United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/S4-86/046 Apr 1987
oEPA Project Summary
Laboratory and Field Evaluation
of the Semi-VOST Method
J. Margeson, J Bursey, J. Steger, M. Palazzolo, D. Benson, J. Homolya,
R. McAllister, J. McGaughey, and D. Wagoner
Laboratory studies and a second field
evaluation have been completed to
assess the performance of the Semi-
Volatile Organic Sampling Train (Semi-
VOST) method for measuring concen-
trations of principal organic hazardous
constituents (POHCs) with boiling
points greater than 100°C emitted from
hazardous waste incineration. The
Semi-VOST Methodology was tested
initially through a series of laboratory
experiments and a field test at a
hazardous waste incinerator. In the first
field test method precision and bias
were shown to be compound specific.
Laboratory studies demonstrated that
method precision and bias were related
to the distribution and recovery of
POHCs throughout the components of
the sampling train. Distribution and re-
covery of a POHC are related to boiling
point, water solubility, and chemical
functionality.
Laboratory experiments have demon-
strated that, since basic organic com-
pounds form salts by reaction with the
acidic and wet environment found in
stack emissions sampled with the
Semi-VOST method, these compounds
can be recovered by appropriate treat-
ment of aqueous train components. A
screening procedure was developed to
select the extraction solvent and to
estimate the analytical recovery of the
test compounds from acidic water con-
densate using the sample preparation
methodology.
A second field test was performed at
a hazardous waste incinerator using five
gaseous deuterated compounds to
dynamically spike four simultaneously
operating semi-VOST trains. Estimates
of method precision and bias are given
for ds-pyridine, ds-toluene, d5-chloro-
benzene, d10-o-xylene, and d2-1,1,2,2-
tetrachloroethane. Results of distribu-
tive volume experiments to compare
chromatographically determined com-
pound retention volumes with values
determined in the field are presented.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering in-
formation at back).
Introduction
The Solid Waste Disposal Act, as
amended by the Resource Conservation
and Recovery Act of 1975 (RCRA), re-
quires that the U.S Environmental Pro-
tection Agency (EPA) establish a national
regulatory program to ensure that haz-
ardous wastes are managed in a manner
which does not endanger human health
or the environment. The statute requires
EPA to promulgate performance stan-
dards for hazardous waste management
Included in the regulations that have
been promulgated are provisions for
waste disposal by incineration and re-
quirements that hazardous waste incin-
erators be so operated that the principal
organic hazardous constituents (POHCs)
are destroyed and/or removed with a
minimum efficiency of 99.99 percent To
determine destruction and removal ef-
ficiency (ORE), EPA has designated the
Semi-Volatile Organic Sampling Train
(Semi-VOST) Method as the method for
measuring flue gas concentrations of
POHCs with boiling points greater than
100°C.
-------
The Quality Assurance Division of the
EPA Environmental Monitoring Systems
Laboratory (EMSL) has responsibility for
evaluating and standardizing EPA source
test methods. Under contract to EMSL,
Radian Corporation is providing technical
assistance in evaluating the Semi-VOST
Method. The objective of the Semi-VOST
program is to provide the data necessary
to determine the bias, precision, applica-
bility, and limitations of the method.
The technical approach used to collect
these data is a multi-task effort involving
literature, laboratory, and field studies.
An initial field test, which utilized a quad-
train approach, provided preliminary data
on method bias and precision. As a result
of the recommendations developed from
the first field study, additional laboratory
studies and a second field test at a
hazardous waste incinerator were
performed.
The experimental design of the labora-
tory evaluations and the second semi-
VOST test was based on experience
gained from the initial test and data
gathered from laboratory studies. The
second test again utilized a quad-train to
evaluate the method performance by
sampling flue gas from a full-scale in-
cinerator while spiking test compounds
into the sampling trains. The objectives
of the program were to:
• design and conduct an experimental
laboratory program to evaluate com-
pound stability in the extraction
solvent and the effect of water
solubility and compound functionality
on the recovery of selected test
compounds;
• develop a screening procedure for
Appendix VIII water-soluble com-
pounds to provide information for
selecting an extraction solvent and
for estimating recovery of water-
soluble compounds from the acidic
water condensate;
• design, construct, and evaluate the
performance of a dynamic spiking
system using deuterated test
compounds;
• carry out a second field test at a
hazardous waste incineration facility
to provide additional data on method
precision and bias for additional test
compounds; and
• compare retention volumes deter-
mined in the laboratory for selected
test compounds on XAD-2® resin
with values determined in the field
by a series of distributive volume
experiments sampling flue gas from
a hazardous waste incinerator.
