EPA/600/A-94/193
Overview of VOC Measurement Technology in the PAMS Program
W. A. Lonneman
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina
ABSTRACT
The PA Ms program began in early 1993 with an Implementation Teleconferencing Workshop for
states and EPA regional offices to discuss the methodologies and programs available for assistance to
support the enhanced monitoring regulations in the 1990 Clean Air Act Amendments. The teleconference
was arranged by the Office of Air Quality Planning and Standards (OAQPS) with technical discussions
from members of the Atmospheric Research and Exposure Assessment Laboratory (AREAL), Since then
several states have commenced VOC measurement programs particularly by the operation of automated
gas chromatograph (gc) systems. Quality assurance programs have been implemented to track
performance of the measurement procedures. A number of issues were discussed at the Teleconference
concerning operational components of the gc methodology. An overview of these issues is presented in
this paper along with laboratory test results to support the conclusions.
This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's
peer and administrative review policies and approved for presentation and publication. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
-------
INTRODUCTION
In late April 1993 the Office of Air Quality Planning and Standards (OAQPS) sponsored a
Photochemical Assessment Monitoring Stations (PAMS) Teleconference Workshop 1 for states and EPA
regional offices to discuss the monitoring and methodology requirements for the Enhanced Ozone
Monitoring Regulation in the 1990 Clean Air Act Amendments.2 Before the workshop, the states and
EPA regional offices received a Technical Assistance Document3 that described current state-of-the-art
methodologies available for "the PAMS program. The use of gas chromatographic (gc) techniques to
measure speciated hydrocarbons is perhaps the most challenging analytical procedures required of the
affected states in the PAMS program. The recommended approach consists of automated gc systems
capable of obtaining hourly hydrocarbon measurements. The regulation2, however, allows an alternate
approach of collecting ambient air in canisters for various integrated time periods of 3, 6, and 24 hours
followed by gc analysis. Regardless of the method of sample collection, the gc analysis systems have
several similarities. All gc systems utilize some type of sample preconcentration to obtain measurable
levels of the hydrocarbons. The preconcentrated sample is injected onto a gc column(s) to separate the
hydrocarbons which are detected by a flame ionization detector (FID). The FID has a sensitive, linear
response to organic compounds that contain carbon and hydrogen. The signal produced in the FID is
amplified and integrated to produce peak area counts that are converted to parts-per-billion carbon (ppbC)
concentration by a calibration factor. These and other operational aspects of the gc system are discussed
in the Technical Assistance Document3 and were expanded upon during the Teleconference Workshop1.
Operational aspects of greatest concern were referred to as issues and included:
~ Uniform Response of the FID
~ Water Management System
~ Canister Storage of Air Samples
~ Quality Assurance Program
~ Data Base Development Approaches
The purpose of this paper is to update and provide more recent information concerning these
issues one year after the Teleconference Workshop1.
EXPERIMENTAL
The C2 to C,2 hydrocarbons were separated on a 60-m x 0.32-mm id DB-1 fused silica capillary
column coated with a 1-^m bonded liquid phase (J & W Scientific, Folsom, CA). Column temp
conditions included a -50°C initial temp held for 2 min followed by temp programming at a rate of 8°C
per min to a final temp of 200°C.4 Air samples for analyses were approximately 500 cm3 prepared by
a manual preconcentration procedure. Details for the gc system, preconcentration procedure and
calibration are provided elsewhere5. All samples were collected in 6-1 stainless steel Summa canisters.
In some tests air samples were dried using a Nafion dryer (Model MD 125-48F, Perma Pure Dryer,
Toms River, NJ). The Nafion dryer consists of concentric tubes with the Nafion membrane inside a
Teflon tube shell. Samples were taken at room temp (— 24°C) at a flowrate of 100 cm3 min'1. During
sample collection a counter-current flow of 200 cm3 min1 zero dry air was routed across the outside
Nafion membrane to facilitate water removal.
-------
RESULT AND DISCUSSION
Uniform Response of the Flame Ionization Detector
Since the development of the FID in 19586, uniform response to hydrocarbons has been observed.
