United States
Environmental Protection
Agency
National Exposure
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/SR-96/051 May 1996
4>EPA Project Summary
VOST Charcoal Specification
Study
A.L. Foster and J.T. Bursey
The volatile organic sampling train
(VOST) methodology incorporates SW-
846 sampling Method 0030 and SW-846
analytical Method 5040 or 5041. VOST
is currently one the leading meth-
odologies available for the sampling
and analysis of volatile principal or-
ganic hazardous constituents and prod-
ucts of incomplete combustion from
stationary sources at very low levels.
However, revisions to the original
method are necessary to maintain VOST
as a viable regulatory tool. Method 0030
states that the VOST sampling tube set
must consist of a front tube containing
Tenax® (a 2,6-diphenylene oxide poly-
mer) and a rear tube containing se-
quential bed of Tenax® and SKC Lot 104
petroleum-based charcoal "or equiva-
lent." However, the method does not
identify a specific equivalent, nor does
the method supply the performance
specifications that would allow deter-
mination of an equivalent. Lot 104 pe-
troleum-based charcoal is no longer
commercially available and has not
been available for several years. Labo-
ratories are presently using a wide
range of substitutes, usually coconut-
based charcoal, and there is a wide
range of performance from batch to
batch of charcoal in one laboratory and
from laboratory to laboratory. To pro-
vide performance specifications and
identify a replacement for SKC Lot 104
charcoal, a VOST charcoal specifica-
tion study was initiated. The following
carbon-based candidate sorbents were
considered: Tenax-GR (a graphitized
Tenax); a Petroleum-based Charcoal;
Ambersorb® XE-340 (hydrophobic car-
bonized resin bead); Anasorb® 747
(beaded active carbon with very regu-
lar pore size); Carbosieve® S-lll (car-
bon molecular sieve); and a Beaded
Activated Charcoal (BAG) with a very
regular pore size.
The results indicated that Tenax-
GR showed significantly poorer per-
formance than the other candidates
in preliminary experimental results.
Ambersorb did not retain the gaseous
volatile organic compounds tested as
well as the others, and recovery of vi-
nyl chloride was very low at all levels
of spiking. Carbosieve was eliminated
as a candidate replacement because of
cost and handling problems. The pe-
troleum-based charcoal was eliminated
because of difficulties in handling a
finely-divided powder. The availability
of Anasorb® 747 proved to be the de-
ciding factor between it and the BAC.
Performance, cost, ease of handling,
and plentiful supply make Anasorb® 747
a good choice for replacement of SKC
Lot 104.
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 (see
Project Report ordering information at
back).
Introduction
The Volatile Organic Sampling Train
(VOST) is used for sampling principal or-
ganic hazardous constituents and prod-
-------
ucts of incomplete combustion from sta-
tionary sources. The sampling protocol for
the VOST is presented in SW-846
Method 0030. The sampling train uses
solid sorbents for collection of volatile or-
ganic compounds. The solid sorbents used
in the VOST are Tenax GC®, a phenylene
oxide polymer, and charcoal. The combi-
nation of sorbents is required because
very volatile organic compounds such as
vinyl chloride are not trapped efficiently by
Tenax® and require the use of a carbon-
based sorbent for efficient collection. VOST
samples are analyzed by thermal desorp-
tion followed by purge and trap gas chro-
matography/mass spectrometry, according
to SW-846 Method 5040 or 5041.
VOST sampling and analytical proto-
cols are periodically revised to address
the development of new technology and
improved procedures. For example,
Method 5041 was written to address the
application of megabore® capillary column
technology to the analysis of VOST tubes.
