EVALUATION OF COMMERCIALLY-AVAILABLE
PORTABLE GAS CHROMATOGRAPHS
R. E. Berkley, Environmental Protection Agency, Atmospheric
Research and Exposure Assessment Laboratory, Research Triangle
Park, NC 27711,
M. Miller and J. C. Chang, IIT Research Institute, Chicago, IL
60616
K. Oliver and C. Fortune, ManTech Environmental Services, Research
Triangle Park, NC 27709.
Six commercially-available portable gas chromatographs (PGC)
were evaluated at a Superfund site during startup of bioremedia-
tion. Concentrations of volatile organic compounds (VOC) were
slightly above ambient background levels. Concurrent colocated
grab samples were collected periodically in Summa-polished canis-
ters. They were analyzed by Method TO-14 using a mass-sensitive
detector. The grab samples served as standards to assess the ac-
curacy of data reported by the PGCs.
Introduction
Portable gas chromatographs (PGC) offer the advantage of pro-
viding immediate data. They can often produce more information at
less cost than laboratory-based methods of analysis. A variety of
PGCs are currently available. During January 1992, we evaluated
five PGCs at the French Limited Superfund Site in Crosby, TX. They
were selected on the basis that they were field-deployable, and the
manufacturers were each willing to provide technical support and a
unit for evaluation.
The French Limited Superfund site is an abandoned sand pit
into which refinery waste has been dumped. Before remediation, ten
feet of sludge underlay twenty-five feet of water, covering an area
-------
of seven.acres. The water was clear, but volatile solvents were
leaching from the sludge into ground water. The French Limited
Task Group (FLTG), formed by the potentially responsible parties,
proposed bioremediation. Their plan was approved by EPA after
successful pilot-testing. They installed a containment barrier
around the site projecting 65 feet downward into a clay layer
below and extending 15 feet above ground to keep out flood water.
A similar barrier divides the pond in half? the two sides are being
treated consecutively. Dredges loosen sludge from the bottom of
the pond and high-speed stirrers mix it with the water. Streams of
pond slurry are being pumped out of the pond, injected with oxygen
gas and nutrients, then pumped back into the pond below the sur-
face. This selectively enhances growth of those strains of indi-
genous bacteria which feed on the sludge.
Experimental
The PGCs (with their detectors) included Photovac 10SPLUS -
10.6 eV photoionization (PID), Microsensor Systems 301 - surface
acoustic wave, Sentex Scentograph - 11.7 ev" Argon ionization, HNU
Model 311 - 10.2 ev PID, and SRI 8610 - 10.2 eV PID and electro-
lytic conductivity (ELCD). A previously-evaluated Photovac 10S70
which is owned by EPA was also included (1/2). All units were op-
erated inside a power-control shed located 20 feet away from, and
15 feet above, the edge of the pond. The interior of the shed was
maintained at about 70°F. All units were connected to 110 volt 60
Hz commercial power. Each unit used its own sample pump to import
outside air through 1/8 inch OD stainless steel tubing. Calibra-
tions were performed periodically using mixtures prepared by dyn-
amic dilution of commercial standards (Alfagaz, Scott) and stored
in 6 liter Summa-polished canisters. Grab samples were taken per-
iodically by opening the valve of an evacuated canister while hold-
ing it as close as possible (within three feet) of the assemblage
of intake tubes while they were collecting samples. Grab canisters
were returned to the laboratory and analyzed by GC/MSD according to
Method TO-14. Canister grab sample data were taken to be true con-
centrations of the compounds analyzed by the PGCs.
Results and Discussion
Detection limits for the PGCs were calculated using data ac-
quired during field calibrations. They are shown in TABLE 1. For
the MSI 301, which doesn't have an identifiable baseline, and the
Scentograph, which doesn't output a baseline signal, it was dif-
ficult to estimate a meaningful detection limit. Baseline distur-
bances can render calculated detection limits meaningless, and such
problems are common in field operations. Any of the instruments
operating uncontaminated in a more sheltered environment might have
shown apparently-lower detection limits.
TABLE 2 contrasts data from three grab samples with corres-
ponding PGC data. Analyte levels were near the detection limits
shown in TABLE 1. The PGCs generally produced results similar to
-------
the grab sample. There were a few flyers, for example the HNU 311
at 10:30 and the Photovac 10S70 at 11:28. These could have been
caused by poor mixing of air, contamination of equipment, or simi-
lar accidents. Agreement between the methods, though not exact,
was close enough to show that all of the PGCs provided reasonable
estimates of the concentrations of compounds which they could
detect and for which they were calibrated.