The earliest documentation of FID response uniformity was reported by Steinberg7. He reported effective
carbon response of unity for paraffinic and aromatic compounds and 0.95 for olefinic compounds. The
only hydrocarbon type not near unity was acetylenic compounds with a 1.30 effective carbon response.
Other later reports observed very similar FID response characteristics8-9. Recently FID response
characteristic were determined with a 16 compound gravimetric standard10 from the National Institute of
Science and Technology (NIST). All compounds exhibited a per carbon response of 1.00 ± 5% with
the exception of acetylene at 1.16.
Because all PAMS target compounds are hydrocarbons, FID calibration is simplified by the use
of a single compound. The approach recommended in the Technical Assistance3 document is the use of
a propane-in-air Standard Reference Material (SRM) available from NIST. Utilization of the 3 ppm
concentration provides the capability of preparing diluted concentration levels in the range of typical
ambient compounds. Some automated GC systems utilize two gc columns and dual FIDs requiring a two
hydrocarbon calibration standard. A propane-benzene mixture is suggested for these systems although
other similar boiling point compound mixtures could be utilized. A two compound mixture is currently
not available from NIST. Such a mixture is available as a Certified Reference Material (CRM) from
several gas manufacturers. NIST previously offered a benzene-in-nitrogen SRM, however, this single
compound standard is no longer available.
Water Management Systems
Other than nitrogen and oxygen, water vapor is the most abundant component in ambient air.
Its concentration is several orders of magnitude higher than the sum of nonmethane hydrocarbons. The
presence of such large quantities of water is a problem to gc systems that utilize sub-ambient temperatures
for both the sample preconcentration and gc column components of an automated gc system. Water vapor
forms ice under these conditions, resulting in trap and/or column plugging and subsequent carrier flowrate
disruptions. Thus some type of water management approach is required. Several approaches to control
water vapor injection onto the gc system are available. A hydrophobic absorbent material in the
preconcentration step can selectively collect the hydrocarbons while passing water through the system.
These absorbent materials however do not efficiently trap the Q-Cj hydrocarbons. Another approach
involves the controlled temp injection of the trapped ambient air components onto a second cryo-focusing
trap prior to column injection to minimize the amount of water injected onto the gc column. With this
approach, care must be taken to insure quantitative transfer of the heavier molecular weight Cg-C,2
compounds. The approach utilized on most of the automated gc systems, particularly at the time of the
Teleconference Workshop1, was a Nafion dryer in the system inlet to selectively remove water vapor
prior to preconcentration of a sample. Nafion is a perfluorocarbon co-polymer membrane that contains
sulfonic acid side chain groups. Water vapor is attracted by the sulfonic acid groups and transported
through the membrane where it is flushed away by a counter-current flow of dry hydrocarbon free air.
While Nafion is commonly used in several automated gc systems, it has not been fully evaluated for the
PAMS target compounds. Water soluble organic compounds such as the carbonyls and alcohols are
expected to be at least partially removed by Nafion. For the PAMS target compounds only oc- and
/3-pinene are expected to be affected by the Nafion dryer. Burns et. al." have reported the transformation
of C10 terpene compounds to other compounds by Nafion.
-------
To test the adequacy of Nafion several types of sample mixtures were analyzed. The gc system
utilized for these tests employees a controlled water vapor injection approach to analyze air samples that
contain water vapor. Canister samples were analyzed with and without the Nafion dryer. Table I
contains the results for the analysis of 7 different canister mixtures that contained PAMS hydrocarbons
as well as other VOCs. Repeat analyses were performed on 2 of the 7 mixtures. Samples taken without
Nafion are identified as direct. In Table I all VOC compounds, including the PAMS compounds, are
included in the total nonmethane organic compounds (TNMOC). Three of the 7 mixtures evaluated
demonstrated practically no effect from the use of the Nafion dryer. These were the PAMS proficiency
mixtures 1 and 2 and the Naphtha HC mixture. The proficiency mixtures consisted of only 18 PAMS
hydrocarbons, none of which were oc- and /3-pinene in ultra pure nitrogen. Trace level concentrations
of about 50 other compounds were also observed in the proficiency mixtures, however these levels were
too insignificant to affect TNMOC. The Naptha mixture consisted of about 112 aliphatic hydrocarbons,
25 of which were PAMS target compounds.
