United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-80-008
April 1980
Air
Guidance for the
Collection and Use of
Ambient Hydrocarbon Species
Data in Development of
Ozone Control Strategies
-------
EPA-450/4-80-008
GUIDANCE FOR THE COLLECTION AND USE
OF AMBIENT HYDROCARBON SPECIES DATA
IN DEVELOPMENT OF OZONE CONTROL STRATEGIES
Final Report
by
Hanwant B. Singh
Atmospheric Science Center
SRI International
Menlo Park, California 94025
Contract 68-02-2548
EPA Project Officer:
Dr. Harold G. Richter
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
-------
This document is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of-readers. Copies are
available free of charge to Federal employees, current EPA contractors
and grantees, and nonprofit organizations - in limited quantities - from
the Library Services Office (MD-35) , U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711; or, for a fee, from the
National Technical Information Service, 5285 Port Royal Road, Springfield,
Virginia 22161.
This report was furnished to the Environmental Protection Agency by
SRI International, 333 Ravenswood Avenue, Menlo Park, California 94025,
in fulfillment of Contract No. 68-02-2548. The contents of this report
are reproduced herein as received from SRI International. The opinions,
findings and conclusions expressed are those of the author and not
necessarily those of the Environmental Protection Agency.
Publication No. EPA-450/4-80-008
11
-------
ABSTRACT
This document provides a set of standard procedures that can be routinely and reliably
depended upon to obtain nonmethane hydrocarbons (NMHC) species composition of the ambient air.
The document discusses the needs that exist for the quantification of ambient hydrocarbon (HC)
species especially as they relate to air quality management applications. Procedures for site
selection, sample collection and storage, calibrations and analysis and quality assurance and control
are described. State of the art HC species analysis techniques are unable to identify 10 to 35 percent
of the ambient NMHC burden. Recommendations are made to fill gaps in our current knowledge.
iii
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgments viii
Executive Summary 1
1. Introduction 5
2. Objectives 6
3. General Approach 7
4. Utility of HC Species Data 8
5. Site Selection, Sample Collection, Storage, and Integrity 10
A. Site Selection 11
B. Presampling Considerations 13
C. Sample Storage and Integrity 14
D. Recommended Overall Sampling Procedure 19
6. Analysis of HC Species 25
A. Mode of Operation 25
B. Instrument Optimization, Sample Preconcentration, and Injection 25
C. Chromatographic Separations and Analyses of HC Species Data 28
D. Routine Identification of HCs 40
E. Primary Standards and NMHC Species Calibrations 42
F. Archiving HC Data 43
7. Quality Control (QC) and Quality Assurance (QA) 44
8. Recommended Means for Obtaining HC Species Data 46
9. Knowledge Gaps and Recommendations 47
Appendix — Analysis of Available Ambient HC Species Data 48
References 53
-------
FIGURES
Number Page
1 Sequence of events for the collection and analysis of ambient HC species data 2
2 Maximum roadway contribution to concentration at different distances 13
3 Specially treated 1-liter stainless-steel sampling vessel for collecting grab samples 18
4 Collection of integrated bag sample and transfer to metal container 24
5 A schematic diagram of sample preconcentration and injection steps 27
6 Light HC (C2-C6) analysis in rural air 31
7 C2-C6 HC species analysis of an auto exhaust air sample 32
8 Typical ambient air chromatogram showing resolution of C2~C5 hydrocarbon fraction ... 33
9 Chromatogram showing resolution of a standard mixture of C5-C10 hydrocarbons
on SE-30 glass capillary column with subambient temperature programming 35
10 Chromatogram showing resolution of a standard mixture of C5-C10 hydrocarbons
on SE-30 glass capillary column with above ambient temperature programming 37
11 Chromatogram of ambient air (500 ml) samples directly from a stainless-steel
container ; 39
12 HPLC-UV separation of a synthetic mixture of hydrazones
of C1-C4 aliphatic aldehydes, benzaldehyde, and acetone 40
13 Ambient air chromatogram showing resolution of Fluorocarbon-11
and other halocarbons 41
A-1 Locations of selected monitoring sites in the Houston area 49
A-2 Southern California sampling sites for 1975 50
VI
-------
TABLES
Number Page
1 Air Quality Models and Their Hydrocarbon Input Requirements 10
2 Advantages and Disadvantages of Selected Collection Mediums
for Ambient HC Species Analysis 15
3 Comparison of HC Species Analysis Measured Directly and After Collection
on Tenax Cartridges 16
4 Stability of Selected Light HCs (C2-C5) in Passivated Stainless Steel Containers 20
5 Stability of Selected Heavy Hydrocarbons (C5-C10) in Passivated
Stainless Steel Containers (ng/m3) 21
6 Stability of Selected Oxygenated Species in Passivated Stainless Steel Containers 22
7 Optimum Operating Conditions for Model 5711A (Hewlett-Packard) FID 26
8 GC Column Selection and Operating Conditions for Detailed HC Species Analysis 30
9 Standard Gas Mixture Measured on SE-30 Capillary Column with Subambient
Temperature Programming 36
10 Standard Gas Mixture Measured on SE-30 Glass Capillary Column with Above
Ambient Temperature Programming 38
A-1 Houston Ambient HC Composition Based on Data Collected by RTI
in August-September 1977 50
A-2 Houston Ambient HC Composition Based on Data Collected by EPA
in September 1978 51
A-3 South Coast Air Basin (SCAB) Ambient HC Composition Based
on Data Collected by ARB in August-September 1975 51
vii
-------
ACKNOWLEDGMENTS
Thanks are due to Dr. Harold Q. Richter, EPA Project Officer, and Mr. E.L. Martinez for their
assistance and encouragement in the preparation of this report.
Cooperation and assistance offered by Dr. Chester W. Spicer (Battelle Columbus Labora-
tories), Dr. George Tsou (California Air Resources Board), Mr. William Lonneman (Environmental
Sciences Research Laboratory), Mr. Martin A. Ferman (GM Research Laboratories), Mr. John M.
Harden (Research Triangle Institute) and Dr. Hal Westberg (Washington State University) is
especially appreciated.
From SRI, the help of Mr. Francis L. Ludwig, Dr. J. Raul Martinez, Mr. Leonard A. Cavanagh,
and Mr. Christopher Maxwell is gratefully acknowledged.
viii
-------
EXECUTIVE SUMMARY
A measure of ambient hydrocarbon (HQ* composition is an essential input to air quality models
used to determine HC reductions required to achieve the federal ozone (03) standard. To date, no
standardized procedures for HC measurements in ambient air exist. The major purpose of this
document is to describe and recommend a set of standardized procedures that can be routinely
depended upon to provide reliable HC species data for air quality management..The detailed
methods presented here, provide the best and simplest means for measuring ambient concentrations of
total nonmethane hydrocarbons (NMHC),* structurally similar HC groups '(e.g.,alkanes, alkenes,
aromatics), and individual HC species. It is not implied, however, that other methods are necessarily
unreliable. This document also addresses the needs of air quality models, possible characterization of
HC composition from limited species measurements (based on analysis of existing data), quality control
procedures, and knowledge gaps.
Air quality models were examined to demonstrate that detailed HC species measurements
provide a good means to satisfy some of the models' needs as well as other air pollution
requirements. Continuous NMHC instruments that have been used in the past and satisfy the needs
of the simplest models (such as EKMA) are considered unreliable at NMHC levels of less than about
1 ppm C. A great deal of ambient NMHC data in typical urban/suburban locations lies below 1 ppm
C. The entire range of ambient NMHC concentrations (typically 0.1 to 3 ppm C) can be accurately
measured only by HC species analysis. HC species data are also highly desirable because they allow
HC reactivity comparisons, act as indicators of sources, and permit the determination of exposure to
certain toxic chemicals (such as benzene). However, species measurements cannot be made
continuously and require specialized instrumentation and personnel. For special modeling
applications, especially in regions where ambient NMHC concentrations are above 1 ppm C,
continuous instrumentation can be highly beneficial.
Figure 1 summarizes the sequence of events that occur during the conduct of HC species
analysis. The overall task of HC species analysis can be broadly divided into the following
categories:
• Site selection
• Sample collection and storage
• Instrumentation and analysis
• Quality control and assurance.
•The terms HC and NMHC can be used interchangeably with VOC (Volatile Organic Compounds) and NMOC (Nonmethane
Organic Compounds), respectively, for the purposes of this document.
-------
SITE SELECTION
SAMPLE COLLECTION
• SELECT SAMPLE INTEGRATION TIME
• COLLECT 75-L INTEGRATED SAMPLE IN BAG; TRANSFER TO
TWO 2-L POLISHED SS CONTAINERS AT 40 PSI
• ANALYZE FOR HC SPECIES WITHIN 1 WEEK; DO NOT
EXCEED 2 WEEKS
SENSITIVITY
Ippbc
PRE-CONCENTRATION
• SAMPLE VOLUME — 500-ml
• PRETRAP — 4 in. X 1/8 in. O.D. SS PACKED
WITH 80/100 MESH GLASS BEADS
• COLLECT SAMPLE AT LIQUID O2 OR Ar
TEMPERATURE
• DESORB AT 90 — 100°C (BOILING WATER)
LIGHT HYDROCARBONS
HEAVY HYDROCARBONS
C2-C6HCs
• 5m X 1/16 in. I.D.. Ni, PHENYL ISOCYANATE/
PORASILC COLUMN
• ISOTHERMAL ANALYSIS AT 40°C
C6-C10(ALSO>C10) HCs
• 30m X 0.25-mm SE30 GLASS CAPILLARY COLUMN
(SUBAMBIENT PROGRAMMED ANALYSIS)
• 50m X 0.5-mm SE30 GLASS CAPILLARY COLUMN
(ABOVE AMBIENT PROGRAMMED ANALYSIS)
CALIBRATIONS
• PROPANE/HEXANE STANDARDS IN AIR
• BENZENE/TOLUENE STANDARDS IN AIR
QUALITY CONTROL
AND
QUALITY ASSURANCE
DATA SYSTEMS
• COMPUTERIZED INTEGRATING
DATA SYSTEM
ERRORS
OVERALL PRECISION ±5%
OVERALL ACCURACY ±10 TO 15%
OUTPUT
• TOTAL NMHC
> HC SPECIES 65 — 90% IDENTIFIED;
REMAINDER UNIDENTIFIED
Figure 1. Sequence of events for the collection .and analysis
of ambient HC species data.
-------
Briefly, a site should be located away from areas having large source gradients and should be
representative of as large an area as possible. Areas with strong source gradients would normally
require a more intensive sampling network. The geographical representativeness of a site must be
estimated. In general, open urban areas away from large and even small sources of HCs (e.g.,
refineries, gas stations, roadways, 03 instrumentation) are preferred. The sampling air inlet sould be
3 to 15 m above ground level and at least 1 m above a support surface. Analysis of available
information (e.g., maps and aerial pictures, population densities, land use, meteorological and
climatological data, emissions inventories, existing monitors) and consultations with those
responsible for selecting existing NMHC monitoring sites are a good practical approach to site
selection.
Once the site has been selected, other details such as power availability and shelter should be
established. A sample integration time (e.g., a 3-h, 1-h, or instantaneous sample) should be fixed
before collecting the air sample. Stainless steel metal bellows pumps (MB-158, MB-151, MB-41,
MB-21) are recommended for sampling. After considering several sampling methods a combination
bag/metal container approach is recommended. A 75 to 100 liter, 2 mil Tedlar bag (or a 5 mil Teflon
bag) provides an easy means to collect an intergrated (say, 3-h) sample. Such bags should be cleaned
thoroughly before testing and kept in a cool «70°F)and dark place before and during sampling. Once
the Tedlar bags are filled, the air should immediately be transferred to two evacuated, 2-liter
electropolished stainless steel canisters at 35 to 40 psi pressure (both canisters should be analyzed
for detailed HC species). Prior to sample pressurization, these evacuated canisters (previously
cleaned) should be flushed with the bag air (5 to 10 liters per canister should be sufficient) and
reevacuated to ensure that any residual air is the same as that in the bag. All 03 is essentially
destroyed during these transfers and no nitric oxide spiking is considered necessary. For collecting
instantaneous samples, the bag step (necessary for time integration) should be eliminated and
stainless steel vessels pressurized directly.
