounds of interest using the flash evaporation technique described in Method 5041 to spike
approximately 50 ng of each analyte onto the paired tubes. Internal standards were spiked into the water of the
purge flask, and spiked tubes were desorbed through the purge flask as a pair. Unspiked tubes were analyzed
as a blank. The analytical system was calibrated by using the metnancl c solutions to spike the purge water,
according to Method 8240. The efficiency of desorption of test compounds from the VOST tubes was
determined by comparing concentrations determined by desorption from i he spiked tubes to the amount spiked
onto the tubes. Five replicate determinations were performed, and the results were analyzed statistically (Table
II). All of the compounds reported in Table II showed acceptable recoveries.
Method detection limits were determined for the VOST analytical method (Method 5041) using a procedure
specified in the Federal Register(5). The method detection limit lor the candidate halogenated compounds from
the Clean Air Act List was estimated to be approximately 10 - 20 ng, Dased on previous experience with the
methodology. Paired VOST tubes were spiked with 20 ng of each analyte in a methanolic solution, using the
flash vaporization technique described in Method 5041. VOST method detection limits for CAAA analytes are
shown in Table III. The method detection limits shown in Table III are acceptable for further evaluation of the
analytes.
Upon completion of the laboratory experiments, a VOST dynamic spiking apparatus was designed and
constructed, and quadruple VOST trains were set up in the laboratory to evaluate the dynamic spiking system
which used a certified cylinder gaseous mixture of the compounds of interest, with the spike in the VOST trains
occurring immediately prior to the entry of the gaseous matrix into the sorbent tubes. Initial evaluation of the
accuracy and reproducibility of the spiking system showed that mass flow controllers used in transmission of
sample from the cylinder to the sampling trains were not functioning properly. The dynamic spiking system was
redesigned to use fine metering valves, using bubble flowmeters to measure flow before and after sampling, and
heat tracing at 130°C of the sample transmission lines from the c:ylinccr regulator to the point of spiking.

-------
Preliminary results demonstrated that, with the modifications made tc the gaseous dynamic spiking system,
accurate and reproducible spiking of target CAAA analytes could be performed on quad VOST trains.
For SemiVOST analytes (CAAA halogenated compounds with boiling points >100°C), the
chromatographic column used was a DB-5, 0.32 mm ID, 30 m length, 1.0 /i film thickness. The column was
programmed at 35°C for 4 min, then 10°C/min to 295°C. The other instrumental conditions followed Method
8270. Reference mass spectra, primary and secondary quantitation ions., and retention times are found in the
complete EPA report on this program(4). Only chloroacetic acid could not be chromatographed and analyzed
successfully; all other candidate analytes were amenable to the analytical methodology.
Recoveries of the compounds from the XAD-2® sorbent used in ;he SemiVOST train were determined
by spiking prepared XAD-2® sampling cartridges with a methylene chloride solution of the appropriate CAAA
halogenated compounds, at a level of approximately 250 ng each with a firal sample volume of 5 mL. Surrogate
compounds from Method 8270 were also spiked onto the XAD-2'® to monitor the performance of the analytical
methodology. Recoveries from the XAD-2® are shown in Table IV. Method detection limits were determined
by spiking prepared XAD-2® with approximately 50 /ig of the appropriate analytes, following the Federal
Register Method(5). Method detection limits are shown in Table V. Problems with calibration were sometimes
encountered with bis(chloromethyl) ether, epichlorohvdrin, anil 3,3 dichlorobenzidine. Recoveries of
hexachlorobutadiene, hexachlorobenzene. pentachloronitrobenzene. chlorobenzilate, and 3,3'-dichlorobenzidine
were outside the range of 50-150%, which is acceptable recover. according to the method. Analytical method
detection limits for the majority of compounds would be in the : amie o 1 /ig/cf.
To evaluate the entire SemiVOST meihod and to prepare lor validation of the methods in the field, a
liquid dynamic spiking apparatus was designed and constructed (4). Ouadruple SemiVOST trains were set up
to test the ability to spike the compounds of interest accurately ind reproducibly. The liquid dynamic spiking
system incorporated a constant flow syringe pump with Teflon * lmes to a glass-lined stainless steel needle
introduced into the sampling lines just behind the probe and immed.ately prior to the heated filter. Pump flow
was adjusted to provide approximately 10 mL of spiking solution over a SomiVOST sampling interval of 2 hours,
with a sampling flow of 0.5 cfm. Dynamic spiking temperature and tow rates were carefully regulated to
provide a droplet of spiking solution at the beveled tip of a das-lined stainless steel needle. The drop could
not be allowed to evaporate completely, nor to drop to the heated glass surface of the train probe. For a
statistical evaluation of the spiking procedures, a Latin Square experimental design was used. In the Latin
Square, four trains, four runs, and four spiking levels were used, under laboratory conditions where nitrogen was
used as the diluting gas to simulate SemiVOST stationary source omissions sampling. Recoveries of spiked
compounds from the Latin Square experiments are shown in Table VI.
CONCLUSIONS
Gas chromatographic retention times, reference mass spectra, and quantitation ions were determined for the
candidate halogenated compounds from the Title III of the Clean Air Act Amendments of 1990. Of the 45
compounds considered as candidates for the methodology, only four could not be chromatographed successfully:
chloroacetic acid, bis(chloromethvl) ether, chloromethyl methyl ether, arid epichlorohydrin. Sorbent recoveries
were determined, and analytical method detection limits were determined. The complete set of laboratory
experiments indicated that a reasonable probability of success in a full field validation could be expected for 67%
of the compounds.
REFERENCES
1.	Test Methods for Evaluating Solid Wastc-Phvsical/Chcmical Method:;. EPA Report SW-846, Third Edition,
Washington, DC.
2.	R.G. FUERST, TJ. LOGAN, M.R. MIDGETT et al.,"Validation Studies of the Protocol for the Volatile
Organic Sampling Train," JAPCA 37(4): 388 (1987).