An aspect of the VOST sampling protocol
that has required revision is the use of the
specified carbon-based sorbent (SKC Lot
104 petroleum-based charcoal). SKC Lot
104 petroleum-based charcoal is no longer
commercially available. Method 0030
specifies that "an equivalent" may be used
but contains no specifications for an
equivalent. To revise the methodology and
provide an appropriate substitute for the
SKC Lot 104 petroleum-based charcoal,
which is no longer available, the U.S. En-
vironmental Protection Agency and Ra-
dian Corporation initiated a VOST Charcoal
Specification Study to define the perfor-
mance characteristics of petroleum-based
charcoal, identify and test potential candi-
date sorbents, and recommend a replace-
ment sorbent for SKC Lot 104 charcoal
for VOST applications.
On the basis of discussions with sor-
bent manufacturers and users, a compre-
hensive literature search, and a preliminary
round of laboratory feasibility testing, five
candidate sorbents were selected for ex-
tensive laboratory testing. These sorbents
were
Reference: Lot 208 petroleum-
based charcoal (SKC);
• Anasorb® 747 (SKC);
• Ambersorb® XE-340 (SKC);
Carbosieve® S-lll (Supelco); and
Beaded activated charcoal (BAG,
from Kureha, Japan).
In the laboratory, the sorbents would be
tested under simulated field conditions.
Results and Discussion
To determine the adsorptive capacity
and recovery efficiency of each of the
candidate sorbents, a thermal desorption
study was performed. Using flash evapo-
ration of the analytes in methanol solution
(the protocol for spiking surrogate com-
pounds onto VOST tubes prior to analysis
according to SW-846 Method 5041), ana-
lytes listed in Method 5041 were spiked
onto VOST tubes containing the candi-
date sorbents. The spiked tubes were ther-
mally desorbed at 250°C, 300°C, and
350°C. Recoveries of spiked analytes were
determined at each desorption tempera-
ture to determine both the optimum des-
orption temperature for each candidate
sorbent and the feasibility of using a single
tube containing only the candidate sor-
bent rather than the tube containing se-
quential beds of Tenax® and carbon-based
sorbent presently used in Method 0030.
For all of the carbon-based sorbents
tested, the laboratory spiking/desorption
study demonstrated that the recovery of
non-gaseous volatile organic analytes us-
ing thermal desorption is poor at any of
the desorption temperatures tested. Since
non-gaseous volatile organic analytes are
efficiently trapped by the carbon-based
sorbents but are not quantitatively released
by thermal desorption, Tenax®-GC must
be used to trap non-gaseous volatile or-
ganic analytes. The tube containing se-
quential beds of Tenax® and charcoal is
used in Method 0030 because organic
analytes such as methylene chloride can
show some distribution on the second tube
of the VOST sampling train. If this second
tube is entirely carbon-based sorbent, the
organic compounds will not be recovered.
If the back tube of the VOST train con-
tains only carbon-based sorbent, two
Tenax® tubes in front of the carbon-based
sorbent will be required to ensure that
non-gaseous volatile organic compounds
can be recovered by thermal desorption.
A sorbent cleaning study was performed
to develop a procedure for initial cleaning/
conditioning for the carbon-based sorbent
and to evaluate the feasibility of regenera-
tion of the sorbent after use. A high tem-
perature oven (ambient to 360°C) was
modified for sorbent cleaning with 20 indi-
vidually regulated stainless steel gas lines,
each capable of 0 to 100 mL/min of inert
gas flow. Tenax® tubes were thermally
desorbed at 250°C for 8-12 hours. This
temperature was selected to be well above
the Tenax® desorption temperature of
180°C in the VOST analytical protocol yet
well below the manufacturer's upper tem-
perature limit of 350°C for Tenax®. Car-
bon-based sorbents, with a higher upper
temperature limit, were cleaned at 300°C
for 18 to 24 hours. All sorbents were des-
orbed under a flow of 80 to 100 mL/min of
high purity nitrogen.
After the cleaning process, a percent-
age of the tubes were checked for cleanli-
ness by gas chromatography/mass
spectrometry (GC/MS) analysis. The clean-
ing conditions described above were suffi-
cient to condition new tubes or to
regenerate used VOST tubes after rea-
sonable use.