Once collected, air samples should be analyzed as quickly as possible, but a 1 week delay may
be tolerated. A delay of 2 weeks is unacceptable. C2 to CioHCs constitute the major identified fraction
of ambient NMHCs. Typically, 10 to 35 percent of the NMHCs are unidentified in ambient HC species
analysis. Because of the established flame ionization detector (FID) response to carbon, the
concentration of unknown NMHCs can be quantified reliably, although this is not possible for
oxygenated species, such as aldehydes, ketones, and acids.
For ease of analysis, it is desirable to divide HC species into two major categories, namely, light
hydrocarbons (C2-C6) and heavy hydrocarbons (C6-Ci2)- The entire NMHC species analysis can be
performed on two gas chromatograph (GC) columns. For modeling purposes, a measurement
sensitivity of 1 ppb Cfor individual HC species is considered acceptable.
The CrC8 HCs are easily measured on a packed GC column (see Figure 1) operated
under isothermal conditions. A sample size of 500 ml is more than adequate. Even under very
humid conditions, water does not pose a problem with this sample size. A glass bead precon-
centration trap held at liquid oxygen or liquid argon temperature is used to trap all HC species
(except methane). Sample volume is measured by accurately monitoring the pressure of a
stainless steel vessel connected to the freezeout trap exit. Desorption is accomplished by
rapidly heating the preconcentration trap electrically or with boiling hot water (90-1 OO'C).
CrC8 analysis under the recommended procedures should not be carried beyond n-hexane.
Species such a 2-methylpentane and n-hexane are also resolved during the heavy HC (>Ce)
analysis. This redundancy is desirable and provides a good internal check.
-------
The complexity of the heavy HC mix requires the use of glass capillary columns for the
desired resolution. A computerized data system is mandatory when capillary columns are used.
An SE-30 glass capillary column is recommended for the analysis of heavy HCs. A sample size
of 300-500 ml is adequate and no water or CO2 interferences are observed. Only HCs in the
C8to Ci0 region (including the natural terpenoid species) are identified. However, the recom-
mended column would also measure HCs higher than C10- For the present, these would be a
part of the unknown HC fraction. Unidentified Cn and higher HC species typically constitute
an extremely small fraction of the total unidentified HC burdens.
No methods currently exist for the reliable measurements of oxygenated species. Alde-
hydes have the potential to play a significant role in photochemical smog formation, even at
very low concentrations, but they do not constitute a large fraction of the ambient HC mix (<5
percent). Interim procedures for the measurement of aldehydes are suggested.
Calibrations and quality assurance procedures are devised to ensure a precision of ±5 percent
for individual species. Samples are collected in two canisters and analyzed in duplicate. Overall
accuracies are limited by the use of average carbon response for calibrations because these can
introduce an error of 5 to 8 percent. In addition, errors on the order of 5 percent in the preparation of
primary standards are to be expected. This would result in an estimated overall accuracy of about 10
percent [(82 + 52)1/2].
It is strongly recommended that all calibration standards be referenced to the National Bureau
of Standards (NBS) propajTe (in air) standard, which has an estimated accuracy of ±2 percent. For
the C2-Ce column, propane is the best standard. Hexane can be measured on both the light HC and
the heavy HC columns and is a good crossover standard. Because ppm levels (e.g., 5 ppm in air) of
benzene and toluene can now be stored in high-pressure aluminum cylinders for more than a year,
these can conveniently be used as standards for the analysis of heavy HCs. An FID carbon response
is an acceptable procedure for the quantification of the concentration of unidentified HCs within an
error of 5 to 8 percent.* Calibration frequencies (twice a day, single point check; once a week,
multipoint check) are recommended recognizing that FIDs are extremely stable and calibrations in
some cases have remained unchanged over a period of several months. It is recommended that at
least 5 percent and preferably 10 percent of the samples be sent for referee analysis.
Practical means to collect HC species data are suggested. A number of shortcomings in the
ability to characterize the ambient HC mix are apparent. Gaps in current knowledge are identified
and recommendations made to fill these gaps. Analysis of samples by contractors is recommended.
'This is generally not true of HCs containing oxygen (oxygenated species) and chlorine atoms (fluorocarbons, etc.).
4
-------
SECTION 1
INTRODUCTION
Ambient hydrocarbons (HCs)* and nitrogen oxides (NOX) are the primary precursors of
ozone (03), a key constituent of photochemical smog. Current strategies for controlling photo-
chemical air pollution depend on HC abatement as the primary means of control. A measure of
ambient HC species concentration is an essential input to determine the HC reductions
required to achieve the federal O3 standard.
Currently, several strategies that employ empirical-kinetic or photochemical diffusion models
are available for purposes of air quality management. However, there are no standardized
procedures for obtaining reliable ambient HC data. The quality of hitherto available data has been
variable at best and even poor at times. Therefore, it is essential to document reliable procedures for
the collection and analysis of HC data with built-in quality control, in an easily usable form. This
document describes a set of procedures that will allow the collection of high quality ambient HC data
at a level of detail that is commensurate with the needs and objectives of the specific application. It
does not summarize methods used by various research laboratories, but rather describes a method
which can be depended upon for routine acquisition of relible HC species data. Other methods not
recommended here are not necessarily invalid. A standardized method is highly desirable, however,
to assure intercomparability of samples collected by different control agencies in different
geographical areas.
'The terms HC and NMHC can be used interchangeably with VOC (Volatile Organic Compounds) and NMOC (Non-
methane Organic Compounds), respectively, for the purposes of this document
-------
SECTION 2
OBJECTIVES
The first and foremost objective is to prepare a guidance document for use by state and local air
pollution agencies that describes and standardizes the procedures for the collection and analysis of
ambient HC species data for the purposes of air quality management. Other important objectives of
this study are to:
. Determine the extent and nature of HC composition data needed to satisfy existing
photochemical air quality models.
• Recommend means for the practical acquisition of HC species data.
• Identify gaps in current knowledge, define uncertainties associated with the best available
methods, and recommend needed future research.
In addition field data have been analyzed to study if a select group of easy-to-measure
HCs can be used to represent the total burden and the composition of a complex ambient mix-
ture of air (Appendix).
6
-------
SECTION 3
GENERAL APPROACH
The contents of this document reflect the cumulative experience of staff members at seven U.S.
laboratories.Detailed information was sought on methods of collecting and analyzing ambient air
samples for hydrocarbon species. These laboratories were requested to provide information in as
much detail as possible, but only on species for which they had direct experience and felt confident
of the reliability of their data. The following organizations were queried:
• Battelle Columbus Laboratories
• California Air Resources Board (ARB)
• EPA-Environmental Sciences Research Laboratory
• General Motors Research Laboratories
• Research Triangle Institute
• SRI International
• Washington State University.
After each group responded to the queries, key investigators were contacted individually for
further discussions. This information was complemented with a review of relevant published
literature. Based on the collected data, a,set of recommended procedures was compiled dealing with
siting, sampling and storage, analysis, and quality control. The task of accurately measuring HC
species depends critically on the satisfactory conduct of each one of these four steps.
In addition, detailed HC species data from a total of 12 sites in Houston and the South Coast air
basin were analyzed. The purpose of this limited analysis of data was to explore whether a few select
HCs, which are easy to measure, might be used to represent the total NMHC burden and/or the
distribution of ambient air at a given site. The possibility of simplifying the otherwise rather complex
process of obtaining NMHC data for those specific applications that allow greater flexibilty was
investigated. The result of this analysis is presented in the Appendix.
-------
SECTION 4
UTILITY OF HC SPECIES DATA
There is general agreement within the scientific community that detailed HC species data are
very useful in designing and evaluating the progress of air pollution control strategies. When reliable
data can be obtained, they have the advantage of meeting many research and air quality
requirements because:
• Detailed HC species measurements have the potential to provide the most reliable measure of
HCs in ambient air. All other methods currently known (e.g., continuous NMHC measure-
ments) are less accurate.
• Detailed HC species data satisfy many of the needs of most air quality models whether
these require as inputs NMHC, alkanes, alkenes, aromatics, aldehydes, C2H4, or a
combination of these chemical groups.
• HC mixtures of widely varying reactivities can be identified and their oxidant forming
potentials studied.
• Detailed HC species data help to characterize and locate sources in an area and can be
used to estimate contributions from these sources.
• Specific HC species (such as benzene) are toxic and determination of exposure to
these species is vital to developing strategies for the control of hazardous chemicals.
Although detailed HC species measurements are needed and desirable, a number of prac-
tical considerations have prevented the widespread application of these techniques, including
the following problems:
• Techniques for measuring HC species require expensive instrumentation as well as expe-
rienced operators.
• These techniques are not standardized and have been used almost exclusively by research
groups.
• Until recently, few people were aware that the continuous NMHC analyzers, which are
relatively easy to operate, generate unreliable data, at least when NMHC levels are less
than 1 ppm C. Unreliable data can be a serious handicap in devising and enforcing sound
control strategies. Continuous NMHC analyzers do have utility under special conditions
when the ambient air is highly contaminated. However, they cannot be relied upon to
provide accurate data over a wide range of NMHC concentrations that are typically
encountered in the ambient air (0.1 to 3 ppm C).
. Species measurement techniques are not continuous and can not satisfy the needs of
models that require continuous HC data.
-------
The HC input requirements of all available models are summarized in Table 1. The follow-
ing HC categories meet the requirements of all air quality models in current use:
. Total NMHCs
• Alkanes
• Alkenes
• Aromatics
• Aldehydes (formaldehyde and higher aldehydes)
• Ethylene.
In addition, species such as acetylene (C2H2) and fluorocarbons are valuable indicators of
sources and degree of contamination. Because no analytical method currently allows the
separate measurement of individual groups of HCs, the only practical solution is the analysis of
all predominant HC species.
-------
TABLE 1. AIR QUALITY MODELS AND THEIR HYDROCARBON INPUT REQUIREMENTS
Model
Type
Empirical-Kinetic
Lagrangian
photochemical
air quality
models
Eulerian Grid cell
photochemical
models
Box models
Name
EKMA
(EPA)'
DIFKIN
(ERT)t
REM
(PES)*
ARTSIM
(ERT)
ELSTAR
(ERT)f
SAI model
(SAI)§
SAI model
(SAI)§
IMPACT
(F and S)**
SULFA3D
(ERT)f
LIRAQ-2
(LLL)ft
EPA
HC Input
Required
Total nonmethane
hydrocarbon (NMHC)
NMHC
C3H6 and a generic
HC that represents
all other nonolefinic
hydrocarbons
Alkanes, alkenes.
aromatics, formalde-
hyde, and higher
aldehydes
Alkanes, alkenes,
aromatics, and
aldehydes..
Alkanes, alkenes.
aromatics, and
aldehydes
Same as SAI
Lagrangian
Same as
ARTSIM
Alkanes, alkenes.
and aldehydes
Similar to SAI model
Remarks
Model uses "NMHC," which is assumed to be a
mixture of surrogate C3H6 and n-C4H10. The
model is validated against the Bureau of Mines Smog
Chamber data. User must specify the surrogate
CjH6/NMHC and n-C4H10/NMHC and aldehyde/
NMHC ratios. Default values for these ratios are
0.25, 0.75, and 0.05, respectively. These ratios
gave the best fit to the smog chamber data but are
otherwise somewhat arbitrary. EKMA is not recom-
mended for absolute results but provides a useful
means for relative comparisons between emission
reduction and oxidant scenarios.
