EPA/600/A-93/171
REPLACEMENT OF CHARCOAL SORBENT IN THE VOST
            Larry D. Johnson and Robert G. Fuerst
               Source Methods Research Branch
         Methods Research and Development Division
   Atmospheric Research and Exposure Assessment Laboratory
            U.S. Environmental Protection Agency
              Research Triangle Park, NC 27711

                A,  L. Foster and J. T. Bursey
                    Radian Corporation
              Research Triangle Park, NC 27709
                       Presented At
                       12th Annual
             International Incineration Conference
                      Knoxville, TN
                       May 5, 1993

-------
                                    ABSTRACT
            REPLACEMENT OF CHARCOAL SORBENT IN THE VOST
       EPA Method 0030,  (VOST), for sampling volatile organics  from stationary sources,
specifies the use of petroleum-base charcoal in the second sorbent tube. Charcoal has proven to
be a marginal performer as a sampling sorbent, partly due to inconsistency in analyte recovery.
In addition, commercial availability of petroleum charcoal for VOST tubes has been variable.
Lack of data on comparability and variability of charcoals for  VOST application has created
uncertainty when other charcoals are substituted.
       Five potential sorbent replacements for charcoal in Method 0030 were evaluated along
with a  reference charcoal. Two of the sorbents tested, Ambersorb XE-340 and Tenax GR, did
not perform well enough to qualify as replacements. Three candidates, Anasorb 747, Carbosieve
S-III and Kureha Beaded Activated Charcoal, performed adequately, and produced statistically
equivalent results. Because Anasorb 747 exhibited  an  excellent combination of performance,
availability, and cost, it was selected for use in field tests to follow this study.

-------
INTRODUCTION

The Volatile Organic Sampling Train (VOST), EPA Method 0030, has been the method of
choice for sampling most volatile organic compounds since its introduction (1,2).  Fig, 1 is a
schematic diagram of the VOST.   Two of the key elements of the train are  the two sorbent
cartridges.  These cartridges are the primary devices for collection and concentration of the
volatile compounds in the flue gas being sampled.  The front sorbent tube contains 1.6 g of
Tenax-GC, a polymer of 2,6-diphenylene oxide, one of many porous polymer bead materials
often used for sampling or for gas  chromatography column packing. The VOST method was
designed with the intent of capturing all of the compounds of interest on the front tube, with the
back tube serving as a secondary trap in case of breakthrough of the more volatile compounds.
In order to minimize potential breakthrough of the second tube, it contains 1.0  g of Tenax-GC,
followed by 1.0 g of petroleum based charcoal.

The combination of Tenax-GC followed by charcoal is effective because the Tenax-GC traps all
but  the most volatile compounds and  prevents them from  being  irreversibly sorbed  by the
charcoal.  The charcoal, in turn, acts as a "safety net" and traps the few compounds which are
not quantitatively sorbed by the Tenax-GC. It would be preferable to load both tubes with a
single sorbent  capable of quantitatively trapping all compounds  of interest  and exhibiting
quantitative release of the compounds upon heat desorption.  Unfortunately, no such ideal sorbent
has yet been found and characterized,

Tenax-GC has served well as the primary  sorbent in the VOST, and fortunately, most of the
compounds of interest from incinerators are trapped on the front tube containing this material (3).
The charcoal has performed adequately, but has been far from ideal in several  respects.  It has
sometimes been contaminated with organic material which has been difficult to remove and
which has complicated analysis and limited detection limits. In addition, charcoal has exhibited
inconsistency of composition and uncertainty of supply.

Method 0030 specifies petroleum  charcoal from SKC Lot 104 or equivalent, but gives  no
procedure for determining equivalency. Petroleum based charcoal was chosen because it is
slightly hydrophobic and, therefore, does not retain water as readily as coconut charcoal and its
sorption properties are not sensitive to the presence of water.  Petroleum charcoal also shows less
tendency than coconut charcoal to irreversible sorption of compounds sampled.  Lot 104 charcoal
became unavailable several years ago, and  samplers have sometimes had difficulties obtaining
petroleum charcoal in  any form.  Coconut charcoal  has  sometimes been substituted, with
uncertain effect upon both sampling and analysis.