Significant differences in TNMOC were observed for the remaining 4 mixtures. The PAMS
hydrocarbon mixture contained n-decane and n-undecane in addition to all 55 PAMS hydrocarbons in
humidified HC free air. The mixture also contained 257.2 and 160.4 ppbC oc- and /3-pinene
respectively. For both sample analyses through Nafion oc - and /3-pinene were completely removed. The
sum of other peaks (—157 ppbC) however could not account for the loss of oc - and /3-pinene. Nafion
appears to effect 2 other PAMS compounds, 2-methyl-l-pentene and styrene. For the two PAMS sample
analyses given in Table I, 15 and 73% 2-methyl-l-pentene respectively were removed by Nafion.
GC/MS analysis suggests that Nafion converts 2-methyl-l-pentene quantitatively to another compound
tentatively identified as 2-methy 1-2-pentene. Styrene loss of 6 and 18% respectively were observed when
the PAMS sample was twice taken through Nafion. No obvious product peaks were observed for styrene
loss. The other 51 PAMS compounds were unaffected by Nafion as evidenced by typical 99% compound
reproducibilities.
A Mexico City fuel vapor sample was also run and was interesting in that it contained 48 PAMS
target compounds all of which excellently reproduced when re-analyzed through Nafion. The only major
loss upon re-analysis through Nafion was methyl-rm-butylether (MTBE), an oxy-fuel additive. MTBE
at 126.1 ppbC was completely removed by the Nafion dryer.
TNMOC for both the diluted auto exhaust and ambient air samples were significantly lowered
when taken through the Nafion dryer. These complex mixtures consisted of more than 250 peaks, some
of which were tailing peaks thought to be oxygenated hydrocarbons. These tailing peaks appeared to be
most affected by Nafion. Chromatograms are compared in Figure 1 to demonstrate the Nafion drier's
effect on peak reproducibility. Tailing peaks at 11.5 and 15.5 min are nearly completely removed when
the sample was taken through Nafion. Other noticeable differences occur in peaks after 30.3 min. The
peak at 30.1 min is n-undecane.
The effect that Nafion has on only the PAMS compounds is shown in Table II. Only the PAMS
hydrocarbon mixture containing the 55 PAMS hydrocarbons was affected by Nafion. As explained earlier
the difference in the two totals can be attributed to the complete conversion of oc - and /3-pinene to other
compounds and the partial losses of 2-methyl-l-pentene and styrene. All other PAMS compounds appear
not to be affected by Nafion. More tests are needed to confirm this conclusion.
-------
Storage of PAMS Compounds In Canisters
For automated gc systems, canister storage is not generally an issue. Samples are directly
introduced into the preconcentration trap over an integrated time period followed by injection onto the
gc column. Some automated gc systems collect an integrated air sample into a canister from which an
aliquot is then taken for analysis. Even with this approach canister storage is not a major concern. The
regulation2, however, does allow for the collection of canister samples over 3, 6, and 24 hour periods.
In fact the regulation requires the collection of a 24-hour integrated sample every six days at PAMS sites.
For these samples, canister storage is an issue since at least more than a day and in most cases several
days storage occur prior to analysis. In previous years canisters have been routinely used to collect air
samples in the field and returned to the laboratory for analysis. It is not uncommon to store canister
samples for a week or more.
Although most storage results are not generally documented and reported, ambient air samples
are found to be stable even though storage periods may be prolonged. To demonstrate the suitability of
canister storage, Table III compares the sum of all peaks, TNMOC and the sum of only the PAMS peaks
for a PAMS mixture stored in two, 6-1 Summa canisters over a 7 month period. Both canisters showed
excellent reproducibility for both TNMOC and PAMS compounds. Although the data are not presented,
all of the 53 PAMS hydrocarbons in the mixtures demonstrated similar, excellent reproducibility.
Figure 2 also demonstrates Summa canister storage stability by comparing the ehromatograms of ambient
air analyzed before and after 85 days of storage. The nearly 3 month storage period had very little effect
on sample integrity. Largest differences occur with tailing peaks at retention times 11.5, 15.0, 25.5, and
30.7 min, none of which are PAMS compounds. The occurrence of change to any peak however
suggests that storage periods are not unlimited. Thus even though air samples appear to store well in
canisters, a minimized storage period prior to analysis is good laboratory practice.