Input is the aggregated concentration of NMHC.
One of the parameters in the chemical package is
adjusted according to the initial NMHC/NOX ratio.
This is a multilayered model.
Model has same generation as DIFKIN. It assumes
a well-mixed vertical column (single layer).
CH4 is excluded from alkanes. Model is multilayered.
CH4 is excluded from alkanes. Model is multilayered.
Both CH4 and C2H6 are excluded from the alkanes
group. C2H4 is assigned to aromatics. Model is
multilayered.
Eulerian models are highly complex, expensive to
operate, and require detailed input data preparation
for study of test cases.
SAI model, IMPACT and SULFA 3D are multi-
layered models.
CH4 is excluded from the alkane group. Aromatic
compounds are assigned to alkenes or alkanes depend-
ing on reactivity. LIRAQ-2 is a single layered model.
Model is used for exploratory modeling studies.
*EPA - Environmental Protection Agency
'''ERT — Environmental Research and Technology
*PES - Pacific Environmental Services
§ SAI — Systems Applications Inc.
"F and S — Forms and Substance Inc.
- Lawrence Livermore Laboratory
10
-------
SECTION 5
SITE SELECTION, SAMPLE COLLECTION,
STORAGE, AND INTEGRITY
Before an analysis is conducted, the task of collecting the air sample must be satisfac-
torily performed. This task is divided into the following four major parts:
• Site selection
• Presampling considerations
• Sample storage and integrity
• Recommended overall sampling procedures.
A. Site Selection
Siting is an important aspect of any HC measurement program especially in areas that
contain a large number of sources. These sources can be located very close and yet be so
small that they often go unnoticed. They may also be diffuse, low level area sources or large
point sources. This report does not provide detailed criteria for site selection (see Ludwig and
Shelar, 1978), but does briefly discuss those practical considerations that greatly facilitate the
process of site selection.
HC species ambient measurements for the purposes of air quality applications are best
utilized when the following two major considerations are met.
• The selected site should be well removed from nearby large HC point sources to
ensure a degree of homogeneity.
• The chosen site should be representative of a well defined geographical area.
Any site selection process should be preceded by a knowledge of the following factors:
• Identification of the region of specific interest
• Purpose to be served by the measurements
• Desirable features of a potential site.
11
-------
Once these decisions have been agreed to, a background information base should be assem-
bled. This data base typically should contain the following information, relevant to the
preselected region:
• Maps and pictures, including an aerial view
• Land use
• Emissions inventories
• Population densities
• Traffic distributions
• Climatological and meteorological data
• Existing monitoring data.
In practice, the task of selecting a desirable site is greatly facilitated by direct communication
with those who were responsible for selecting existing HC monitoring sites. Often, available
information from such sites is in itself a good indicator of the desirability of a given site. In general,
the selection of a site should not be finalized until strong nearby sources are identified and their
impact on the potential site assessed at least roughly. Regions of strong source gradients should
either be avoided or enough sampling sites installed to characterize these gradients.
As a general rule, a site should be as well removed as possible from the impact of local
sources. Five steps should be followed:
• Large point sources should be characterized and their impact on a potential site
assessed using simple Gaussian models. It may be preferable to avoid sites that can
be directly impacted by large point sources.
• A minimum separation from nearby .roadways should be maintained. Figure 2 can be
used to specify the separation (on a worst case basis) between monitoring sites and
roadways with different levels of average daily traffic (ADT).
. A site should be as far removed as possible (at least several hundred meters) from
such ground level sources as gasoline stations, drycleaners, surface coating opera-
tions, and refineries.
• The inlet of the sampling manifold should be 3 to 15 m above ground level (the higher
the better) and at least 1 m above any support surface. In addition, it should be
separated from surrounding obstacles by distances at least twice the height of the obs-
tacle.
• An open urban location should be found. Street canyons should be avoided.
A problem sometimes encountered by researchers is contamination from other air quality
instrumentation operating in the vicinity. For example, C2H4 from chemiluminescent 03 analyzers
has appeared in samples at concentrations of several hundred ppb's. Preferably, a sampling site
should be well removed from and upwind of any air quality instrumentation. In addition, all ozone
monitors should be fitted with C2H4 exhaust oxidizers.
In summary, selection of a proper site or sites should be made after a careful analysis of
available information and consultation with those experienced in selecting monitoring sites. A
12
-------
20.000
2000
O
Q
z
O
CO
tr
<
o
o
tz
o
200
20
ASSUMPTIONS:
1. PEAK-HOUR TRAFFIC EQUAL 10% OF ADT
2. SLIGHTLY STABLE ATMOSPHERE
3. INITIAL VERTICAL DISPERSION EQUAL TO 1.5m
4. EMISSION RATE EQUAL TO 4 g mile'1
10 20 40
SOURCE: Ludwig and Shtlar (1978)
100 200 400 1000
DISTANCE FROM HIGHWAY — m
2000
4000 10.000
Figure 2. Maximum roadway contribution to concentration
at different distances.
Knowledge of the performance of existing networks or air quality stations is a valuable guide in
site selection. Guidelines such as those published by Ludwig and Shelar (1978) should be fol-
lowed when specific applications are in mind.
B. Presampling Considerations
Once a site (or sites) has been selected, the task of sample collection begins. The opera-
tor must answer the following questions:
• What sample integration (averaging) times are required by the model?
• How much sample volume and how many samples are required?
13
-------
• What kind of a sampling platform is to be utilized for ground level and/or airborne sam-
pling?
Procedures recommended in this document for collecting air samples allow more than
adequate sample volume for HC species analysis, are as applicable at ground level as aboard
airborne platforms, and are easily adjusted to any sample averaging time. Integrated samples
with the following averaging times are often needed:
• 3-hour integrated sample (e.g., 6 to 9 am)
• 1-hour integrated sample
• Instantaneous (or less than 10 min) sample.
Proper choice of a site when coupled with samples integrated over a long time period (e.g.,
3 h) greatly increase the geographical representativeness of a sample. This is a desired goal in
most air quality applications.
C. Sample Storage and Integrity
The most desirable features of a sampling vessel are:
• Its ability to contain an air sample without impairing its integrity for as long as possible.
• Its ability to allow collection of a desired volume of air sample with a preselected
averaging time.
• Ease of handling and reliability of the vessel material.
The deterioration in integrity of an air sample because of interactions with the container
walls frequently occurs during storage. In addition, the desorption of chemicals from walls of
the container, permeation through walls (especially bags), as well as further reactions in the
gas phase, need to be considered. The problem becomes especially difficult when a complex
mixture is involved. Even though the total NMHC concentration is maintained, an internal
chemical differentiation may occur. Additional steps are often necessary to minimize these
problems. The two major sampling options available are:
• Collection of trace HCs on solid sorbent tubes
• Collection of whole air samples in bags and metal containers.
1. Collection of Trace HCs on Solid Sorbent Tubes
In principle, all trace constituents in the ambient air can be collected by passing air
through a bed of a suitable solid sorbent maintained at a desired temperature. Typical sorbent
materials used are charcoals, silica gel, alumina, and Tenax. The problems of collecting and
retaining all components of a complex mixture of trace chemicals have not yet been resolved
for any of these materials. Recently, Tenax (a polymeric material) has received the most atten-
tion because it can collect ambient species at room temperature. Low breakthrough volumes
for water and CO2 allow the preconcentration of HCs from extremely large volumes of air. How-
ever, many of the light HCs cannot be collected on Tenax. Overall, the advantages offered by
solid sorbent traps do not outweigh their disadvantages. The advantages and disadvantages of
cartridges in general and Tenax cartridges in particular are listed in Table 2.
14
-------
TABLE 2. ADVANTAGES AND DISADVANTAGES OF SELECTED COLLECTION MEDIUMS
FOR AMBIENT HC SPECIES ANALYSIS
Sampling Type
Advantages
Disadvantages
Solid sorbent tubes
(Tenax)
Plastic bags [e.g.,
Mylar, FEP Teflon
and Tedlar (PVC)
; films]
I
Passivated stain-
less steel canister
• Adequate breakthrough volumes at
ambient temperatures for some species
• No significant retention of CO2 and H2O
• Allows collection of very large sample
volumes (<104 liters) for some species
• Integrated samples over a period of
minutes to days easily obtained
• Thermal desorption of chemicals by heat-
ing at 400°C for several minutes easily
accomplished
• Sampling cartridges are light weight, low
volume, rugged, and easily handled and
transported
• Easy availability
• Convenient for collecting integrated
samples over any predefined time
interval
• Good stability of chemicals
• Pressurization allows the collection of a
desired volume of air « 20 liters) in a
compact mode.
• No chance of leakage or permeation of
chemicals through metal walls. Sunlight
is kept out.
• Trace chemicals show excellent stability
(see Table 4-6)
• Easy to obtain, clean, transport, and
handle
• Breakthrough volumes differ greatly and
many species are only partially retained
(see Table 3)
• Serious artifact problems are encountered.
Ozone and oxygen appear to oxidize Tenax
resulting in significant amounts of artifact
contamination
• In case of HC species analysis the major
advantage of Tenax cartridges (collection
of a very large volume) is not a necessary
feature
• Reduced precision due to less than quanti-
tative absorption and desorption of
chemicals
• Poor handling properties, such as pinholes
and cracks encountered during handling
and transportation
• Permeation of chemicals out of and into
plastic film bags during storage
• Self-contamination from compounds
released due to outgassing from bag films
• Limited total volume can be collected
«20 liters STP)
• Higher initial cost
• Integrated samples not easily collected
15
-------
Table 3 clearly shows that for many C6-C10 HCs Tenax cartridges do not yield accurate
quantification, which results in poor reliabilty. Results for CrCs HCs are even poorer. To date,
the problem of artifacts remains poorly defined and highly unpredictable. Thus, cartridges filled
wih solid sorbents (such as Tenax), at their present state of development, do not offer a desir-
able sampling means for the accurate quantification of hydrocabon species data.
TABLE 3. COMPARISON OF HC SPECIES ANALYSIS MEASURED DIRECTLY
AND AFTER COLLECTION ON TENAX CARTRIDGES
Chemical Species
2-Methylpentane
3-Methylpentane
n-Hexane
Methylcyclopentane
2,4-Dimethylpentane
Benzene
Cyclohexane
2-Methylhexane and
2,3-Dimethylpentane
3-Methylhexane
2,2,4-Trimethylpentane
n- Heptane
Methyl Cyclohexane
Toluene
2-Methylpentane
3-ethylhexane
n-Octane
Ethylbenzene
m/p-Xylene
o-Xylene
n-Nonane
n-Propylbenzene
p-Ethyltoluene
m-Ethyltoluene
1 ,3,5-Trimethylbenzene
o-Ethyltoluene
1 ,2 ,4-Tri methyl benzene
1 ,2,3-Trimethylbenzene
1 ,3-Diethyibenzene
1 ,4-Diethylbenzene
C-1 0 Substituted benzene
Tenax*
(Aig/m3)
12.8
8.4
11.7
3.0
5.1
38.6
5.2
13.2
6.4
7.7
7.7
6.2
52.0
4.3
7.8
6.5
13.7
54.1
20.9
7.2
5.4
21.1
9.5
12.0
15.5
42.8
8.1
9.8
8.9
8.6
Direct
(Aig/m3)
23.7
16.5
20.4
13.5
7.6
66.1
6.2
24.5
12.6
13.5
13.6
10.3
86.7
6.3
8.6
6.6
13.5
51.5
19.3
6.6
5.3
19.5
9.2
12.5
11.9
42.0
7.8
9.2
8.6
8.6
Ratio of
Tenax/Direct
.54
.51
.58
.59
.67
.58
.83
.54
.51
.57
.57
.60
.60
.68
.91
.99
1.01
1.05
1.08
1.09
1.02
1.08
1.04
.96
1.30
1.02
1.03
1.06
1.03
1.00
500 ml sample air per gram of Tenax.