The goal of this project was to find and test a  sorbent with sorption and recovery behavior at
least as good as petroleum charcoal, with better consistency of properties and availability, and
hopefully with lower contaminant levels and water sensitivity.

EXPERIMENTAL

The project was carried out in three experimental phases.  Phase One was a  screening study

-------
 designed to determine suitability of a series of sorbents for more extensive testing. Phase Two
 involved upgrading the stack sampling simulation, and showed the need for improvements in the
 sorbent spiking procedure and  the analysis method.  Phase Three included determination of
 optimum  desorption temperature, development of an improved sorbent cleaning procedure,
 determination of the best calibration method, and dynamic spiking and recovery studies of a
 modified VOST.  The dynamic  spiking experiments of Phase Three will be described in detail,
 but only the conclusions of the  other portions of that phase will be presented.

 Phase One.  Five sorbents were tested in the initial screening study. Petroleum based charcoal
 (SKC Lot 208) was chosen  as  the reference material to represent the original SKC Lot 104,
 which is no longer available. The other four sorbents tested were Ambersorb XE-340 (available
 from SKC), Anasorb 747 (SKC  Lot 645), Carbosieve S-UI (available from Supelco) and Tenax-
 GR  (available from Alltech).   Ambersorb XE-340 is a hydrophobic carbonized resin bead.
 Anasorb 747  is a beaded active carbon with a very regular pore size and structure which should
 contribute to consistent performance characteristics. Carbosieve S-III is a carbon molecular sieve
 which is widely used in conjunction with Carbopak sorbent in commercial analytical desorption
 traps.  Tenax-GR, not to be confused with Tenax-GC, is a new material only recently made
 commercially available.  Tenax-GR consists of Tenax with graphite incorporated  into the
 particles.  The intent of the manufacturer was apparently to produce a sorbent with the virtues
 of both constituents.

 Twelve test compounds shown  in Table I were selected to serve as a representative  set of
 volatile organics spanning the practical limits of Method 0030.  Three compounds were selected
 to serve as surrogates for all three phases, and are the first compounds listed in Table III. The
 Phase One analyte recovery and detection limit study utilized VOST tube pairs prepared as
 specified in Method 0030.  The front tube was packed with 1.6 g of Tenax-GC and 1.0 g each
 of Tenax-GC and the candidate sorbent were loaded into the rear tube. Analytes and surrogates
 were loaded onto clean VOST tube pairs by a flash vaporization technique.  The paired tubes
 were connected to  a heated injection port so that the  flow of high purity nitrogen carrier gas
 entered the front Tenax-GC tube and exited from the rear of the sequential sorbent tube. The
 target analytes (25 ng) and surrogates (250 ng) in methanol solution were then injected into the
 injection port with  a microlker syringe.  The carrier gas swept the volatilized compounds onto
 the sorbent tubes.  The tube pair was placed in an ambient temperature purging apparatus, and
 was purged for 20 min with a total of 20 L of high purity nitrogen in order to simulate sampling
 conditions. Sample tubes were  sealed and  stored to await analysis.

 Analysis of these compounds on the sample tubes prepared by the procedure described above,
was performed using a slightly  modified version of Method 5041  (4).  Method 5041  includes
 heat desorption, purge and trap, and gas chromatography/ mass spectrometry.  Readers interested
 in further details on the modifications made to Method 5041 and the reasons for them, are
 referred to the EPA report on this project (5). A detection limit study for the test analytes was
performed for each sorbent.  Each  study  was  carried out according to instructions given in
 Reference 6.  Seven replicate sets  of tubes  were analyzed for each candidate sorbent.

 Phase Two.  Phase Two of the project was designed to provide additional,  more realistic
testing for the candidate sorbents which survived Phase I. The VOST sample pairs were

-------
prepared containing Tenax-GC in the front tube and only candidate sorbent in the rear tube.
The paired tubes were cleaned, blank checked, and then spiked with relatively low levels of
target compounds in a dynamic atmosphere generator to simulate the hot, wet conditions of a
stack. The spiked VOST pairs were analyzed front and back separately to determine both
recovery and distribution of target compounds.