Procedures
In the application of the Semi-VOST
method, gaseous and paniculate compo-
nents are isokinetically withdrawn from
an emission source and collected in a
multicomponent sampling train. Key
elements of the train include a high-
efficiency particulate filter and a packed
bed of a porous polymeric adsorbent resin
(XAD-2®). The filter is used to separate
stack gas particles from gaseous sub-
stances which are then adsorbed onto
XAD-2®. Following sample collection, the
train components are extracted with sol-
vent, the extracts concentrated, and
specific compounds are identified/quanti-
fied using high resolution gas chromato-
graphy coupled with low resolution mass
spectrometry. After returning the samples
to the laboratory, all fractions are spiked
with surrogate standards prior to extrac-
tion, concentration, and analysis. Surro-
gates provide a measure of the efficiency
of the sample preparation procedures.
The surrogate compounds should be
stable isotopically-labeled analogs of the
compounds of interest or compounds that
would exhibit properties similar to the
compounds of interest, be easily chro-
matographed, and not interfer with the
analysis. Potential surrogate spiking
compounds include: deuterated ethyl-
benzene, naphthalene, chrysene, phenol,
nitrobenzene, chlorobenzene and toluene.
Previous evaluation of the Semi-VOST
method has shown that method precision
and bias are compound specific. Labora-
tory experiments demonstrated that
method precision and bias are related to
the distribution and recovery of a par-
ticular POHC throughout the components
of the sampling train. Sample distribution
and recovery of a POHC are related to
boiling point, water solubility, and chemi-
cal functionality.
With these parameters in mind the
following laboratory experiments were
performed prior to the second field test in
order to:
• Define XAD-2® adsorbent resin in-
teractions and/or analytical proce-
dure interferences responsible for
poor and/or variable recoveries of
particular POHCs.
• Evaluate method interferences due
to interaction with or in the solvent
matrix by determining the stability
of aniline, pyridine, phenol, and
resorcinol in the extraction solvent
(methylene chloride).
• Develop a method for screening
Appendix VIII POHCs to identify in-
teractions with Semi-VOST analyti-
cal procedures.
• Determine the retention volumes of
toluene, 1,1,2,2-tetrachloroethane,
chlorobenzene, pyridine, and q-
xylene on XAD-2®resin to estimate
breakthrough volumes for field dis-
tributive volume experiments.
The second Semi-VOST field evaluation
was designed based on experience and
data gained from the initial field and
laboratory studies. Both the first and
second field tests utilized a Semi-VOST
train in a quad train configuration to
evaluate method performance by sam-
pling flue gas from a full-scale incinerator
while spiking test compounds into the
sampling trains. The objectives of the
second field test were to: 1) provide addi-
tional data on method bias and precision,
2) compare laboratory retention volume
data for selected compounds to field data,
and 3) evaluate the stability of selected
compounds spiked onto the XAD-2®resin
by both dynamic gaseous spiking and
liquid spiking.
Results and Discussion
Toluene, 1,1,2,2-tetrachloroethane, and
chlorobenzene were included in the first
field test. Pyridine was selected based on
its potential to form salts in the presence
of hydrochloric acid. o-Xylene was added
due to its large retention volume on XAD-
2® resin and its use as an internal stan-
dard in the Semi-VOST laboratory experi-
ments. To assess any interferences due
to solvent interactions, the stabilities of
aniline, pyridine, phenol, and resorcinol
in methylene chloride were evaluated.
The concentrations of the four test com-
pounds remained within 10 percent of
the original value for up to 118 days.
Therefore, no major solvent-compound
interactions were observed during the
typical Semi-VOST sample storage time.