In TABLE 3 the degree of agreement between PGC and canister
data is analyzed in terms of the absolute values of the differences
between them. Averages of absolute differences for each unit for
each compound are shown with their standard deviations (in paren-
theses) . A low average difference indicates good agreement between
canister and PGC data. The standard deviation, considered together
with the range, which is defined by the maximum and the minimum
values which are shown, indicates how consistent the agreement was
between PGC and canister data. A small average difference with a
smaller standard deviation and a narrow range would indicate close
agreement between the two methods. A large average difference with
a small standard deviation and a narrow range could be due to sys-
tematic error, perhaps an inaccurate calibration standard. A small
average difference with a standard deviation of comparable magni-
tude and a narrow range would indicate that the PGC was producing
data of reasonable accuracy but mediocre precision. That would be
expected when analyzing concentrations which are near detection
limits. Most data in TABLE 3 are of that type. A larger average
difference with a still larger standard deviation and a very broad
range would suggest a data set which contains a flyer. Examples in
TABLE 3 are Photovac 10S70 (benzene), MSI 301 #10 (toluene), and
HNU 311 (benzene and toluene). A large average difference with a
broad range and a standard deviation comparable in magnitude to the
range would indicate little or no agreement between methods. That
pattern is not seen in TABLE 3. Zero minimum values result from at
least one case of exact agreement between the two methods, but the
multitude of zero minima in TABLE 3 actually resulted from runs in
which neither method detected anything.
V
All units performed as expected^reasonably well. Examination
of TABLES 2 and 3 shows agreement to better than an order of magni-
tude among all methods, except for the four bad points. This is
encouraging, since these instruments were built according to dif-
ferent design criteria and intended for different applications. In
view of this general agreement, it would be futile afctatnyt to rank
the instruments arbitrarily on the basis of the results obtained,
which in this case reflects only their performance in one environ-
ment. Concentrations of VOCs encountered during this study were
much lower than expected, and the range of concentrations was quite
narrow. A study carried out in a different environment might have
produced a similar body of data differing only in detail and pos-
sibly yielding no additional knowledge about relative capabilities.
-------
Conclusions
The instruments evaluated in this study all performed satis-
factorily according to claims for their capabilities. All of them
were able to detect the levels of compounds encountered at the
French Limited Superfund Site, usually with a reasonable degree of
accuracy. Choosing one of them for a particular application should
be based upon consideration of its particular features and capabil-
ities.
References
1. R. E. Berkley, K. Kronmiller, and K. Oliver, Proceedings of the
1990 EPA/AWMA International Symposium: Measurement of Toxic and
Related Air Pollutants, 849, 1990.
2. R. E. Berkley, J. L. Varns, and J. Pleil, Environ. Sci.
Technol., Hi, 1439 (1991).
Disclaimer
The information in this document has been funded by the United
States Environmental Protection Agency. It has been subjected to
agency review and approved for publication.
TABLE 1. DETECTION LIMITS FOR PORTABLE GAS CHROMATOGRAPHS
CALCULATED FROM FIELD CALIBRATIONS
(parts per
billion by volume)
Benzene Toluene
Photovac 10SPLUS
MSI 301
Sentex Scentograph
HNU 311
SRI 8610 PID
SRI 8610 ELCD
0.5
6.7
3.8
2.7
0.4
NR
1.2
20.5
4.3
3.6
0.3
NR
Tetrachloro-
ethylene
0.5
INT
3.4
4.9
0.2
2.4
Chi or o-
benzene
1.5
INT
7.6
4.2
0.2
5.2
INT Interference. Another peak or an elevated baseline made it
impossible to calculate detection limit.
NR No response to electrolytic conductivity detector.
-------
TABLE 2. COMPARISON OF CANISTER GRAB SAMPLE WITH
PORTABLE CHROMATOGRAPH DATA
Several simultaneous colocated samples collected and analyzed on
January 18, 1992 at the French Limited Superfund Site.
(parts per billion by volume)
Time
Benzene
Trichloro- Chloro-
Toluene ethylene benzene
10:30 Canister Grab Sample
Photovac 10S70
Photovac IDS PLUS
MSI 301 #06
MSI 301 |10
Sentex Scentograph
HNU 311
SRI 8610 PID
SRI 8610 ELCD
10:59 Canister Grab Sample
Photovac 10S70
Photovac 10SPLUS
MSI 301 #06
MSI 301 #10
Sentex Scentograph
HNU 311
SRI 8610 PID
SRI 8610 ELCD
11:28 Canister Grab Sample
Photovac 10S70
Photovac 10SPLUS
MSI 301 #06
MSI 301 #10
Sentex Scentograph
HNU 311
SRI 8610 PID
SRI 8610 ELCD
3.3
3.9
4.2
5.0
2.0
11.0
163.0
3.8
NR
1.6
NA
ND
3,0
1.0
ND
ND
2.0
NR
2.7
22.9
3.2
4.0
1.0
ND
ND
3.6
NR
3.1
7.1
3.9
2.0
1.0
ND
88.0
4.6
NR
1.3
NA
2.0
1.0
1.0
ND
ND
2.0
NR
2.5
5.1
2.7
1.0
1.0
ND
1.2
4.1
NR
0.2
ND
0.5
ND
ND
ND
ND
0.0
ND
ND
NA
ND
ND
ND
ND
0.2
ND
0.5
0.1
ND
ND
ND
ND
ND
0.3
ND
0.4
0.3
0.3
0.9
ND
ND
ND
ND
0.8
ND
0.1
NA
ND
ND
ND
ND
0.4
0.8
ND
0.2
4.0
0.5
ND
ND
ND
0.1
0.7
ND
ND Not detected.