A dynamic spiking apparatus was designed to simulate stationary  source stack conditions as
closely as possible.  Ambient high-purity nitrogen carrier gas was admitted into the front of the
system and then passed throuj;h a heated glass coil. Immediately after this gas pre-heating point,
purified analyte-free water was admitted into the system at a constant rate metered by an HPLC
pump. The gas-water mixture was then passed through a second heated glass coil to complete
vaporization. After this second heating coil, target compounds were spiked into the system at
a constant rate as metered by mass flow controllers. Bromochloromethane and chloroethane were
added to the list of test compounds used in Phase  One, bringing the total to fourteen for Phase
Two. The target compounds were in a gaseous state and administered  from a certified Scott
cylinder.  The moist, pollutant-laden  stream then  passed out of the  dynamic spiking apparatus
and into  a standard VOST train.   Tubes were spiked with 100 ng of each compound.  As in
Phase One, samples were analyzed by a slightly modified Method 5041.

Phase Three. A fifth sorbent was added to the four successful candidates from Phases One and
Two. Kureha Beaded Activated Carbon (BAC) was obtained from  Kureha Chemical Industry
Co., Japan.  BAC has been used for VOST testing with apparent  success, and seems to be
similar to Anasorb 747 in structure and performance. Phase Three included determination of the
optimum desorption temperature,  development of an improved sorbent cleaning procedure,
determination of the best calibration method for the analytical system, and dynamic spiking and
recovery studies of a modified VOST. Space only permits description of the dynamic spiking
experiments,  but conclusions of the other portions will be briefly presented in a later section.
For complete experimental details  of Phase Three, see Reference 5.

The triplicate dynamic spiking study was designed to provide a  comprehensive picture of the
performance characteristics of each of the candidate sorbents in realistic VOST applications. A
new dynamic spiking apparatus was  built and tested, and it  was shown that VOST cartridges
could be  reproducibly spiked with  various levels of analytes.  A  modified VOST configuration
was adopted in order to give greater protection against breakthrough of the non-gaseous analytes
which had displayed poor recovery  from the candidate sorbents in the early phases of the project.
This configuration consisted of front and middle tubes containing Tenax-GC, and a full tube of
candidate sorbent  in the back.

The dynamic spiking apparatus consisted of four major sections,  a carrier gas moisturizing
section, a spiking and primary mixing chamber,  a secondary mixing chamber and a sampling
manifold.  The system operated at approximately 175°C and 11%  moisture, in order to be
consistent with a hot moist flue gas.

The dynamic spiking system was sampled  using the three  tube modified  VOST, following
procedures in Method 0030. Triplicate runs were conducted for each candidate sorbent at three
spiking levels, approximately 100 ng.,  200 ng, and 300 ng. These spike levels represent normal

-------
levels encountered during actual field sampling with the VOST.  Lower spike levels were used
in Phase One because a detection limit study was an important part of the initial screening.
Detection limit studies, because of their very nature, must be carried out at the lower end of the
operating range.  The expanded list of twenty three test compounds and three surrogates is shown
in Table III.  The VOST tubes collected IL/min of sample for 20 minutes, giving a 20L sample.
The tubes were then sealed, and either analyzed immediately or stored appropriately.  Each three
tube modified VOST  functioned as a unit, but the tubes were analyzed separately  in order to
provide analyte distribution information.  Analysis was by the modification of Method 5041 used
in Phase One, except  that carbonaceous sorbents and Tenax-GC were heat desorbed at 250C.

RESULTS AND DISCUSSION

Phase One.  As previously mentioned, seven replicate sets of tubes were analyzed in Phase One.
Average values are used in this paper, but all data is included in Reference 4.  Table I presents
the detection limits determin
-------
CONCLUSIONS

Phase One.  Tenax-GR performed poorly with respect to recovery of low boilers, and was
eliminated from the later phases  of  the project.   Ambersorb XE-340 exhibited  marginal
performance for low boilers, but was carried forward, partly because of excellent performance
with respect to the compounds in the  middle range of boiling point applicability.  The other
candidates performed in a siniilar manner to each other and were carried forward to Phase Two,

Phase Two.   Inconsistencies in Phase Two  data  showed the need for development and
improvement in the dynamic spiking equipment and procedure, in the sorbent cleaning procedure,
and in the analysis method.