Aniline and pyridine, which are basic
organic compounds, will react in an acidic
environment to form salts. These salts
are insoluble in methylene chloride but
are soluble in water, which is also present
in the system. These reactions to form
salts and the Soxhlet extraction procedure
are partially responsible for poor and
variable recoveries from XAD-2® resin.
Acid/base extraction must be performed
on the aqueous phase from the XAD-2®
resin to obtain good recovery of salt-
forming compounds.
Breakthrough or specific retention
volume (Vg) is the sample volume required
to elute a compound which is introduced
onto a XAD-2® resin bed. Before break-
-------
through occurs, the amount of spiked
material adsorbed on the resin bed is
directly proportional to the volume sam-
pled. Moreover, the adsorbate-sample
volume relationship must pass through
the origin. After the breakthrough volume
has been reached, the amount of ad-
sorbate is no longer proportional to the
sample volume, but approaches a hori-
zontal asymptotic value equal to the
equilibrium loading.
In the present study, laboratory mea-
surements were made at several tem-
peratures and values of In (V/T) vs. I/T.
These values were fitted by linear regres-
sion to obtain an equation from which
the breakthrough volume estimates could
be made. The resin employed for the
laboratory studies was from the same lot
as the resin used for the subsequent field
study. Carrier flow rate for the laboratory
studies was selected to simulate the face
velocity at the adsorbent cartridge. Lab-
oratory retention volumes, breakthrough
volumes, and breakthrough times for five
compounds are summarized in Table 1.
The data in Table 1 indicate that for the
2-hour precision and bias tests performed
in the field, none of the compounds was
expected to break through the XAD-2®
resin except pyridine. Pyridine is water
soluble and would be expected to pass
through the resin and be collected in the
condensate.
Three quad-probe distributive volume
tests were conducted to compare the
results of laboratory breakthrough volume
studies to data obtained in the field. Dur-
ing these tests, three of the trains were
spiked at 100 MOL with five deuterated
compounds and one train was operated
unspiked. The three spiked trains provided
a triplicate set of samples for breakthrough
analysis and the unspiked train a blank or
background sample. Table 2 presents the
distributive volume data generated from
the field test.
The data of Table 2 were tested to
determine if the breakthrough volume
had been reached using an F statistic
which measured the lack of fit of the data
to a straight line through the origin. If
there were no lack of fit, the breakthrough
volume had not been reached. If there
were a lack of fit to the linear relationship
the breakthrough volume had been ex-
ceeded. The field test distributive volume
data demonstrated that d10-o-xylene and
d2-tetrachloroethane did not break
through the XAD-2® resin bed and that
the retention volume determined experi-
mentally in the laboratory was not ex-
ceeded. The field test distributive volume
results for ds-toluene and d5-chloroben-
Table 1. Retention Volume
Compound
Toluene
1 ,1 ,2,2-Tetra-
chloroethane
Chlorobenzene
Pyridine
o-Xylene
Retention
Volume
mL/gat20°C
137,277
1,430,672
450.473
68.896
1.021,676
Sampling
Flow Rate
cfm'
0.50
0.50
0.50
0.50
0.50
Breakthrough
Volume
nr>(ft3)
2.72(96)
28.6 (101 1}
9.001318)
1.36(48)
20.4 (720)
Time for
Breakthrough
h
3.2
33.7
10.6
1.6
24.0
'0.50 cubic foot/minute (cfm) is equivalent to O.OO0236 rrf/s.
Table 2. Distributive Volume, Adsorption Results
Sample
Volume
m3
1.66
1.64
2.66
5.05
5.00
5.00
9.98
10.04
10.14
Adsorbate
da-toluene
ng
934
954
1291
2297
1698
2071
3041
1698
2037
d5 -chlorobenzene
ng
174
173
323
632
553
615
781
647
607
d;0-o-xylene
ng
937
924
1315
2764
2553
2805
6908
5584
4737
d2-tetrachloroethane
ng
94
92
192
291
290
270
492
530
462
zene confirmed that, for sample volumes
greater than the retention volumes deter-
mined in the laboratory, the total com-
pound mass collected on the resin departs
from a linear correlation with the sample
volume. For the distributive volumes used
in the field test, about 88 percent of the
d5-pyridine was found in the water
impingers following the XAD-2® adsorp-
tion bed. These results show that break-
through had occurred, confirming the
laboratory measurements.