-------
3.	J.H. MARGESON, J.E. KNOLL, M. R. MIDGETT, et aI.,"An Evaluation of the Semi-VOST Method for
Determining Emissions from Hazardous Waste Incinerators," J A PC/, 37(9) 1067 (1987).
4.	"Laboratory Validation of VOST and SemiVOST for Halogenated Hydrocarbons from the Clean Air Act
Amendments List," EPA Report, Research Triangle Park, NC 27709.
5.	Code of Federal Regulations, 40CFR, Part 136, Appendix B. Wl.
6.	"Protocol for the Field Validation of Emission Concentrations from S:ationary Sources," EPA Report 450/4-
90-015, Research Triangle Park, NC, Feb. 1991.
The information in this document has been funded by the United States Environmental Protection
Agency under contract 68-D1-0010 to Radian Corporation, It has bei:n subjected to Agency review and approved
for publication. Mention of trade names or commerical products does not constitute endorsement or
recommendation for use.

-------
Table I. CAAA Halogenated Compounds Investigated.
Compound
Boiling point.°C
VOST
SemiVOST
Allyl chloride
44 - 46
X

bis(Chloromethyl) ether
lOo
X
X
Carbon Tetrachloride
77
X

Chlorobenzene
132
X
X
Chloroform
60.5 - 61.5
X

Chloromethyl methyl ether
55 - 57
X

Chloroprene
59.4
X

1,3-Dichloropropene
105 - 106/730mrr;
X
X
Epichlorohydrin
115 - ir
X
X
Ethyl chloride
12'
X

Ethylene dibromide
131 - 132
X
X
Ethylene dichloride
83
X

Ethylidene dichloride
57
X

Methyl bromide
4*
X

Methyl chloride
¦24.2*
X

Methyl chloroform
74 - 76
X

Methylene chloride
*
oc
rh
X

Methyl iodide
41 - 43
X

Propylene dichloride
95 - 96
X

T etrachloroethylene
121
X
X
1,1,2-T richloroethane
111) - 115
X
X
T richloroethylene
86.')
X

Vinyl bromide
16/750mm*
X

Vinyl chloride
•13.4-
X

Vinylidene chloride
30-32
X

Benzotrichloride
219 •• 223

X
Benzyl chloride
177 - 181

X
Bromoform
150 - 151

X
Chloroacetic acid
18')

X
2-Chloroacetophenone
244 - 245

X
Chlorobenzilate
147

X
l,2-Dibromo-3-chloropropane
1%

X
1,4-Dichlorobenzene
173

X
3,3'-Dichlorobenzidine
mp = 165

X
Dichloroethyl ether
65-67/15mm

X
Hexachlorobenzene
323 - 326

X
Hexachlorobutadiene
210 -220

X
Hexachlorocyclopentadiene
239

X
Hexachloroethane
186

X
Pentachloronitrobenzene
328

X
Pentachlorophenol
309.5

X
1,1,2,2-T etrachloroethane
147

X
1,2,4-T richlorobenzene
214

X
2,4,5-T richlorophenol
248/740mm

X
2,4,6-T richlorophenol
24/.