Quality Assurance and Data Base Development
Operational components of the gc system that affect quality assurance were discussed at the
Teleconference Workshop1. These issues primarily focused on the operation of gc integrators and data
processing systems such that accurate gc peak areas and correct peak identifications would result. GC
systems, automated or manual, that preconcentrate 200 cm3 of ambient air are capable of detecting and
resolving peaks below 1.0 ppbC. Typical ambient air ehromatograms contain more than 130 peaks
ranging" in concentrations as much as two orders of magnitude above and one order below the 1.0 ppbC
level. Several peaks are only partially resolved from neighboring peaks, and on occasion peaks co-elute
on the tail of polar VOC peaks. Careful initial selection of integrator peak detection and integration
parameters are well worth the effort for accurate peak area determinations. Forced baseline points and
tangential skims are needed for tailing, coeluting and other peaks of a typical ambient air chromatogram.
After initial setup, operators will need to routinely look at the baselines of all or a sub-set of the daily
sample ehromatograms to determine the adequacy of integration parameters. The purpose of this QA
procedure is to refine the gc integration parameters to yield the best possible results.
Correct peak naming should be a companion part of this QA activity. Verification of each
individual gc peak is desirable but perhaps too tedious particularly for gc operations generating 24
samples daily and 168 samples weekly. A more suitable approach would be the verification of correct
reference peak selection. Peak naming software uses reference peaks to make retention time corrections
before look-up of retention times in a calibration table of retention times and peak names. The reference
peaks are prominent peak which are always present. If reference peaks are properly selected then most
likely all other peaks will be properly named.
-------
It is strongly recommended that both peak integration and reference peak naming verification
procedures be implemented early in the setup of the PAMS gc systems. Likewise collected data should
not be incorporated into a data base system such as AIRS until all data results are properly verified.
REFERENCES
1. National Photochemical Assessment Monitoring Stations Teleconference Workshop, Sponsored
by Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, held
at North Carolina State University, April 27-29, 1993.
2. U.S. Environmental Protection Agency, Code of Federal Regulations, Title 40, Part 58,
Enhanced Ozone Monitoring Regulations, Washington, DC, Office of the Federal Register,
August 23, 1991.
3. U.S. Environmental Protection Agency, "Technical Assistance Document for Sampling and
Analysis of Ozone Precursors," EPA/600/8-91-215, October 1991.
4. Lonneman, W. A., R. L. Seila, J. V. Daughtridge, and H. G. Richter, "Results from the
Canister Sampling Program Conducted During the 1990 Atlanta Precursor Study," in Proceedings
of the 84th Annual Meeting of the Air and Waste Management Association. Vancouver, B. C.,
1991, Paper 91-68.2.
5. Seila, R. L., W. A. Lonneman, "Determination of Q to C,2 Ambient air Hydrocarbons in 39
U.S. Cities from 1984 through 1986," EPA/600/3-89-058, 1989.
6. Harley, J., W. Nel, and V. Pretorius, Nature. Vol. 181, p. 177, 1958.
7. Sternberg, J. C., W, S. Gallaway, and D. T. Jones, International Gas Chromatography
Symposium. 1961.
8. Dietz, W. A., "Response Factors for Gas Chromatographic Analyses," J. Gas Chromatoe. 5. p.
68, 1967.
9. Blades, A. T., "Ion Formation in Hydrogen Flames," Can. J. Chem.. 54, p. 2919, 1976.
10. Apel, E. C., J. G. Calvert, R. Zika, M. O. Rodgers, V. P. Aneja, J. F. Meagher, W. A.
Lonneman, "Hydrocarbon Measurements During the 1992 Southern Oxidants Study, Atlanta
Intensive: Protocol and Quality Assurance," submitted to Journal Air Waste Management, April
1994.
11. Burns, W. F., D. T. Tingey, R. C. Evans, and E. H. Bates, "Problems with a Nafion Membrane
Dryer for Drying Chromatographic Samples," J. Chromatog. 269, p. 1, 1983.