Five tests were performed with the
quad-probe to assess method precision
and bias. Five deuterated compounds
were spiked into the sample trains at the
100 MDL level during the five tests to
provide data on precision and bias. Table
3 summarizes the results of precision
and bias determinations. Bias was cal-
culated relative to percent recovery of the
spiked compound, and precision in terms
of standard deviations and percent coef-
ficients of variation of the percent re-
coveries for each spiked compound. The
bias ranged from -29.0 percent for d5-
pyridine to 0.9 percent for d,0-g-xylene.
Pyridine was the only compound tested
which has significant water solubility.
Because of the high moisture content of
the stack gases, several liters of water
condensate were extracted for prepara-
tion of the samples. Recovery of a water-
soluble compound is far less efficient
when it must be extracted from a large
volume using multiple extraction and
concentration steps. An analysis of vari-
ance of compound recovery showed the
mean percent recoveries were not equal
at a level of significance less than 0.005.
Method bias, therefore, was compound
specific, or at least specific by classes,
e.g., water-soluble.
Precision in terms of percent coefficient
of variation ranged from 8.1 percent for
d,0-q-xylene to 32.9 percent for d2-
tetrachloroethane. A Bartlett Test for
homogeneity of compound variances of
percent recovery showed that there was
essentially zero probability that the com-
pound variances were equal. From the
results of this study it was not possible to
give an overall, or pooled, variance, stan-
dard deviation, or coefficient of variation
for all compounds tested. Rather, the
precision was compound specific. The
percent coefficient of variation for ds-
pyridine was not significantly different
from that of d8-toluene or d5-chloroben-
zene for these three compounds, as the
precision did not seem to be related to
the water-solubility characteristics.
Two quad-probe samples were collected
to provide information on sample stability.
Two trains were spiked with the five
deuterated compounds using the dynamic
spiking system. The other two trains re-
mained unspiked during the flue gas
-------
Table 3. Summary of Precision and Bias Results
Precision
Compound
ds-pyridine
dg-toluene
ds-chlorobenzene
d,0-o_-xylene
d2-tetrachloroethane
Number
of Valid
Samples
16'
13*
13
13
13
Mean%
Recovery
71.0
85.1
86.2
99.1
81.5
Percent
Bias3
-29.0
-14.9
-13.8
-0.9
-18.5
Standard
Deviation
12.6
12.4
11.8
8.0
26.8
Percent
Coefficient
of Variation
17.7
14.6
13.7
8.1
32.9
' Run 03B samples went to dryness in preparation; data were discarded.
2 Runs 01 A. 01C. and 01D were discarded because of the sequence of extractions and
concentrations. For all of the remaining samples, the Soxhlet extract of the XAD-21* was the final
aliquot concentrated.
3 Percent bias = Mean % recovery-100
sampling, but were spiked with a liquid
containing the same five compounds
using a syringe immediately after the
sampling. The liquid spiked samples were
analyzed with other samples collected
during the testing and were used to
assess compound loss due to sample
transport, preparation and analysis. The
two dynamically spiked samples were
stored for five weeks and nine weeks and
then analyzed. Results are shown in Table
4. The data from the two quad-probe
stability samples show that the holding
time (up to 63 days) between sampling
and analysis does not significantly affect
the results obtained. Greater variation
was encountered between different sets
of recoveries obtained from duplicate
liquid spiking experiments than when
two sets of recoveries obtained from
dynamic spiking are compared or when
dynamic spiking was compared to liquid
spiking.
Conclusions and
Recommendations
Several conclusions and recommenda-
tions have been made regarding the use
of the Semi-VOST method for the mea-
surement of POHCs with boiling points
greater than 100°C that are emitted from
hazardous waste incinerators. These
include:
• Method precision and bias were
shown to be compound specific and
could not be generalized for all
classes of organic compounds. Lab-
oratory and field experiments dem-
onstrated that method precision and
bias are related to the distribution
and recovery of a particular POHC
throughout the components of the
sampling train.