Source: Holdren et al. (1979)
16
-------
2. Collection of Whole Air Sampling
in Bags and Metal Containers
Collection of whole air samples is desirable when the volume requirements are less than
100 liters and preferably less than 25 liters. For the analysis of HC species, a 10-liter sample
is considered to be more than adequate; thus, whole air sampling is an acceptable option. Two
types of devices are commonly used for this purpose: plastic bags and metal containers.
For many years plastic bags have been used as storage and reaction containers for
atmospheric photochemical studies. Although a wide variety of bag materials has been tried,
the most common materials have been such commercial polymer films (2 to 5 mil) as Mylar,
FEP Teflon, and Tedlar (polyvinyl fluoride material). The most desirable feature of the bags is
their flexibility, which allows easy collection of an integrated sample. The major advantages
and disadvantages of using these bags are summarized in Table 2.
The collective experience of all research groups supports the view that if the bags are
handled carefully, they can provide a reliable means of collecting an air sample for HC species
analysis. However, problems of outgassing and permeation into and out of the bags can create
serious problems (see Seila et a!., 1976). All the problems associated with bag sampling are
greatly minimized by the following precautions:
• Checking to assure against a "bad batch" of bag material by flushing it with HC free air
and analyzing the stored air.
• Thorough cleaning of bags.
• Storage of bags (before and after sampling) in dark areas and at relatively cool tem-
peratures (<70"F).
• Sample storage times of less than one day.
• Careful and minimal handling of film bags.
Long-term storage in the bags or transport in bags of the sample from the field to central
laboratories is at best an undesirable practice. Although some have claimed that bags yield
reliable answers over a 2-week period, this delay is not recommended. Certainly, fluorocarbon
tracers and several HCs are expected to permeate (both ways) through bag walls. For storage
times of several hours, both outgassing and permeation problems are minimal. The flexibility
with which bags can be easily used to collect whole air integrated samples is their major asset.
Because of the shortcomings of bags, solid wall containers have become increasingly
popular in recent years. Glass, spectra-sealed aluminum, and stainless steel sampling canis-
ters are available. Currently, the reliability of anodized aluminum cylinders for purposes of HC
species sampling is uncertain and glass samplers can break. The most desirable sampling
devices are passivated stainless steel canisters. Electropolished by the SUMMA process
(patented), these lightweight containers can be safely pressurized to 50 psi. A 2-liter canister
weighs less than 1.5 kg. Figure 3 shows a 1-liter passivated sampling canister. These are
available in 0.5 to 6 liter sizes in cylindrical or spherical shapes. A 2-liter size is considered
optimum for detailed HC species analysis.
The advantages and the disadvantages of these metal containers are listed in Table 2.
The major disadvantage of these canisters is the inability to collect an integrated sample in a
17
-------
Figure 3. Specially treated 1-liter stainless-steel sampling vessel
for collecting grab samples
18
-------
reproducible fashion. * For collecting instantaneous grab samples (i.e.,sampling time less than
10 minutes), the metal container can be directly pressurized. It is clear from Table 2 that the
major disadvantage of metal containers—the inability to collect an integrated sample—is pre-
cisely the advantage that the bags offer.
Polished stainless steel containers (PSSC) have been extensively tested with satisfactory
results by virtually all research groups. Tables 4 and 5 show the stability of C2 to C5 and C5 to
C10 HCs, respectively, in these sampling vessels for periods of 1 to 3 weeks. In addition, the
stability of selected natural HCs in these canisters is evident from Table 5. Similar resuls in
Tedlar bags have also been reported by Lonneman and Bufalini (I979). Table 6 further shows
that oxygenated species are also stable in these canisters. Other tracers, such as
fluorocarbons-1 2 and 11, have maintained their integrity in these canisters for several years
(Singh et al.. 1979).
D. Recommended Overall Sampling Procedure
Clearly, metal containers provide all the desirable features of a sampling vessel, except for
the ability to directly and easily collect a linearly integrated sample of air. This, however, is the
best feature of bags. Therefore, a combination bag/metal container approach for sample col-
lection is recommended. This approach draws on the best features of both sampling options.
Any sampling procedure must be tested to ensure that no contamination of air occurs at
any step. Bags and polished stainless steel containers must be specially and thoroughly
cleaned before any sample collection begins. The PSSC are best cleaned by heating them at
100°C and purging them with cryogenically cleaned (at liquid 02 temperature) ultrapure air at
a flow rate of 0.5 liter per minute for a period of approximately 30 minutes. Purging with an
inert gas such as helium or nitrogen is not quite as effective. The effectiveness of air is partly
due to the oxidative properties of oxygen. In actual practice, two or three of these canisters
can be connected in series to save time and gas. After this cleaning process, the PSSC should
be stored under a positive pressure. Other cleaning processes, where heated canisters are
filled with clean air and then evacuated, with the procedure repeated several times, are also
acceptable. Depending upon the number of samplers, ten or more PSSC can be cleaned simul-
taneously.
In the sampling procedure very few bags are required; two or three bags per site are ade-
quate. Whenever possible, 75 to 100 liter bags (2 mil Tedlar or 5 mil Teflon) should be used.
These bags must be individually and thoroughly cleaned. The bags should be filled with ultra-
clean air and evacuated. This procedure should be repeated several times until the bags are
clean.
In summary, the following overall steps in sampling are involved:
• Select a sample integration (averaging) period.
• Procure a 75 to 100 liter bag that has been previously cleaned and protect it from sun-
light
'Attempts have been made to directly collect integrated samples in metal containers (Harden, 1979). An evacuated
vessel is slowly filled by using a critical orifice that controls flow. This results in a non-linearly integrated sample since
the flow rate varies as the backpressure increases. Therefore, this technique cannot be recommended for quantitative
and accurate HC species analysis.
19
-------
TABLE 4. STABILITY OF SELECTED LIGHT HCs (C2-C5) IN PASSIVATED STAINLESS STEEL CONTAINERS
Sample Number
Date of Collection
Date of Analysis
Ethane
Ethylene
Propane
Acetylene
Isobutane
n-Butane
Propylene
Isopentane
Pentane
Sum
Mean
Standard Deviation
A-84
7/6/78 (0730-0910 EST)
7/7 7/7 7/18 7/25
12.8 13.2 13.6 13.0
104.8 103.5 98.2 100.5
6.2 6.6 7.1 6.6
138.1 136.9 126.6 119.4
7.0 9.1 6.3 7.6
38.6 38.8 34.6 37.2
34.3 31.3 28.6 28.6
65.1 59.3 58.5 56.3
44.3 42.2 40.5 37.6
451.2 440.9 414.1 406.8
428.2
21.2(*5%)
A 176
7/6/78
7/6 7/7 7/18 7/25
13.3 12.4 12.4 11.9
102.5 99.0 94.6 97.9
6.0 5.8 5.0 6.3
149.0 130.2 124.5 120.3
7.8 6.2 7.4 8.7
38.7 35.8 35.1 36.0
32.3 29.5 27.0 27.7
62.3 60.0 55.4 55.8
38.5 38.1 38.3 45.1
450.5 417.0 399.7 409.7
419.2
22.0 (*<5%)
A-190
7/6/78
7/6 7/6 7/18 7/25
12.7 13.8 12.7 12.5
108.9 106.6 100.4 103.1
6.0 6.6 6.4 6.0
149.5 142.2 131.2 127.8
10.4 6.4 7.5 7.6
40.4 38.0 38.8 36.9
35.8 34.1 31.3 31.2
61.0 61.2 55.1 59.5
42.4 32.7 39.0 38.4
467.1 441.6 422.4 423.0
438.5
21.0 («5%)
Mean
12.9
101.7
6.2
133.0
7.7
37.4
31.0
59.1
39.8
428.7
428.6
Standard
Deviation
0.5 (P*4%)
4.0 («4%)
0.5 (=*8%)
10.2 (*8%)
1.2 (« 15$)
1.8(^5%)
2.8 (**9%)
3.1 («5%)
3.4 (**&£)
21.0(«5%)
9.7 (*2%)'
to
o
Source: Harden (1979)
-------
TABLE 5. STABILITY OF SELECTED HEAVY HYDROCARBONS (C5-C10)
IN PASSIVATED STAINLESS STEEL CONTAINERS
(Mg/m3)
Compound
2-Methylpentane
3-Methylpentane
Hexane
Methylcyclopentane
2,4-Dimethylpentane
Benzene
Cyclohexane
2-Methylhexane
3-Methylhexane
Dimethylcyclopentane
2,2,4-Trimethylpentane
Methylcyclohexane
Toluene
2-Methylheptane
3-Methylheptane
Octane
Ethylbenzene
m, p-Xylene
o-Xylene
4-Ethyltoluene
2-Ethyltoluene
1 ,2,4-Trimethylbenzene
1 ,2,3-Trimethylbenzene
Isoprene*
a-Pinene*
0-Pinene*
Limonene*
Initial
Concentration
6.0
3.7
5.6
3.3
2.0
6.7
1.3
7.5
3.0
5.9
2.7
2.0
14.8
1.9
1.3
1.0
2.5
10.1
3.4
2.5
2.2
3.9
3.9
26.4
13.1
9.4
8.4
Concen-
tration
After
1 Week
5.9
3.8
5.2
3.2
2.0
6.2
1.3
7.0
2.8
5.8
2.9
1.8
14.3
1.8
1.3
1.0
2.5
10.1
3.3
2.5
2.0
4.0
3.3
25.4
11.9
7.9
6.1
Percentage
Change
-2%
+3
-7
-3
0
-7
0
-7
-7
-2
+7
-10
-3
-5
0
0
0
0
-3
0
-9
+3
-15
-4
-9
-16
-27
Concen-
tration
After
2 Weeks
5.7
3.9
4.9
3.1
1.9
5.9
1.2
6.9
2.6
5.6
2.5
1.7
13.7
1.8
1.3
1.0
2.4
9.7
3.4
2.4
1.8
3.9
2.8
23.9
9.6
5.6
1.0
Percentage
Change
-5%
+5
-13
-6
-5
-12
-8
-8
-13
-5
-7
-15
-7
-5
0
0
-4
-4
0
-4
-18
0
-28
-10
-27
-40
-88
Concen-
tration
After
3 Weeks
5.4
3.6
4.7
3.0
1.8
5.4
1.1
6.4
2.5
5.1
2.1
1.7
11.9
1.6
1.2
.9
2.1
8.4
3.0
2.6
1.6
2.9
2.4
21.9
5.3
2.5
<1.0
Percentage
Change
-10%
-3
-16
-9
-10
-19
-15
-15
-17
-14
-22
-15
-20
-16
-8
-10
-16
-17
-12
+4
-27
-26
-38
-17
-60
-73
-100
Source: Holdren et al. (1979)
'Canister artificially spiked with these compounds.
21
-------
TABLE 6. STABILITY OF SELECTED OXYGENATED SPECIES IN
PASSIVATED STAINLESS STEEL CONTAINERS
Compound*
Acetone
Neohexane*
Butyraldehyde
2-Butanone
Crotonaldehyde
Valeraldehyde
2-Pentanone
Furfural
2-Heptanone
Heptaldehyde
Benzaldehyde
2-Octanone
Percentage Change After 3 Weeks Storage
Can 1
+10
0
- 7
+ 9
+ 9
- 2
+10
- 2
+15
+ 6
+ 6
+10
Can 2
+13
0
-17
-14
- 6
-31
- 1
-22
-13
-50
-14
-46
Can 3
+12
0
-27
-19
- 9
-35
+ 2
-49
- 7
-76
-13
+24
Source: Holdren et al. (1979).