X
* Below the recommended lower boiling point limit of 30°C for VOST
K~

-------
Table II. Recoveries of Compounds from VGST Sorbents.
(Tenax GC® - Tenax GC®/petroleum-based charcoal)
Compound
Mean*
SD**
%cov***
Ethyl chloride
95.8
10.73
11.20
Ethylene dichloride
123.0
5.61
4.56
Methyl iodide
108.4
5.73
5.28
Allyl chloride
127.2
6.91
5.43
Methylene chloride
101.6
2M
.2.84
Ethylidene dichloride
108.8
4.32
3.97
Chloroprene
104.2
4.49
4.31
Methyl chloride
101.2
8.20
8.10
Chloroform
117.4
5.77
4.92
Carbon tetrachloride
108.4
16.22
14.97
1,2-Dichloroethane
95.8
5.93
6.19
Vinyl Chloride
90.4
lO.Sf
12.01
T richloroethylene
110.0
7.56
6.88
Propylene dichloride
98.0
9.33
9.52
cis- 1,3-Dichloropropene
109.0
15.S1
14.59
trans- 1,3-Dichloropropene
96.6
17.39
18.00
1,1.2-T richloroethane
106.4
14.58
13.71
T etrachloroethylene
111.6
7.50
6.72
Ethylene dibromide
97.0
14.42
14.86
Methyl bromide
97.4
9.53
9.78
Chlorobenzene
94.2
13.72
14.56
Vinyl bromide
110.8
10.30
9.30
Methyl chloroform
103.4
12.70
12.28
'Average of 5 values
*'Standard deviation
***%Coefficient of variation

-------
Table III. VOST Method Detection Limits.
Compound
Mean*
SD<:"
MDLCnel
Ethyl chloride
34.20
8 XI
24.59
Ethyl dichloride
24.90
6..2
17.26
Methyl iodide
30.50
7.-1
20.05
Allyl chloride
29.80
5.:.4
14.49
Methylene chloride
42.10
8.62
24.32
Ethylidene dichloride
31.90
6.3 L
17.81
Chloroprene
29.80
7.<18
21.10
Methyl chloride
92.64
203)
46.30
Chloroform
36.40
5.-W
15.46
Carbon tetrachloride
30.30
5/ki
15.40
1,2-Pichloroethane
33.30
7.07
19.96
Vinyl Chloride
31.90
7.68
21.67
T richloroethylene
28.20
4..S4
12.81
Propylene dichloride
30.60
5.58
15.75
cis- 1,3-Dichloropropene
31.60
5.5()
15.69
trans- 1,3-Dichloropropene
31.10
5.(i8
16.04
1,1,2-T richloroethane
32.80
5.92
16.71
T etrachloroethylene
29.30
5.42
15.28
Ethylene dibromide
29.80
5.75
16.22
Methyl bromide
43.70

28.74
Chlorobenzene
29.80
4.64
13.08
Vinyl bromide
30.60
6.£W 1
18.05
Methyl chloroform
43.80
7.86
22.16
* Average of 10
** Standard deviation
*** Minimum detection level
n