NA Not analyzed. Photovac 10S70 calibrated automatically.
NR No response to electrolytic conductivity detector.
-------
TABLE 3. ABSOLUTE VALUES OF DIFFERENCES BETWEEN
CANISTER TO-14 REFERENCE METHOD AND
PORTABLE GAS CHROMATOGRAPH DATA
Absolute differences between concentrations found by the portable
gas chromatograph and concentrations found in a simultaneous
colocated canister grab sample. Samples collected at French
Limited Superfund Site during startup of bioremediation.
January 11 - 19, 1992 (parts per billion by volume)
Benzene
Photovac 10S70 (14 samples)
Maximum 20.3
Mean (STD) 2.5 (5.0)
Minimum 0.3
Photovac 10SPLUS (14 samples)
Maximum 2 . 5
Mean (STD) 1.0 (0.6)
Minimum 0.1
MSI 301 #06 (13 samples)
Maximum 12 . 0
Mean (STD) 2,3 (2.9)
Minimum 0 . 5
MSI 301 #10 (13 samples)
Maximum 4 . 4
Mean (STD) 1.7 (1.1)
Minimum 0 . 2
Toluene
4.1
1.3 (1.1)
0.0
2.0
0.8 (0.6)
0.0
3.2
1.2 (0.7)
0.3
41.0
6.8 (13.3)
0.0
Sentex Scentograph (13 samples)
Maximum 7.7 3.5
Mean (STD) 2.2 (1.9)
Minimum 0.4
HNU 311 (15 samples)
Maximum 159 . 7
Mean(STD) 11.9 (39.5)
Minimum 0 . 1
SRI 8600 PID (13 samples)
Maximum 6 . 2
Mean (STD) 1.6 (1.5)
Minimum 0 . 1
SRI 8610 ELCD (13 samples)
Maximum
Mean (STD) NA
Minimum
1.9 (0.8)
1.0
84.9
7.1 (20.8)
0.4
7.6
1.8 (1.8)
0.2
NA
Trichloro-
ethene
0.3
0.1 (0.1)
0.0
2.7
0.3 (0.7)
0.0
0.2
0.1 (0.1)
0.0
0.2
0.1 (0.1)
0.0
0.2
0.1 (0.1)
0.0
0.2
0.1 (0.1)
0.0
1.4
0.3 (0.4)
0.0
3.4
0.5 (0.9)
0.0
Chloro-
benzene
7.8
1.2 (2.1)
0.0
1.7
0.4 (0.5)
0.0
0.3
0.1 (0.1)
0.0
0.3
0.1 (0.1)
0.0
0.3
0.1 (0.1)
0.0
0.3
0.1 (0.1)
0.0
1.6
1.0 (0.3)
0.6
7.2
2.1 (2.9)
0.0
NR No response to electrolytic conductivity detector.
-------
TECHNICAL REPORT DATA
1. REPORT NO.
EPA/600/A-92/249
PB93- 121051
4. TITLE AND SUBTITLE
Evaluation of Commercially-Available Portable Gas
Chromatographs
5.REPORT DATE
6.PERFORMINa ORGANIZATION CODE
7. AUTHOR(S)
R. Berkley, M. Miller, J, Chang, K. Oliver,C.
Fortune
8.PERFORMING ORGANIZATION REPORT
NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Atmospheric Research and Exposure Assessment
Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
1Q.PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-D8-0002
12. SPONSORING AGENCY NAME AND ADDRESS
Atmospheric Research and Exposure Assessment Lab
RTF
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD COVERED
Conference Proceedings
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Six commercially-available portable gas Chromatographs (PGC) were evaluated at
a Superfund site during startup of bioremediation. Concentrations of volatile
organic compounds (VOC) were slightly above ambient background levels. Concurrent
colocated grab samples were collected periodically in Summa-polished canisters.
They were analyzed by Method TO-14 using a mass-sensitive detector. The grab
samples served as standards to assess the accuracy of data reported by the PGCs.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED
TERMS
c.COSATI
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (Ms Report)
UNCLASSIFIED
21 .NO. OF PAGES
7
20. SECURITY CLASS (Hits Page)
UNCLASSIFIED
22. PRICE
1
-------
------- |