A preliminary dynamic spiking study was
performed to determine correct parameters
for a major dynamic spiking study. It was
also necessary to evaluate the reproduc-
ibility and accuracy of spiking VOST tubes
through a simulator at various analyte con-
centration levels. Using a modified three-
tube VOST configuration (Tenax®/Tenax®/
carbon-based sorbent), triplicate spiking
experiments were conducted at three spik-
ing levels (approximately 100 ng, 200 ng,
and 300 ng). The VOST tube sets were
analyzed separately (front/middle/back) by
GC/MS to determine both total recovery
and distribution of spiked analytes among
the three cartridges.
Conclusions
The data were subjected to statistical
analysis to differentiate the performance
of the candidate sorbents. No statistically
significant differences were found for the
whole range of analytes at any spiking
level. Ambersorb® XE-340 demonstrated
very poor recoveries for vinyl chloride com-
pared to the reference sorbent and the
other candidate sorbents. The other three
candidate sorbents all behaved similarly
and were comparable in performance to
SKC Lot 208. Sorption/recovery data that
can be used as a basis for selection of an
alternative to SKC Carbon Lot 104 are
shown in Table 1.
In Table 1, chloromethane shows the
high and erratic recoveries that have char-
acterized this compound in laboratory and
field studies that include a carbon-based
sorbent. The compound is apparently be-
ing formed on the sorbent tubes. The other
gaseous organic analytes, vinyl chloride,
bromomethane, and chloroethane, show
reasonably reproducible recoveries at all
concentration levels.
The standard SW-846 Method 5041
VOST calibration procedure, where ana-
lytes, surrogate compounds, and internal
standards are spiked onto the tubes by
flash evaporation immediately prior to
analysis, should not be used in the analy-
sis of a sorbent tube containing only car-
bon-based sorbent because analyte
recoveries are not quantitative with increas-
ing boiling point of the analyte. A calibra-
tion study was therefore conducted to
determine an appropriate method of cali-
bration for VOST analyses using the modi-
fied three-tube configuration. Five-point
-------
Table 1. Mean Percent Recoveries from Sampling Runs for Reference Sorbent (SKC Lot 208) and
Candidate Carbon-Based Sorbents (100/200/300 ng Spiking Level)
Organic Analyte
chloromethane
vinyl chloride
bromomethane
chloroethane
trichlorofluoromethane
1,1-dichloroethene
methylene chloride
iodomethane
1,1-dichloroethane
chloroform
1,1, 1 -trichloroethane
carbon tetrachloride
benzene
1 ,2-dichloroethane
trichloroethene
1 ,2-dichloropropane
cis- 1 , 3-dichloropropene
toluene
trans-1 , 3-dichloropropene
1,1,2-trichloroethane
tetrachloroethene
n-octane
chlorobenzene
SKC Lot 208
5784/4055/2014
69/65/85
78/61/46
47/52/60
97/99/120
79/80/96
139/97/100
56/70/74
83/80/93
89/78/86
80/75/87
70/71/82
108/87/100
78/68/76
96/87/97
86/79/88
51/56/61
99/83/79
63/71/76
85/23/89
80/83/94
133/115/93
79/79/86
Anasorb® 747
646/410/470
71/87/88
45/38/35
37/49/69
93/98/113
77/84/99
125/101/105
71/69/81
76/85/94
79/80/88
78/78/91
68/74/86
90/96/99
71/70/77
81/90/94
73/79/86
47/55/60
78/95/88
59/65/74
74/80/85
68/84/91
95/104/95
64/78/83
Ambersorb®
XE-340
117/142/181
3/3/13
30/42/49
56/59/90
126/126/142
86/72/111
154/106/111
70/64/78
87/83/95
92/81/89
90/77/91
78/70/86
179/147/164
80/69/75
90/85/92
83/81/86
52/57/62
112/114/105
64/70/80
82/85/89
79/88/91
105/108/106
78/77/84
Carbosieve®
S-lll
738/371/364
74/82/91
38/54/56
38/49/74
97/100/117
85/77/106
129/96/106
68/57/70
84/80/97
92/77/89
81/72/91
74/63/85
125/84/99
86/68/77
94/86/96
86/80/87
59/57/60
100/91/88
70/70/74
89/84/86
88/88/94
108/104/102
84/81/86
BAG
Kureha
1197/397/161
74/74/75
20/22/35
37/38/58
100/88/104
96/80/101
152/94/101
68/61/80
89/89/96
95/85/89
88/78/92
80/76/88
109/94/103
84/77/80
97/97/99
83/87/88
49/61/60
93/83/92
57/76/79
78/90/89
87/94/95
108/93/100
82/88/87
Note: All sorbent tubes were desorbed at 250°C. Mean recoveries are calculated from duplicate sampling runs at each spiking level.