Phase Three.  This phase produced a number of important conclusions. Even though previous
discussion only supports the conclusions related to Ihe dynamic spiking  experiments, several
others are included for added perspective.

The carbon-based sorbents tested in this project showed strong adsorptive properties toward non-
gaseous VOST analytes.  These analytes could be recovered at only 50% or less,  even  at
desorption temperatures of 350°C.  This means that it is still essential to include Tenax-GC
ahead of the carbonaceous sorbents to prevent analyte loss.  A thermal desorption temperature
of 250°C was sufficient to desorb the light compounds from the candidate sorbent tubes. Higher
temperatures gave no significantly increased recoveries.

Tenax-GC should be cleaned at 250°C for eight to  twelve hours and carbon-based sorbents
should be cleaned at 300°C tor twelve to  eighteen hours. All tubes should be purged with inert
gas while being thermally cleaned,

A calibration curve with analytes and internal standards purged from water was  superior to the
other methods tested for initial calibrations and daily  QC analyses.

VOST sample  sets in the modified configuration could be accurately and reprodueibly spiked
with target analytes through  the use of a dynamic spiking apparatus which simulated a stack
environment.

A dynamic spiking study demonstrated that Ambersorb XE-340 performed very poorly for vinyl
chloride and is not an acceptable replacement for charcoal in the VOST.  A statistical evaluation
of the sorbent  recoveries confirmed  that at a given dynamic spiking level, performance
differences for the other three: candidates and the reference charcoal measured in terms of percent
recoveries of analytes  were not  significant.  Based on performance alone, any of these three
would be an acceptable replacement for charcoal in the VOST.

Because Anasorb 747 exhibited an excellent combination of performance, availability, and cost,
it was selected for use in field tests of the three-tube VOST.

-------
FURTHER WORK

Two other significant related projects have been completed and will be published in the near
future.  A draft three-tube VOST method using Anasorb 747 has been written, and is under
review as an alternative to Method 0030.  A field test in which  the new VOST method was
evaluated along with Method 0030 has been carried out

NOTICE

The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency under Contract 68-D1-0010 to  Radian Corporation.  It has
been subjected to the Agency's peer and administrative review, and it has been approved for
publication.  Mention of trade names or commercial products does not constitute endorsement
or recommendation for use.

REFERENCES

1.  G.A. Jungclaus, P.O. Gorman, G. Vaughn, G.W. Scheil, FJ. Bergman, L.D. Johnson
   and D. Friedman,  "Development of A Volatile Organic Sampling Train," presented at
   Ninth Annual Research Symposium on Land Disposal, Incineration, and Treatment of
   Hazardous  Waste, Ft. Mitchell, KY, May 1983.  In Proceedings, EPA-600/9-84-015,
   PB84-234525, July 1985,
2.  "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, DC, November 1986.
3. M.D. Jackson, J.E. Knoll, M.R. Midgett, J.T. Bursey, R.A. McAllister and R.G. Merrill,
   "Evaluation of VOST and Semi VOST Methods for Halogenated Compounds in the Clean Air
   Act Amendments Title III, Bench and Laboratory Studies,"  In Proceedings of the National
   Air and Waste Management Association Meeting, Kansas City, MO, June  1992.
4.  "SW-846 Manual, Second Update to the 3rd ed."  U.S. Government Printing Office,
   Washington, DC, November 1990.
5.  "VOST Charcoal Specification Study," Radian Corp. Project Report to USEPA, Contract
   68-D1-0010, RTF, NC, 1993.  NTIS Report undergoing clearance.
6.  "Definition and Procedure for the Determination of the Method Detection Limit," Code of
   Federal Regulations, 40CFR Part 136, Appendix B, U.S. Government Printing Office,
   Washington, DC, 1991,