'individual concentrations varied between 20 and
* Internal standard.
• Fill this bag at a constant flow rate so that a 70-liter air volume is collected over the
selected averaging period.
• Transfer this bag air sample into two previously evacuated 2-liter PSSC at atmospheric
pressure and evacuate these PSSC. Repeat this procedure twice (to ensure that any
remaining air in the container has the same composition as the bag air).
• Fill the two containers with bag air at 35 to 40 psi pressure.
For each integrated sample collected in a bag, two separate canisters should be pressur-
ized. While one 2-liter canister at about 40 psi is adequate for complete HC species analysis,
quality assurance is greatly facilitated by collecting two canisters. Sample containers prior to
sampling should be maintained at a positive pressure. The evacuation should be performed at
the field site.
It is recommended that tubing employed during sampling should be precleaned stainless
steel or Teflon. All fittings should be stainless steel, brass, or glass. Pumps for pressurizing
and collection have been known to be key causes of contamination. Metal bellows (MB), Car-
bon Vane, and Komhyr "all Teflon" pumps are acceptable (Westberg, 1979). Metal bellows
22
-------
pumps, in particular, the MB-158, a stainless steel pump that has all the features that a typical
sampling program may require, is recommended. These metal bellow pumps can be used to
pressurize sampling containers to 50 psi (typically 40-50 psi). Both battery operated and AC
pumps are available. When an AC pump is being used, a long cord (200 m) must be carried for
contingencies. The same pump (MB-158) can be used for evacuation of canisters. The canis-
ters should be evacuated to as low a pressure as possible. Whenever a manifold is used for
drawing inlet air, glass or stainless steel manifolds are desirable.
Self-contamination from pumps, tubing, valves, and containers is an important problem that
can also vary widely depending upon the source as well as the batch of materials. No shortcut to a
systematic evaluation of potential problem areas exists. It is recommended that ultrapure air be sampled
through the entire sampling system and be analyzed before and after collection to ensure the
cleanliness of the many parts that routinely constitute a sampling system. The same procedure should
be used with a complex HC mixture (containing alkanes, alkenes, aromatics,etc.)to ensure that there are
no surface removal problems.
Figure 4 provides a schematic diagram of an integrated bag sample collection and its
transfer to the sampling container. A controller can be used to shut off the pump at the end of
the selected sampling period. This could also be accomplished manually. Once the two sam-
ples have been collected, they should be completely labeled (sampler numbers, date, start
time-end time, location, etc.) and stored for analysis.
It has been suggested often that sampling of alkenes, which are known to react with 0^ is
best performed when the O3 is quickly destroyed. Injection of NO has been used to effect this
destruction. As long as the bag is kept in a dark place, O3 generation from NO2 photolysis
(also generated from NO injection) can be eliminated. Practical experience indicates that this
approach is unnecessary. For short integrating periods (<3 h), there is insufficient time for
reaction during collection. Also, most pumps, including the one recommended here, destroy
more than half of the 03 during sampling; the bag surface destroys some more (Lonneman and
Bufalini, 1979). A second transfer into a canister ensures almost complete destruction of 63.
Thus, for the sampling procedure recommended here, the Injection of NO to destroy 03 is con-
sidered unnecessary and could be omitted.
23
-------
2-MIL TEDLAR
OR
5-MIL TEFLON
PRESSURE GAUGE-
CRITICAL ORIFICE,
FLOWRATE = 1000ml/min (1h SAMPLE)
= 350ml/min (3h SAMPLE)
(a) INTEGRATED BAG SAMPLE
NUPRO 4H4 VALVES
PRESSURE GAUGE
1 TO 5 LITER POLISHED SS CONTAINERS
(PREFERRED SIZE =2 LITERS)
(MAXIMUM PRESSURE 40-50 PSI)
METAL BELLOWS
(MB-158 PUMP)
CONTROLLER 1
MANIFOLD-
M8-158 PUMP
(b) INSTANTANEOUS TRANSFER FROM BAG TO CONTAINER
NOTE: AM tubing should be stainless steel (SS) or Teflon: connectors and valves should be SS, brass, or glass.
Figure 4. Collection of integrated bag sample
and transfer to metal container.
24
-------
SECTION 6
ANALYSIS OF HC SPECIES
The analysis of HC species is often the most critical step, requiring sophisticated instru-
mentation and experienced personnel. A set of procedures that can be depended upon to pro-
vide detailed HC composition data for air quality management are described here. These pro-
cedures are kept as simple as possible while still reliably determining the composition of a
highly complex mixture of ambient air. A detectability limit of 1 ppb C for individual species is
assumed to be adequate.
A. Mode of Operation
As stated earlier, once samples have been collected, the delay between collection and
analysis should be kept to an absolute minimum. This delay should be kept to an absolute
minimum, not to exceed 1 week. Two modes of operation are generally feasible:
• An instrumented mobile laboratory is stationed in the vicinity of the sampling site.
• Samples are transported, sometimes over long distances, for analysis at central labora-
tories.
The first method is clearly the most desirable. Depending upon the number of samples,
analysis time can be kept down to less than a day or two. Also, any problems or local features
indicating hitherto unknown compounds are identified quickly and corrective action can be
taken. The major disadvantage of stationing an instrumented field laboratory is usually the
expense involved. However, this mode of operation should always be considered and used
when feasible.
Because the practical needs of air quality monitoring far exceed the availability of instru-
mented mobile laboratories, the samples must often be analyzed at specialized laboratories.
The major delay is usually transportation. However, through proper planning these delays can
be minimized and kept to less than 1 week.
B. Instrument Optimization, Sample Preconcentration
and Injection
While many sophisticated tools such as GC-MS have been used for the identification of
ambient HCs during the developmental phases of measurement technology, the GC-FID tech-
niques are considered satisfactory and the most practical means for HC species data collec-
tion. The flame ionization detector (FID) is a well-established and highly stable detector that
possesses a remarkable linear dynamic range of 106. In addition, a unique advantage of the
FID is its near linear response to the number of carbon atoms in organic compounds.
Thus, calibrations do not have to be conducted for each and every species being
25
-------
measured, and unknown peaks can be quantified. The GC-FID provides the best and most
economical commercially available means for the .analysis of HC species.
Instruction manuals for GC equipped with FID detectors usually describe the conditions
for optimum analysis. These conditions, however, vary with the design of the FID. The .operator
is advised to consult the GC-FID manual for selection of optimum parameters. Table 7 shows
optimum operating conditions (fairly typical of most FIDs) for the Hewlett-Packard Model 5711
A FID detector.
TABLE 7. OPTIMUM OPERATING CONDITIONS FOR
MODEL 5711A (Hewlett-Packard) FID
Total Carrier Flow*
Hydrogen Flow (20 psi)
Carrier/Hydrogen Ratio
Air Flow (20 psi)
Nitrogen
Carrier
51 ml/min
39 ml/min
1.31
240 ml/min
Helium
Carrier
56 ml/min
34 ml/min
1.62
240 ml/min
'Auxiliary carrier gas must be added to the detector when the column
flow is below this value.
For a sensitivity approaching 1 ppb C, a 500-ml air sample is considered optimum. This
will allow measurement of Cj-C12 HCs without causing problems due to clogging from water
and CO 3. To inject a 500-ml sample of air, the HCs must be trapped on a sorbent material,
which will allow the quantitative collection of all trace species without contamination and will
desorb them quantitatively and easily. A 4-inch long bed of 80/100 mesh glass beads packed
in an 1/8 in. o.d. stainless steel tubing serves as an excellent preconcentration trap for all HCs
of interest. Although a 0.5-liter sample size is recommended, doubling this is not likely to
create any problems. A 0.5-liter sample under extremely humid conditions (relative humidity >
90 percent) is easily retained on such a preconcentration trap.
Sample preconcentration is accomplished at liquid oxygen (B.P.-183*C) or liquid argon
(B.P.-186°C) temperatures. Liquid argon is safer to use and is to be preferred. Liquid oxygen is
more easily available and can also be prepared at the site with liquid nitrogen. For field opera-
tions, liquid nitrogen should be carried and liquid oxygen prepared as needed on site.
Desorption from glass beads for all HCs of interest is accomplished by heating the
freezeout trap with boiling water for 3 minutes or by electrical heating at 90*C. The sample is
transferred into the GC-FID unit by the carrier gas itself. Figure 5 shows the sampling and
injection modes of a preconcentration trap.
26
-------
COMPOUND GAUGE PRESSURE INDICATOR
(WALLACE AND TIERNAN)
MODEL 62A-6B-0030
-380 TO 0 TO +380 mm Hg PRESSURE RANGE
ACCURACY = ±0.4 mm Hg
0-50 psig
2-LITER PASSIVATED
STAINLESS STEEL
CANISTER
SAMPLE FREEZEOUT LOOP
(a) FILL POSITION
SAMPLE CANISTER
•#•
SAMPLE FREEZEOUT LOOP
(bl INJECTION POSITION
Figure 5. A schematic diagram of sample preconcentration
and injection steps.
27
-------
The simplest procedure for measuring the volume of air sampled is by installing a 2-liter
stainless steel canister at the freezeout trap exit. This canister can be the same as the sam-
pling canisters but its volume must be accurately determined. The actual volume sampled can
be calculated by carefully monitoring the pressure change in this canister. Figure 5 shows the
specifications of one recommended vacuum and compound range pressure gauge. This 8-1/2
inch diameter precision gauge (Wallace and Tiernan. Inc., Belleville, New Jersey) is available in
the -380 to +380 mm Hg pressure range and has an accuracy of ±0.4 mm Hg and a sensitivity
of ±0.06 mm Hg. Typically, when the sampling canisters are at 30 to 40 psig pressure, the
vacuum part of the gauge is not necessary- In a 2-liter vessel a pressure change of 190 mm
Hg (at a 760 mm Hg total pressure) would correspond to a 500 ml sample volume. When the
sample canister is poorly pressurized (5 to 10 psig) it may be desirable to operate the volume
monitoring canister at a slight vacuum. This would ensure smooth flow through the freezeout
trap.
While an accurate volume of the injection air sample is essential, the volume need not be
the same all the time. An overall sampling accuracy of better than ± 3 percent is attainable.
Note, however, that neglect of water vapor in air can cause an error of ±0 to 2 percent. For
determining air quality, this is not significant.
To perform an actual transfer operation refer to Figure 5. Valve A is closed and valves B
and C opened to ensure a leak-proof system as indicated by pressure gauge PG1. If so, valves
A and D can now be opened. The freezeout loop should be immersed in hot water and purged
with the sample to flush out the system for 20 seconds. Valves B and D should now be closed
and the pressure on gauge PG2 read carefully. The hot water should now be replaced with
liquid oxygen (or argon) and valve B quickly opened. The sampling valve should be changed to
the inject position and liquid oxygen quickly replaced with hot water for 3 minutes. At the end
of the sampling period final pressure on PG2 should be noted and valve B closed. It should be
added that the final pressure on PG2 can be predetermined once a sample size has been esta-
blished. The actual sample volume preconcentrated can be calculated as follows:
V, • 298
volume sampled - — . . APG2
Ps • u T 27o/
where AV is the volume sampled at standard atmospheric pressure Ps (760 mm Hg) and tem-
perature (25*C), APG2 is the pressure change on gauge PG2, V, is the vessel volume, and t is
the sample temperature.
C. Chromatographic Separations and Analyses
of HC Species Data
In the previous section the procedures for collection and preconcentration of a known
volume of air on a freezeout trap were defined. The trace chemicals are thermally desorbed
and directly injected into the GC. Details of column packings and other procedures that must
be followed to obtain HC species data shall be provided here.