-------
Table IV. Recoveries of Compounds from SemiVOST Sorbent.
(XAD-2® resin)
Compound
bis(Chloromethyl) ether
Epichlorohydrin
cis-l,3-Dichloropropene
trans- 1,3-Dichloropropene
1,1,2-T richloroethane
Ethylene dibromide
T etrachloroethylene
Chlorobenzene
Bromoform
1,1,2,2-Tetrachloroethane
Dichloroethyl ether
1,4-Dichlorobenzene
Benzyl chloride
Hexachloroethane
l,2-Dibromo-3-chloropropane
1.2.4-T	richlorobenzene
Hexachlorobutadiene
Benzotrichloride
Chloroacetophenone
Hexachlorocyclopentadiene
2,4,6-T richlorophenol
2.4.5-T	richlorophenol
Hexachlorobenzene
Pentachlorophenol
Pentachloronitrobenzene
Chlorobenzilate
3,3'-Dichlorobenzidine
•Average of five values
*'Standard deviation
*** % Coefficient of variation
Mean*	SD** %COV**'1
59.3
8.10
13.67
75.2
11.10
14.76
71.0
10.46
14.74
79.4
12.01
15.13
78.8
9.98
12.67
89.2
12.56
14.08
61.1
7.66
12.20
96.6
12.10
12.52
80.8
11.3(1
13.99
102.0
14.05
13.78
104.4
11.80
11.30
95.0
12.43
13.08
103.2
13.08
12.68
87.4
12.46
14.26
92.0
13.2^
14.42
90.6
13.35
14.74
47.8
6.42
13.43
76.8
11.80
15.36
141.6
21.43
15.14
53.0
9.51
17.95
93.8
15.16
16.16
108.2
15.24
14.08
45.8
5.63
12.29
69.8
10.55
15.11
38.0
4.58
12.06
47.6
6.88
14.45
275.0
55.83
20.31

-------
Table V. SemiVOST Method Detection Limits.
Compound
ue/mL
Total us
bis(Chloromethyl) ether
11.4
57.0
Epichlorohydrin
9.8
49.0
cis- 1,3-Dichloropropene
5.8
29.0
trans- 1,3-Dichloropropene
6.5
32.5
1,1,2-T richloroethane
9.0
45.0
Ethylene dibromide
10.7
53.5
T etrachloroethylene
13.4
67.0
Chlorobenzene
9.5
' 47.5
Bromoform
10.6
53.0
1,1,2,2-Tetrachloroethane
8.2
41.0
Dichloroethyl ether
11.0
55.0
1,4-Dichlorobenzene
12.9
64.5
Benzyl chloride
12.0
60.0
Hexachloroethane
10.9
54.5
l,2-Dibromo-3-chloropropane
12.6
63.0
1,2,4-T richlorobenzene
13.1
65.5
Hexachlorobutadiene
15.7
78.5
Benzotrichloride
12.7
63.5
Chloroacetophenone
13.9
69.5
Hexachlorocyclopentadiene
14.5
72.5
2,4,6-T richlorophenol
11.6
58.0
2,4,5-T richlorophenol
16.5
58.0
Hexachlorobenzene
13.4
67.0
Pentachlorophenol
30.7
153.5
Pentachloronitrobenzene
13.0
65.0
Chlorobenzilate
15.6
78.0
3,3'-DichIorobenzidine
19.3
96.5

-------
Table VI. Recoveries of SemiVOST Compounds from the Latin Square Experiments
Compound
Averaee*
SD**
NL*_I
bis(Chloromethyl) ether
18.28
9.22
14
Epichlorohydrin
75.20
24.11
14
cis-l,3-Dichloropropene
21.90
6.55
14
trans-l,3-Dichloropropene
20.35
5.8(1
14
1,1,2-T richloroethane
53.13
14.82
14
Ethylene dibromide
66.31
14.56
14
T etrachloroethylene
49.68
14.48
14
Chlorobenzene
75.%
13.46
14
Bromoform
99.27
22.25
14
1,1,2,2-Tetrachloroethane
81.05
12.77
14
Dichloroethyl ether
75.78
11.99
14
1,4-Dichlorobenzene
68.16
10.90
14
Benzyl chloride
78.72
20.43
14
Hexachloroethane
85.43
35.16
14
l,2-Dibromo-3-chloropropane
66.24
6.91
14
1,2,4-T richlorobenzene
58.20
10.94
14
Hexachlorobutadiene
58.34
10.69
14
Benzotrichloride
67.02
16.58
14
Chloroacetophenone
79.64
18.03
14
Hexachlorocyclopentadiene
513.04
245.26
14
2,4,6-T richlorophenol
45.61
16.30
14
2,4,5-T richlorophenol
52.69
37.98
14
Hexachlorobenzene
32.85
18.35
14
Pentachlorophenol
8.93
10.50
14
Pentachloronitrobenzene
38.24
20.66
14
Chlorobenzilate
43.63
35.49
14
3,3'-Dichlorobenzidine
86.42
165.82
14
Four quadruple runs were performed (total of 16 samples); two set? of results were rejected as outliers, leaving
14 samples.
•Standard deviation
** Number of values
10

-------