calibration curves were created under four
sets of conditions:
Analytes, surrogate compounds,
and internal standards spiked in
water with the sorbent tube des-
orbed thermally according to the
Method 5041 protocol;
Analytes, surrogate compounds,
and internal standards spiked on
one Anasorb® tube, which is then
desorbed thermally according to the
Method 5041 protocol;
Analytes, surrogate compounds,
and internal standards spiked on
one Tenax® tube, which is then de-
sorbed thermally according to the
Method 5041 protocol; and
Analytes, surrogate compounds,
and internal standards spiked on a
paired set of Tenax® front tube and
Anasorb® back tube, with the pair
then desorbed thermally according
to the Method 5041 protocol.
Calibration curves obtained under each
set of conditions were evaluated for ac-
ceptable compound recoveries and per-
cent standard deviation for response
factors for the spiked analytes. The evalu-
ation showed that the calibration curve
determined from spiking analytes and stan-
dards into water was superior in recovery
and reproducibility and that the curve gen-
erated with paired Tenax® and Anasorb®
was acceptable within the specifications
of Method 5041. Calibration curves origi-
nating from Tenax® only and from
Anasorb® only did not meet the accep-
tance criteria of Method 5041.
On the basis of the laboratory test re-
sults and the following considerations, a
replacement carbon-based sorbent could
be selected:
Anasorb® 747 showed consistent
performance, low cost, and avail-
ability of an abundant supply.
SKC Lot 208 petroleum-based char-
coal, the reference sorbent, showed
performance equivalent to all of the
other candidate sorbents. However,
batch-to-batch variability of petro-
leum-based charcoal has historically
been high. If alternative sorbents
with equivalent performance are
-------
available, another choice of sorbent
is preferable to petroleum-based
charcoal.
Because of its poor recovery for
vinyl chloride, Ambersorb® XE-340
was eliminated.
Because of its high cost (twenty
times as expensive as the other
candidate sorbents), Carbosieve®
S-lll was eliminated; other less
costly sorbents demonstrated
equivalent performance.
BAG charcoal was eliminated be-
cause of concern over long-term
availability from a foreign supplier.
Anasorb® 747 was selected as the re-
placement sorbent for petroleum-based
charcoal because of its consistent perfor-
mance, low cost, and abundant supply for
the foreseeable future. A field study with
dynamic spiking is required to compare
the performance of the modified VOST
train (three sorbent tubes) with the stan-
dard VOST train. The GC/MS calibration
should be performed with analytes and
internal standards spiked into water, with
surrogate compounds spiked on the
tube(s) being analyzed. VOST tubes from
the modified train may be analyzed indi-
vidually, or the Tenax® tubes may be
paired and the Anasorb® tube analyzed
individually.
A.L. Foster and J. T. Bursey, are with Radian Corporation, Research Triangle Park,
NC 27709.
Robert G. Fuerst is the EPA Project Officer (see below).
The complete report, entitled "VOST Charcoal Specification Study," (Order No.
PB96-175252; Cost: $47.00, 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
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection Agency
National Risk Management
Research Laboratory (G-72)
Cincinnati, OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use
$300
EPA/600/SR-96/051
------- |