KEY WORDS
Clean Air Act Amendments, cryogenic preconcentration, Enhanced Ozone Monitoring Regulations, FID,
flame ionization detection, gas chromatography, hydrocarbons, Nafion Dryer, PAMS, peak
identification, peak integration, quality assurance, speciated hydrocarbons, VOCs.
-------
Table I. Effect of Nafion on TNMOC.
Sample \ Analysis
Direct Analysis, ppbC
Analysis through Nafion, ppbC
PAMS HC mixture
8,753
8,344
PAMS HC mixture
8,766
8,313
Naphtha HC mixture
1,089
1,073
PAMS proficiency #1
259
259
PAMS proficiency #2
231
232
Mexico fuel vapor
1,161
1,033
Dilute auto exhaust
7,554
7,141
Dilute auto exhaust
7,414
7,079
Ambient air
430
367
Table II. Effect of Nafion on PAMS compounds.
Sample \ Analysis
Direct Analysis, ppbC
Analysis through Nafion, ppbC
PAMS HC mixture
8,357
7,756
PAMS HC mixture
8,210
7,718
Naphtha HC mixture
690
680
PAMS proficiency #1
228
223
PAMS proficiency #2
219
221
Mexico fuel vapor
932
921
Dilute auto exhaust
6,046
6,007
Dilute auto exhaust
5,933
5,974
Ambient air
259
267
Table in. Storage of PAMS compounds in Summa canisters.
Canister ID
Measurement
Initial Cone, ppbC
Final Cone, ppbC
Canister A
TNMOC
1,035
1,155
E PAMS cmpds
955
968
Canister B
TNMOC
1,103
1,018
E PAMS cmpds
1,136
1,007
-------
w
cn
DtJLECT
CO
Jb
"i 1 1 1 1 1 1 1 1 i r-
12 14 16 18 20 22
LtjUlM— '
illjl—uxil
li
DIRECT |
1 .1 nmu NATION j
i r 1 1 1 r
22 24 26 21 30 32 34
RETENTION TIME minutes
Figure 1. Analysis of a PAMS Compound Mixture
Direct and Through a Nafion Dryer
-------
• 2 4 t S 19 13 14
W
t/D
£
o
(/}
<
s
p*
u
o
MarcfeSl.DM
xjl lljLalji- A_iL A
UaJ 7 S.1W4
RETENTION TIME minutes
Figure 2. Storage of an RTP,NC Ambient Air Sample
in a Summa Polished Canister
9
-------
TECHNICAL REPORT DATA
1. REPORT NO.
EPA/600/A-94/193
2 .
- mi mi iiiiii mi
PB95-122644
4. TITLE AND SUBTITLE
OVERVIEW OF VOC MEASUREMENT TECHNOLOGY IN THE PAMS
PROGRAM
5,REPORT DATE
S.PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
William A. Lonneman
8.PERFORMING ORGANIZATION REPORT
NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Atmospheric Research and Exposure Assessment Lab
Office of Research and Development
U.S. Environmental Protection Agency-
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Atmospheric Research and Exposure Assessment Lab
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13.TYPE OF REPORT AND PERIOD
COVERED
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The PAMS program began in early 1993 with an Implementation Teleconferencing
Workshop for states and EPA regional offices to discuss the methodologies and
programs available for assistance to support the enhanced monitoring regulations in
the 1990 Clean Air Act Amendments. The teleconference was arranged by the Office
of Air Quality Planning and Standards (OAQPS) with technical discussions from
members of the Atmospheric Research and Exposure Assessment Laboratory (AREAL).
Since then several states have commenced VOC measurement programs particularly by
the operation of automated gas chromatograph (gc) systems. Quality assurance
programs have been implemented to track performance of the measurement procedures.
A number of issues were discussed at the Teleconference concerning operational
components of the gc methodology. An overview of these issues is presented in this
paper along with laboratory test results to support the conclusions.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/ OPEN ENDED
TERMS
C.COSATI
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This
Report)
UNCLASSIFIED
21.NO. OF PAGES
20. SECURITY CLASS (This
Page)
UNCLASSIFIED
22. PRICE
------- |