-------
               Table I

 An alyte Detection Limits in Nanograms
Phase One Method Detection Limit Study
Target Compounds
Chloromethane
Chlorobenzene
Vinyl Chloride
1,1-Dichloroethene
Chloroform
Toluene
Methylene Chloride
1 ,1 ,1 -Trichloroethane
Carbon Tetrachloride
Benzene
Trichloroethene
Tetrachloroethene
Petroleum
Charcoal
82
8
5
16
10
122
35
6
7
16
7
7
Ambersorb
XE-340
64
16
26
17
5
117
306
14
10
43
6
10
Anasorb
747
482
9
10
6
6
132
77
13
15
60
5
5
Carbosieve
S-lll
490
10
9
14
9
79
386
39
33
9
8
11
Tenax-GR
12
32
3
52
43
102
46
44
43
75
44
40

-------
                  Table II

Average Percent Recoveries For Target Analytes
   Phase One Method Detection Limit Study
Target Compounds
Chloromethane
Chlorobenzene
Vinyl Chloride
1,1-Dichloroethene
Chloroform
Toluene
Methylene Chloride
1,1,1 -Trichloroethane
Carbon Tetrachloride
Benzene
Trichloroethene
Tetrachloroethene
Petroleum
Charcoal
247
80
87
102
102
484
176
101
102
117
98
96
Ambersorb
XE-340
164
76
47
90
100
340
389
114
109
247
100
91
Anasorb
747
1017
76
94
101
88
299
215
98
88
138
89
88
Carbosieve
S-lll
710
79
77
87
92
313
491
120
105
95
87
85
Tenax-GR
15
87
2
99
102
327
222
104
100
197
96
94

-------
                      Table HI

Average Percent Recoveries from Triplicate Sampling Rons
                200 ng Spiking Level
                    Phase Three

1.
2.
3.
4.
5.
6.
7.
8.
9,
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Target Compound
d4-1 ,2-DiehIoroethane (Surr.)
dS-Toluene (Surr.)
p-Bromofluorobenzene
(Surr.)
Crtloromethane
Vinyl Chloride
Bromomethane
Chloroethane
Triehlorofluoromethane
1 ,1 -Dichloroethene
Methylene Chloride
lodomethane
1 ,1 -Dichloroethane
Chloroform
1,1,1 -Trichloroethane
Carbon Tetrachloride
Benzene
1 ,2-Dichloroethane
Trichloroethene
1 ,2-Dichloropropane
cis-1 ,3-DiehIoropropene
Toluene
trans-1 ,3-Dichloropropene
1,1,2-Trichloroethane
Tetrachloroethene
n-Octane
Chlorobenzene
Petroleum
Charcoal
NA
NA
NA
4055
65
61
52
99
80
97
70
80
78
75
71
87
68
87
79
56
83
71
23
83
115
79
Anasorb
747
NA
NA
NA
410
87
38
49
98
84
101
69
85
80
78
74
96
70
90
79
55
95
65
80
84
104
78
Ambersorb
XE-340
NA
NA
NA
142
3
42
59
126
72
106
64
83
81
77
70
147
69
85
81
57
114
70
85
88
108
77
Carbosieve
S-lll
NA
NA
NA
371
82
54
49
100
77
96
57
80
77
72
63
84
68
86
80
57
91
70
84
88
104
81
Kureha
BAG
NA
NA
NA
397
74
22
38
88
80
94
61
89
85
78
76
94
77
97
87
61
83
76
90
94
93
88

-------
Glass
Wool
Filter
 Stack
                                  Three Way
                                  Glass/Teflon
                                    Valve
Spiking Gas In
To Bubble
Flow Meter
               JL
          Heated Glass •
          Lined Probe  :
r                            Three Way
                            Glass/Teflon
                              Valv/o
                           Charcoal
                            Trap
                        Y
                                         1/4" Teflon
                                           Tube
                                 *"*  "    *        —. .
                                 wDi>C>€>«*'S«E"«.S       c?iiiOe
                                    Trap          Gel
                                Sampling Module
                                                          Meter Box
                    Figure 1. Volatile Organic Sampling Train

-------