Dividing HC species into two major groups, light hydrocarbons (Cj-Ce) and heavy hydro-
carbons (Cg-C12 HCs), is feasible. Some duplication occurs in the CrC6 analysis and Ce-C12
28
-------
analysis. This redundancy is a good internal check on the reliability of the two methods.
Methane (CHJ plays an unimportant role in photochemical 03 generation in polluted atmo-
sphere and is not considered to be a hydrocarbon of interest here. Methane is not collected on
the freezeout traps under the preconcentrating conditions recommended here.
1. Light Hydrocarbon Species (CrC6 HCs)
Light HC species are easily measured by injecting a 500-ml sample of air into a GC-FID
micropacked column system. Table 8 summarizes the GC conditions that are recommended for
CrCQ HC species analysis. Column No. 1 is recommended primarily because it allows isother-
mal operation and greatly simplifies the analysis of light HCs. Column 1, however, cannot be
heated above 60*C and all column preconditioning should be done at 50*C. Column No. 2 is an
equally reliable column but temperature programming is absolutely essential. Subambient tem-
perature programming is used on Column 2 and hence this operation is more tedious. Figures
6 and 7 shows chromatograms of CrC6 HC species analysis with the recommended Column 1.
No two columns are exactly identical. For example, Figure 6 and 7, although obtained from
nearly identical GC columns, do not produce identical chromatograms. Each column should be
tested and optimized for best possible operation at each laboratory. The excellent performance
of Column 2 with subambient temperature programming is also demonstrated in Figure 8. The
simplicity of the operation of Column 1 makes it highly preferable. Generally, Column 1 should
not be used after hexane has been resolved. As will become clear, 2-methylpentane, hexane
etc. are also resolved during heavy hydrocarbon species analysis. This overlap provides a
good internal check on the reliability of two columns.
2. Heavy Hydrocarbons (Cg-C 12 HCs)
Because of the complexity of the Ce-C12 HC mix in ambient air, it is essential to use capil-
lary columns. Three major advantages compared to non-capillary columns are realized:
• Analysis of all Cg-C12 HCs is accomplished in a single operation.
• Higher resolution for the same analysis time or equal or better resolution in much less
time is obtained.
• Higher sensitivity.
In the past, for example, CS-C10 aliphatics and aromatics required at least two GC
columns for effective separation (Lonneman et al., 1974). The complexity of Cs-C12 HCs is
ideally resolved and quantified with the help of high resolution capillary columns. However,
capillary columns suffer from a lower operational life and require careful handling when
compared to packed columns. The advantages of capillary columns far outweigh their disad-
vantages. For the analysis of higher than C6 HCs, Columns 3 and 4 (Table 8) are recom-
mended. Although the dimensions are different, Columns 3 and 4 are both SE-30 glass capil-
lary columns and therefore very similar. Both columns can resolve major CrC12 HCs satisfac-
torily, although Column 3 operation is expected to give superior resolution. While temperature
programming is essential for both Columns 3 and 4 operations, the former operation requires
subambient temperature programming (-30*C to 80*C). As discussed in subsequent para-
graphs, Column 3 operation is superior and is recommended when possible. Column 4 opera-
tion is clearly simpler (above ambient temperature programming) and also acceptable.
29
-------
TABLE 8. GC COLUMN SELECTION AND OPERATING CONDITIONS FOR DETAILED HC SPECIES ANALYSIS
Category
Light
Hydrocarbons
(C2 Ca HCsl
Heavy
Hydrocarbons
(Ce C,j HCll
Tracer
Column
No.
1
2
3
4
4
Column Description
5m x 1/16-in.i.d.. Ni, Dura-
pack Phenyl-isocyanate/
PorasilC. 100/120 mesh
(micropacked)
6m x 1/16 in. i.d.. SS,
Qurapack n-octane/Porasil C.
100/1 20 mesh
30m SE-30 glass capillary.
0.25 mm i.d.. W.C.O.T.
50m SE-30 gla» cipillwy.
0.5 mm i.d.. W.C.O.T.
2m « I/Bin. 10%DC-200on
80/100 mesh-carbowax
W IA/WI plus 3 in. x I/Bin.
Ascarite trap
Operation
Isothermal
operation at
40°C
Programmed tem-
perature opera-
lion necessary
-70°C lo 65°C
at 16°C/min-
hold at 65° C for
25 min.
-30°C for 8 min.
-30°C to 80°C
at 4°C/min, 15
min at 80° C
45°C lor B min.
45 to ISO°Cat
8°C/min, 10 min
at 150°C
Isothermal
operation at
45"C
Total Time
per Sample
(minutes)
40
40
60
60
20
Typical
Sample Size
(ml)
500'
500*
300-500'
300-500*
5
Detector
Type
FID'
FID
FID
FID
ECO*
Temp.
I°CI
250
250
250
250
250
Typical
Carrier Flow
Rate (ml/min)
30
5-10
1-2
1-2
40
Remarks
Recommended column. Requires no temperature programming.
Maximum column temperature - 60° C. Column conditioning
to be done at 50° C. Low bleed. Good operation at all relative
humidities.
Without temperature programming either C2He/C2H4 or
CjHj/CjHg cannot be resolved. Maximum column tempera-
ture is 176 C. Column conditioning to be done at 150°C.
Low bleed. Good operation at all relative humidities.
Column 3 when used under prescribed conditions is recommended
for high quality data collection. Subambienl temperature pro-
gramming is essential. Good operation occurs at all relative
humidities. SE 30 is to be preferred over 0V 101. which has
been traditionally used and suffers from bleed and artifact
problems.
Column 4 is very similar to Column 3. Because of programming
at above ambient temperatures the peak shapes may be interior.
Column 4 usage is acceptable if subambient operation is to be
avoided.
A single tracer such as CCljFIFIllis acceptable. Fll is measured
in 5 min, but other chlorinated species are expected. Ascarite
trap is used to remove water. Measurement of a tracer during HC
species analysis is optional but usually quite valuable.
W
o
'Collected on a 4 x 1/Bin. 80/100 mesh glass head trap at liquid O2 (or liquid Ar) temperature
"FID'Flame formation Detector; ECD-Eluctron Capture Detector.
-------
COLUMN No.1
TEMPERATURE
500-ml SAMPLE
25° C
SOURCE: SPICER (1979)
Figure 6. Light HC (C2-C6) analysis in rural air.
31
-------
ETHYLENE-
ETHANE
ACETYLENE-
PROPENE-
n-BUTANE
ISOBUTANE'
PROPANE -
INJECTION
N_
TRACE OF
METHANE
\J
COLUMN NO. 1
TEMPERATURE 45°C
-2-METHYLBUTANE
n-PENTANE + t-2-BUTENE + ISOBUTENE
PROBABLY t-2-BUTENE
USUALLY NOT SEPARATED
2-METHYLPENTANE
n-HEXANE
t-2-PENTENE
ISOPRENE
2.3 DIMETHYLBUTANE
2-METHYL-2-8UTENE
1-8UTENE 2,2 OIMETHYLBUTANE
SOURCE. HARDEN (1979)
Figure 7. C2-C6 HC species analysis of an auto exhaust air sample
32
-------
60J-
20
t- a.
w O
0
-7O
6
+ 26
10
+65 —»HOLO
12 14 16 16 20 22
MINUTES
TEMPERATURE
2«
26
JO
32
34
(a) COLUMN 2: STANDARD GAS MIXTURE 70 TO 65°C, 16°/min
ui
Ul
z
Ul
-------
Subambient GC operations add an additional complexity that in itself is neither difficult
nor problematic, however. In reality, once initiated subambient GC operation is quite routine
and at the very least allows a greater range of temperature programming.
The carrier flow rate through a capillary column is only 1 to 2 ml/min, and this is used to
flush the trapped sample into the GC. In Column 3 operation, because of the subambient
column temperature (-30°C), the trace HCs are trapped at the head of the column and then
released during temperature programming of the GC oven. This results in sharper peaks and
better resolution. Chromatograms in Figures 9 (also Table 9) and 10 (also Table 10) show the
resolution of Ce-C10 HCs on capillary Columns 3 and 4, respectively. A greater number of
peaks are shown to be well resolved in Figure 9. Most of the important peaks are also rela-
tively well resolved in Figure 10 where subambient temperature is not used. Considering that
10 to 35 percent of the NMHC burden is currently unidentified; both Column 3 and Column 4
operations are considered acceptable. In the future, as more of the similar peaks are identified,
Column 3 operation would be clearly preferable. Note that in both Figures 9 and 10 the GCs
show excellent baseline stability.
In both cases, however, only HCs up to C10are identified. (Note that not all HCs up to C10
have been identified.) This HC fraction constitutes a major part of the ambient HC burden.
However, Columns 3 and 4 would also measure HCs greater than C10, but these are not rou-
tinely found to be present in large amounts. Similarly, natural terpenoid HCs are also resolved
on Columns 3 and 4 (Table 8) (at the very end of the Figure 9 and 10 Chromatograms), but are
not generally found In significant quantities. HCs above C10 would be resolved and measured
on Columns 3 and 4, but would be part of the "unknowns." Typically, however, HCs greater than
C10 would form a very small fraction of the NMHC as well as of the overall "unknown" fraction.
As demonstrated in Figure 11, at early stages of HC resolutions on Columns 3 and 4, the
Chromatograms of ambient air mixtures are highly complex. Often it is desirable to measure
some species (such as 2-methylpentane or hexane) on both columns for internal checks. Thus
CrC6 HCs up to (or including) 2-methylpentane are measured on Column 1 and C6-C12 HCs
after (or including) 2-methylpentane on Column 3 or 4.
3. NMHC Measurements
For many applications (such as the EKMA model), the only input required is total NMHCs
(6 to 9 am) and therefore separation and identification of individual species need not be
stressed. The only reason for conducting detailed species analysis in such instances arises
from the unreliability of continuous NMHC measuring instruments at concentrations below 0.5-
1.0 ppm C. In that case, Columns 1 and 4 are recommended. A 2-methylpentane cutoff can be
used in both columns. Thus, all peaks to 2-methylpentane can be added up on Columns 1 and
the remainder of Column 4. The individual Column 1 and 4 responses can then be converted
to the concentrations (based on carbon response) and added to obtain total NMHC concentra-
tion (e.g., ppb C).
4. Aldehydes and Other Oxygenated Species
While oxygenated species are expected to constitute only a small fraction (<5 percent) of the
total NMHC burden, these species (especially aldehydes) can make a disproportionately large
contribution to photochemical smog formation. Unfortunately, no reliable instrumental measurement
method is currently available.
34
-------
u
en
UJ
V)
z
2
t/i
UJ
cc
r
L
1
1
5
2
4
H u*
10
9
11
12
16
1314
15
18
r
2<
2
19
LVL •
23
24
26
l^-
34
33
29
43'
.
v«
38
49
43
44
•
39
^_
40
J
.
45
46
r
^
49A
COLUMN 3
STANDARD GAS MIXTURE
TEMPERATURE = -30 TO +80°C
60
82 55 M 61 66
65
1KB I67A
1 66 A
64 I
1 70 |73
1 1 1 L lliL69^°A A
INCREASING TIME AND TEMPERATURE
NOTE: For pMk Identification raf«r to T«bl« 9.
SOURCE: WEST8EHQ 11979)
Figure 9. Chromatogram showing resolution of a standard mixture of C5-C)0 hydrocarbons
on SE-30 glass capillary column with subambient temperature programming.
-------
TABLE 9. STANDARD GAS MIXTURE MEASURED ON SE-30 CAPILLARY
COLUMN WITH SUBAMBIENT TEMPERATURE PROGRAMING
Peak No.*
1.
2..