• Sample distribution and recovery of
POHCs are related to boiling point,
water solubility, and chemical func-
tionality. Field studies verified these
conclusions using a variety of POHCs
exhibiting a range of boiling points,
water solubilities, and chemical
functionalities.
• XAD-2® adsorbent samples must be
extracted immediately after removal
from the resin cartridge for optimum
analyte recoveries. Samples that are
allowed to stand in Soxhlets at room
temperature prior to extraction may
lose compounds. The compound loss
due to volatilization is related to its
vapor pressure.
• Dynamic spiking of the Semi-VOST
train using deuterated compounds
was demonstrated as a viable ap-
proach to determine method bias
under field conditions. In the absence
of measurable stack emissions of
POHCs, dynamic spiking can be used
to measure method precision.
• Method precision for d5-pyridine was
17.7 percent coefficient of variation.
This value was derived from 16
determinations of percent recovery
from field test data using the dynamic
spiking system. Method bias cal-
culated from these data was -29.0
percent.
• Method precision for d8-toluene was
14.6 percent coefficient of variation
and method bias was -14.9 percent,
derived from 13 determinations of
percent recovery during the second
field test. These values compare
favorably with a method bias of = 1.0
percent determined for toluene in
the first field study.
• Method precision for d,0-o-xylene
was 8.1 percent coefficient of varia-
tion and method bias was -0.9 per-
cent, derived from 14 determinations
of percent recovery for the same
field test.
• Method precision for d2-tetrachloroe-
thane was 32.9 percent coefficient
of variation and method bias was
-18.5 percent, derived from 13 deter-
minations of percent recovery for
the same field test. An estimate of
method bias from the first field study
was -16.0 percent.
• Method precision for ds-chloroben-
zene was 13.7 percent coefficient of
variation and method bias was -13.8
percent, derived from 13 determina-
tions of percent recovery for the
same field test. In the first field
study, method precision for chloro-
benzene was estimated to be 19.9
percent. Method bias for chloroben-
zene was not determined in the first
field study because of the large con-
centration of chlorobenzene in the
stack gas.
• All compounds spiked on XAD-2®
resin with the exception of d2-
tetrachloroethane exhibited better
precision and recovery than pyridine,
which was collected in the con-
densate and impingers. The poorer
precision for d2-tetrachloroethane is
in part due to the variability in the
measurement of this compound near
the method detection limit.
• The field test distributive volume
experiments demonstrated that d,0-
o-xylene and d2-tetrachloroethane
did not break through the XAD-2®
resin bed and that the retention
volume determined experimentally
in the laboratory was not exceeded.
Therefore, the Semi-VOST method
is adequate for the sampling and
analysis of d10-g-xylene and d2-
tetrachloroethane.
• The field test distributive volume
experiments for d8-toluene and d6-
chlorobenzene confirmed that, for
sample volumes greater than the
retention volumes determined in the
laboratory, the total compound mass
collected on the resin departs from a
linear correlation with the sample
volume. The result verifies that
laboratory data can be used to pre-
dict field retention volumes for
POHCs.
• Retention volumes determined
chromatographically in the laboratory
were used to determine sampling
volumes for the field test. Experi-
mental data from the field test sup-
port the use of laboratory-determined
-------
Table 4. Semi- I/OS T Field Test Percent Recovery Results
Run 08
Dynamic Spike
Liquid Spike
Compound
ds-pyridine
da -toluene
d5 -chlorobenzene
dw-o-xylene
d2-tetrach/oroethane
Train A
54.3
948
956
997
1146
Tram B
699
990
983
101 2
1070
Train C
1043
111 6
859
972
86 1
Tram D
1049
1129
1000
989
91 1
Run 08
Compound
Dynamic Sp/ke
Liquid Spike
Train A
Train B
Train C
Train D
ds -pyr/dme
d8- toluene
ds- chlorobenzene
dw-o-xylene
d2 - tetrachloroethane
840
785
91 8
1063
1439
1002
850
943
107 1
1349
944
67 1
633
77 7
554
856
709
704
798
822
retention volumes to predict field
breakthrough volumes.
Pyridme is a basic organic compound
that reacts with aqueous hydrochloric
acid to form an organic chloride salt,
which is highly water soluble Ap-
proximately 88 percent of the
pyridme is collected in the back-half
(impinger solutions) of the Semi-
VOSTtram.
1 A difference in holding time before
extraction for dynamically spiked
XAD-2® samples (35 days vs. 63
days) does not result in a significant
difference in recoveries for all of the
compounds spiked
If the volume of condensate requires
multiple extractions and concentra-
tions, the extract of the aqueous
tram components should be com-
bined with the XAD-2® extract at
the final concentration step
A comparison was made between
observed values and expected levels
of five compounds spiked onto XAD-
2® quality control samples provided
by the Research Triangle Institute
The percent bias ranged from -1.84
percent to -8.90 percent with refer-
ence to the RTI values Average per-
cent bias was -4 68 percent. The
agreement is excellent, showing that
the recovery of organic compounds
from XAD-21"1 resin using the Semi-
VOST methodology is favorable.
The use of surrogate standards dur-
ing sample preparation and analysis
is necessary to obtain the best repre-
sentation of test compound re-
coveries Deuterated analogs, when
available, are the best possible sur-
rogates because they behave exactly
the same as the compounds of in-
terest and will be truly representative
of test compound recoveries.
On the basis of laboratory and field
studies, the following recommendations
are made1
• A literature review should be con-
ducted to categorize Appendix VIII
compounds according to water
solubility, chemical similarity and
adsorption strength (retention vol-
ume) on XAD-2® to allow estimates
of method precision and bias for
compound classes, since method
bias and precision have been shown
to be compound specific. Predicted
values should be verified by
experiment
• Laboratory studies have demon-
strated that low compound recoveries
can be attributed to either poor
chromatographic properties, selec-
tion of inappropriate solvent for
extraction, or compound reaction
with the aqueous components of the
sampling train which become acidic
during sample collection Further
laboratory studies are warranted to
extend these observations and to
use Semi-VOST screening proce-
dures to determine the applicability
of the sample preparation and
analysis steps to other Appendix VIII
compounds or classes of compounds
• Earlier laboratory studies demon-
strated low compound recoveries
associated with NOX flue gas levels.
Variation in compound recovery with
NOX level should be studied using
solutions of pitric acid to simulate
condensate and impinger contents
for evaluating compound reactivity
with NOX The results of the labora-
tory studies could then be used to
plan an additional field evaluation of
the method to include artifact
formation
Laboratory and field studies should
be performed to differentiate be-
tween resin artifacts and Products
of Incomplete Combustion (PICs)
present in XAD-2® resin resulting
from sampling of stack emissions
from hazardous waste combustion
Paniculate filters from selected
Semi VOST trains should be chal-
lenged with known POHC concen-
trations to evaluate POHC-particulate
matter interactions during sample
collection
Further laboratory and field evalua-
tion studies should include an
evaluation of the applicability of the
Semi-VOST method to the measure-
ment of Appendix VIM aldehydes,
aromatic amines such as diphenyla-
mine, and compounds that can only
be analyzed by high performance
liquid chromatography (HPLC) with
a selective detector
-------
r/je EPA author, John Margeson (also the EPA Project Officer, see below)
is with Environmental Monitoring Systems Laboratory, Research Triangle
Park, NC27711; andJ. Bursey, J. Steger, M. Palazzolo. D. Benson, J. Homolya.
R. McAllister. J. McGaughey, and D. Wagoner are with Radian Corporation,
Research Triangle Park, NC 27709.
The complete report, entitled "Laboratory and Field Evaluation of the Semi-
Vost Method, "(Order No. PB 87- 145934/AS; Cost: $30.95, subject to change)
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:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use S300
EPA/600/S4-86/046
0000329 PS
U S ENVIR PROTECTION AGENCY
!!soTol.!iS5Sa«!T5EET
CHICAGO IL 60604
------- |