3.
4.
5.
6.
! 7.
o
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23-
24.
25.
26.
27-
28.
29.
30.
31.
Chemical Name
2-Methyl-1-butene
2-Methylbutane
Halocarbon
1 -Pentene
3-Methyl-1-butene
n-Pentane
Isoprene
Carbon disulfide
t-2-Pentene
c-2-Pentene
2-Methyl-2-butene
2,2-Dimethylbutane
Cyclopentene
Cyclopentane
4-Methyl-1-pentene
2,3-Dimethylbutane
2-Methylpentane
t-4-Methyl-2-pentene
c-4-Methyl-2-pentene
3-Methylpentane
Peak No.*
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
49 A.
50.
51.
52.
53.
54.
55.
56.
2-Methyl-1-pentene
1-Hexene
n-Hexane
t-2-Hexene
2-Methyl-2-pentene
t-3-Methyl-2-pentene
c-2-Hexene
Methylcyclopentane
c-3-Methyl -2-pentene
2,2,3-Trimetnylbutane
1,1,1 -Trichloroethane
2,4-Dimethylpentane
Benzene
1-Methylcyclopentene
Cyciohexane
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
66A.
67.
67A.
68.
32. 2-Methylhexane 69.
33.
34.
35.
36.
37.
38.
2,3-Dimethylpentane 70.
3-Methylhexane
70A.
Dimethylcyclopentane 71.
Dimethylcyclopentane
Dimethylcyclopentane
2,2,4-Trimethylpentane
72.
73.
Chemical Name
n-Heptane
Methylcyclohexane
Trimethylcyciopentane
Ethylcyclopentane
2,5-Dimethylhexane
2,4-Dimethylhexane
2,3,4-Trimethylpentane
Toluene
2,3-Dimethylhexane
2-Methylheptane
3-Ethylhexane
3-Methylheptane
2,2,5-Trimethylhexane
C-9 Alkane
Dimethylcyclohexane
n-Octane
Ethylcyclohexane
C-9 Alkane
Ethylbenzene
p-Xylene
m-Xylene
Styrene
o-Xylene
n-Nonane
i-Propylbenzene
n-Propylbenzene
3-Ethyltoluene
2-Ethyltoluene
1 ,3,5-Trimethylbenzene
1-Ethyltoluene
1 ,2,4-Trimethylbenzene
sec-Butylbenzene
1 ,2,3-Trimethylbenzene
n-Decane
Methylstyrene •
1,3-Diethylbenzene •
1,4-Diethylbenzene i
1 ,2-Diethylbenzene |
C-10 Substituted benzene
C-10 Substituted benzene i
Undecane
'Key to Figures 9 and 11.
36
-------
Tf!'1!' |!!! !'i I! i I! I! n i 11!!'! '! |! |, i 11
! •! |i ' ! i ' : r l ^ • r ' i ! ! : • ' ! I i ' : ' i
I i ' • • _ :_' ' i ' ' • .
r-uvju
... «i«>' i
I _l_U.O_ l—l-J. 1 .i ..I..
r peak Identification refer to Table 10,
HARDEN (1979)
COLUMN 4
TEMPERATURE = 40°C TO 150°C
i
I 1
1
*-t-T3
1
1
; !
1 ; ~
i ! H-
l ' : i
• ;
^ '• : :
" !~i r
! : I
; '
: ! j
ijH
i_j ^
rTrnj
•1
t
3
I
,
!
!
i
i
i
-
-
M
+-
• 1
'
'
) i
l 3-9
0 1 43
.
\_ V
44
^
Li"
"'
'-.
r
=
•:
-
F= =
.^~
n? =
i 7;
_ =.
- :=
^
. nH
•; T
z ~.
1. E
f '~7
L —
jt =
tt ^
g s
± =
Figure 10. Chromatogram showing resolution of a standard mixture of C5-C10 hydrocarbons
on SE-30 glass capillary column with above ambient temperature programming.
-------
TABLE 10. STANDARD GAS MIXTURE MEASURED ON SE-30 GLASS CAPILLARY COLUMN
WITH ABOVE AMBIENT TEMPERATURE PROGRAMMING
Peak No.*
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Chemical Name
Propane + Propene
Propyne
Isobutane + 1 -Butane
—
Cis-2-butene + 2,2-dimethylpentane
1,4-Pentadiene
2-Methylbutane
Isoprene
n-Pentane
t-2-Pentene
2-Methyl-2-butene
2,2-Dimethylbutane
4-Methyl-1-pentene
2,3-Dimethylbutane
2-Methylpentane
3-Methylpentane
2-Methyl-1-pentene
n-Hexane
1-Hexene
2,4-Dimethylpentane
Benzene
2-Methylhexane
3-Methylhexane
n-Heptane
Peak No.*
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
Chemical Name
2,2,4-Trimethylpentane
2,5-Dimethylhexane
2,3,4-Trimethylpentane
Toluene
2-Methylheptane
3-Methylheptane
n-Octane
Ethylbenzene
p/m Xylene
o-Xylene
n-Nonane
Alpha Pinene
n-Propy I benzene
Beta Pinene
m-Ethyl toluene
1 ,3,5-Trimethylbenzene
o-Ethyltoluene
1 ,2,4-Trimethylbenzene
s-Butylbenzene
n-Decane
1 ,2,3-Trimethylbenzene
Limonene
n-Butylbenzene
Key to Figure 10.
In the interim, wet chemical methods may be used to measure total aliphatic aldehydes.
These involve changes in the color of a chemical solution whose absorbance is measured with
a spectophotometer and compared with standards. A sensitivity of 1 ppb is possible, but many
potential interferences exist. Total water soluble aliphatic aldehydes are measured as formal-
dehyde in ambient air. A 10-ml volume of 0.05 percent MBTH (3-methyl-2-benzo thiazolone
hydrazone hydrochloride) is used as a collecting medium. Aliphatic aldehydes react with
MBTH in the presence of ferric chloride to form a blue dye in acidic media (U.S. Public Health
Service, 1965; Joshi, 1978). A 100-liter air sample is necessary to obtain a 1 ppb sensitivity.
Several promising procedures are being actively developed. One of these methods (Grosjean,
1979) entails sampling of large volumes of air on dinitrophenylhydrazine cartridges and analysis by
high performance liquid chromatography (HPLC). The method could potentially measure less than 1
ppb of aliphatic and aromatic aldehydes. Figure 12 shows the separation of a synthetic mixture of
hydrazones of aldehydes. The method, however, has yet to be field tested.
38
-------
to
CO
COLUMN 3
(FOR PEAK IDENTIFICATION
REFER TO TABEL 9)
16 20 29
18
II
24
28
32-33
INCREASING TIME AND TEMPERATURE
SOURCE: HOLDREN. et al. (1979)
Figure 11. Chromatogram of ambient air (500 ml) samples directly from a stainless-steel container.
-------
Ul
5 4
•BENZALDEHYDE
•BUTYRALDEHYDE
ACETONE
• PROPIONALDEHYDE
ACETALDEHYDE
•FORMALDEHYDE
10
20
30 40
U.V. ABSORPTION
50
60
SOURCE: Grosjean (1979)
Figure 12. HPLC-UV separation of a synthetic mixture of hydrazones
of C-|-C4 aliphatic aldehydes, benzaldehyde, and acetone.
5. Atmospheric Tracers
Typically, the analysis of one inert tracer serves as a useful indicator of the degree of
urban contamination. The analysis of fluorocarbon-11 (CC/3f) is recommended. It is one of the
most inert chemicals in the environment and is easily measured. A 5-ml direct injection of air
into a GC-EC system with a 2 m DC-200 column (see Table 8) is adequate for the analysis of
CC/3F Figure 13 shows a chromatogram of CC/3F separation from ambient air. A standard per-
meation tube can be used to establish the CC/3F concentration in a secondary standard (Singh
et al., 1979). In the canisters recommended here for sampling, CC/3Fis stable for a period of
several years. One canister could be set aside as a secondary standard for CC/3F analysis.
Such tracers (along with HC tracers such as C2H2) can often provide useful information about
the nature of the HC mix but do not themselves contribute to photochemical smog.
D. Routine Identification of HCs
Extensive past research has led to the identification of a large number of ambient HCs,
mostly with the help of GC/MS techniques. For routine analysis, retention time can be used as
a reliable identifier of GC peaks. As is clear from Figures 9 and 10, retention times must be
carefully determined and labelled if GC peaks eluting close to each other are not to be con-
fused. Most of the unknown HC species are typically those present in small amounts. There-
fore, direct identification and measurement of only the major species that collectively form a
significant fraction (65-90 percent) of NMHCs is possible at this time.
40
-------
138 ppt
10.0-ml SAMPLE
SUNDAY OCT. 5, 1975
ATT. 2
MENLO PARK, CA
C2HCI3
620 ppt
C2CI4
110 ppt
I
40 36 32 28
SOURCE: SINGH at al (1979)
24
20
MINUTES
16
12
Figure 13. Ambient air chromatogram showing resolution of Fluorocarbon-11
and other halocarbons.
A computerized data system is mandatory when utilizing capillary columns. The operator
should always possess a well-identified marking system containing a list of HCs such as shown
in Table 9. The retention times of individual peaks must be known. For day-to-day operation, a
standard spectrum is extremely useful in assigning the correct chemical name to the peaks
based on retention times. A number of data systems available in the market (e.g., Hewlett
Packard, Perkin Elmer, Spectra Physics), plot the chromatogram and mark the exact retention
time. Relative retention times based on one or more well-characterized species (e.g., benzene
or toluene) are calculated and corresponding peaks identified.
41
-------
Such a process should not be fully automated. Manual checks of measured chromato-
grams with that of the standard should be made frequently. In practice, both retention times
and relative retention times can be reproduced with extreme precision. However, an experi-
enced operator can visually recognize the major peaks and chances of mislabeiing a peak-are
greatly reduced. Automatic identification (based on retention times) using available data sys-
tems is feasible. For the analysis of Ca-C12 HCs, however, we strongly recommend that manual
visual identifications also be made to reduce any chances of error. These problems are much
less severe for C2-C6 HC analysis, largely because of the smaller number of chemical species.
E. Primary Standards and NMHC Species Calibrations
The uniform carbon response of an FID can be used for the calibration of NMHCs with an
accuracy of ±5 to ±8 percent* (Lonneman, 1979). This error is quite reasonable when one
considers that primary standards with an accuracy of much better than ±5 percent are rarely
available. Thus, calibrations carried out with a select group of MCs can be used for the
quantification of all NMHC species.
The single most desired standard is the traceable National Bureau of Standards (NBS)
propane standard. Best NBS estimates suggest that the mean concentration of the NBS pro-
pane standard is within ±2 percent (95 percent confidence limit) of the specified value. NBS
propane standards in air are available in concentrations of 1 ppm, 3 ppm, 10 ppm, 50 ppm, 100
ppm and 500 ppm. Standards are stored in,high-pressure stainless steel cylinders. Propane is
stable over a period of at least 1 year. All standards, whether internally generated or externally
obtained, must be referred to the NBS propane standard.
At SRI a 5-ppm mixture of benzene and toluene in high-pressure aluminum cylinders
(Scott-Marrin, Riverside, California) has remained unchanged (±3 percent) over a 1-year
period. It is recommended that any laboratory conducting HC species analysis should possess
the following standards (preferably in 5-ppm concentrations) for NMHC analysis:
• NBS propane standard in air (STDI)
• Propane + hexane standard in air (STD2)
• Benzene + toluene standards in air (STD3).
All of these are commercially obtained. Most commercial suppliers refer to the NBS pro-
pane standard (STDI), but each laboratory must verify the concentration(s) independently. The
NBS propane standard can be directly used for the calibration of CrC6 HCs. However, STD2
is recommended as a duplicate. STD2 can also be used for C6-Ci2 HC species analysis. Hex-
ane offers a clear advantage because it is conveniently analyzed on columns for light as well
as heavy HC species analysis. Benzene and toluene standard (STD3) is an independent check
on the C6-C12 HC species carbon response.
The 10e linear dynamic range of the FID ensures a linear response over the entire low-
ppm to low-ppb concentration range. This, however, should be routinely tested. Standards
'This is generally not true of HCs containing oxygen (oxygenated species) and chlorine atoms (fluorocarbons. etc.).
42
-------
should always be stored at ppm concentration levels. Any dilutions can be a single step operation
and should be conducted on the day of detailed calibrations. FID response has been found to be
unchanged over a period of many months. A single standard, however, must be used frequently to
confirm the proper operation of the FID system. Detailed calibrations and multiconcentration levels
should be conducted at least once a week. More frequent calibrations may be warranted if the FID
response is found to change, but this should not occur under normal operating conditions.
F. Archiving HC Data
If only a few ambient air samples are analyzed for HC species, the data can be written on individual
sample sheets and examined by simple inspection. If many air samples are analyzed, on the other hand,
particularly over a time period which may be measured in years, such a data storage system is clearly
unsatisfactory. The data must be stored on cards or magnetic tape.
Several years ago, EPA developed AEROS (Aerometric and Emissions Reporting System) for the
purpose of archiving aerometric and emissions data. The sub-system developed for storing pollutant
data is SAROAD (Storage and Retrieval of Aerometric Data). SAROAD works well for the criteria pol-
lutants such as SC>2, 03, NC^, etc.; but, while the sub-system can be made to accept HC species data
from ambient air samples, it is very awkward to do so. Retrieving HC species data from AEROS for analy-
sis is particularly tedious and time-consuming. Because of this, very little of the HC species data from
past studies can be found in AEROS. Instead, the data are stored on sample data sheets, or in project
reports, and therefore have been incompletely analyzed.
An alternative storage system has been devised recently (Nitz and Martinez, 1980) which shows
promise of minimizing some of the problems associated with archiving HC species data. The new system
stores the data by assigning a number to each identified (or unknown) HC species in a sample, which
itself is uniquely identified by sampling time and location. Simple software can be written to analyze the
data from many samples. EPA is storing HC data from its summer studies using this system, and plans to
do so in the future. It is recommended that State and local agencies give serious consideration to storing
their data in this manner also.
43
-------
SECTION 7
QUALITY CONTROL (QC) AND QUALITY ASSURANCE (QA)
Because of the complexity of the ambient air mixture and the relative sophistication of
analytical techniques, a great number of pitfalls are possible. Because of the expense and
effort involved, it is essential to ensure that the collected HC species data meet stringent
requirements. Detailed QC and QA procedures should be developed after the program designs
(including equipment and operational procedures) have been decided. Many of the special pre-
cautions and procedures are addressed in this report. This section describes a sample proto-
col to serve as an example.
Certain areas of the operational procedures lend themselves to systematic checks to
achieve high quality data. These areas are as follows:
• Sample collection
• Sample preconcentration
• Calibration standards
• Calibration of gas chromatograph response
• Verification of gas chromatograph linearity
• Data collection reporting and HC identification
• Referee samples for interlab comparison.
Sample Collection—Loss of data quality due to the sample collection process is usually
attributable to chemical loss on the surfaces of the pressurizing pump, tubing, or sample canis-
ters as well as outgassing of species from sampling lines. Samples must be collected by
means of a noncontaminating pump (stainless steel bellows pump such as the MB 158) and 2-
liter containers fabricated from electropolished (passivated) stainless steel. Sample canisters
should be baked (100'C) and flushed with zero air before use to remove any HCs present as a
result of the manufacturing process. Ultrapure air should be analyzed before and after sam-
pling to ensure cleanliness of the entire sampling protocol. The same procedure should be
repeated with a complex HC mixture to ascertain the absence of surface removal problems.
Three sample canisters or 10 percent of the inventory of canisters, whichever is greater, should
be checked to verify storage stability of ambient air. This air, after storage of 1 week in the
sample canister, should exhibit a combined HC concentration within ± 5 percent of that at the
start. Ambient sample storage in stainless steel canisters should be limited to 1 week prior to
analysis. Samples are collected in duplicate and all collected samples must be analyzed.
Sample Preconcentration—Errors in HC measurements are sometimes attributable to the
preconcentration system, or, with far less likelihood, sample toss on the media used for cryo-
genic preconcentration. Leak checks should be conducted on the preconcentration system
routinely or whenever calibration information indicates an instrumental change or malfunction.
44
-------
Loss of sample during preconcentration can be verified by comparing instrument response to
calibration standards with the preconcentration system bypassed and in-line. A daily stan-
dards check will be helpful in the identification of these problems.
Calibration Standards—Potential measurement errors can occur due to changes in the HC
concentration in multicomponent mixtures contained in compressed gas cylinders. Com-
pressed gas standards are used to calibrate instrument response to specific HCs, to calibrate
flame detector linearity, and to evaluate system sample loss. System sample loss may be
caused by leaks within the plumbing by adsorption or reaction of the sample with the internal
surfaces of the system. Two types of calibration standards can be employed. First, two com-
pressed gas working standards containing a mixture of propane/hexane and benzene/
toluene in air in individual anodlzed aluminum cylinders at a nominal concentration of 5 ppm,
can be used. Second, a standard compressed gas cylinder containing propane at a 5 ppm con-
centration certified by the NBS shall be used as the reference standard. The NBS propane
standard should be used as a primary standard within the period of certification (about 1 year)
and will serve as a reference to evaluate the concentration and the stability of the multicom-
ponent working standard.
Calibration of Gas Chromatograph Response—Calibration of instrument response by injec-
tion of the multicomponent working standard should be performed after every fifth ambient
sample, or twice per day, whichever is more frequent. The precision of the single point calibra-
tion should be ±5 percent or better. If the precision of the calibration is not within ±5 percent,
then the system should be recalibrated, checked for leaks, and appropriate trouble shooting
procedures applied. Calibration of gas Chromatograph linearity is also recommended whenever
the precision of the single point calibration deviates by more than 5 percent from previous cali-
bration.
Calibration of Gas Chromatograph Linearity— The linearity of the gas Chromatograph should
be verified once each week of operation. A diluted gas standard is prepared using an evacu-
ated 1 to 5 liter stainless steel canister. A known volume of working calibration standard is
injected through a diaphragm (or with a syringe) into the canister and the canister is pressured,
by zero air, to 40 psig. A 100 to 200 ppb concentration standard should be prepared in this
way. A series of sample loops from 50 ml to 1 ml (or cryogenic trapping of 10 to 100 ml
volumes) are used to obtain instrument response with at least five HC concentrations. When
sample size versus instrument response is plotted, the results should fall along a straight line.
Data Collection, Reporting and HC Identification—The reports should contain records of gas
Chromatograph response to single point calibration standards and to multipoint calibrations for
instrument linearity. The results of ambient sample analysis should contain tabulations of HC
identification (if known), elution time, and gas Chromatograph response. Samples of important
HCs on the urban atmosphere list should be obtained and used to verify elution time during
analysis on the equipment of interest.
Reference Samples for Interlab Comparison—All samples should be collected and analyzed
in duplicate. A total of 5 percent of the ambient HC samples collected will be used for interla-
boratory comparison. Specific details as to distribution of referee samples will be developed
within the HC measurement program. A maximum delay from collection to analysis may not
exceed 1 week. The referee sample processing will provide independent audit of each
laboratory's analytical procedures.
45
-------
SECTION 8
RECOMMENDED MEANS FOR OBTAINING HC SPECIES DATA
Samples of ambient air are collected on selected days in summer months for analysis of
their NMHC concentrations. It is recommended that such data could be most efficiently
acquired by outside contractual procurements. In-house capabilities should be considered but
the expense involved in acquiring sophisticated instrumentation as well as the difficulties
involved in effective use of personnel with dedicated expertise in HC species analysis makes
this an expensive option. Outside contractors can bring together single or multi-institutional
experience and are available when needed. This makes it highly cost-effective. Hence the use
of outside contractors is recommended.
Currently, the number of outside contractors that could handle future needs of state and
local agencies are somewhat limited and generally tend to be research-oriented organizations.
State and local organizations should obtain guidance from their regional offices for suggestions
regarding the suitability of private contractors.
Cost of analysis is highly variable and depends upon the number of samples to be
analyzed. For detailed C2-C12 HC species the average cost can be expected to range between
150-300 dollars per sample.
46
-------
SECTION 9
KNOWLEDGE GAPS AND RECOMMENDATIONS
Several recommendations became evident during the conduct of this study. These arise
because HC species identification and measurement procedures are still deficient and need to
be further improved. These improvements are likely to increase the complexity as well as the
expense of analysis. Alternate methodologies that simplify these complex procedures should
therefore be vigorously pursued. The following specific recommendations are proposed:
• Continuous NMHC analyzers provide an inexpensive and routine means of HC data
collection. At this time, there is no clear concentration cutoff above which these become
acceptable, although 1 ppm C is commonly assumed. If technological improvements (e.g.,
prior removal of CHU) can be implemented to improve their reliability such options should be
further considered.
• An unacceptably high fraction (10 to 35 percent) of urban HCs remain unidentified.
Additional research efforts especially with the GC-MS should be conducted to reduce this to
less than 5 percent. Until then, specific methods to deal with unidentified HCs should be
developed for air quality model applications
. To date, C2 to C6 HCs have been measured with packed columns and many small
peaks (e.g., butenes) are inadequately resolved. Capillary columns should be tested for
the analysis of C2-C6 HC species to better resolve these light HC species.
• Currently, oxygenated HCs are measured with questionable accuracy. Additional research,
especially for the analysis of aldehydes, is highly recommended. Even though aldehydes
may constitute only a small fraction of total NMHC «5 percent), their impact on smog
formation can be disproportionately large.
» Preliminary analysis of available (Appendix) data indicates that the CrC5 HC fraction,
which is also the easiest to measure, is a relatively good indicator of total NMHCs and
the associated chemical groups. Because of the tremendous simplification this brings,
detailed studies should be performed to analyze available HC species data to test,
improve and verify this concept. We recommend that these analyses also include
meteorological factors to further improve CrC5 HC relationships with total NMHCs and
individual chemical groups. Such data analysis should also be used to better define
the unknown HC fraction in various regions.
47
-------
Appendix
ANALYSIS OF AVAILABLE AMBIENT HC SPECIES DATA
A program of detailed HC species analysis requires expensive instrumentation, experi-
enced personnel, and much time, at best. To test whether a select group of easy-to-measure
HCs can be used to characterize the total NMHC concentration of ambient air and the distribu-
tion of groups of HCs (e.g., alkanes, alkenes), detailed HC species data were analyzed from two
regions:
• Houston, Texas
• South Coast air basin (SCAB), California.
The Houston data were obtained for summer 1977 from the Research Triangle-Institute
(RTI) (Denyszyn et al., 1979) and summer 1978 from EPA (Nitz and Martinez, 1980), and the
SCAB data were collected in summer 1975 (Mayrsohn et al., 1976). Locations of these sites
ate shown in Figures A-1 and A-2. For each case four sites were selected and 0600-0900
average data acquired for a period of 10 days per site. The 0600-0900 h integrated average
was selected because of the well-established importance of this high emission period in air
quality applications. Ten days were randomly selected (in some cases, only 10 days of data
were available) and no attempt was made to select only those days with similar wind conditions
during this preliminary analysis.
Tables A-1, A-2, and A-3 summarize this information in some detail. The percentages of
alkanes, alkenes, C^Hj, aromatics, and aldehydes, and the standard deviation (cr) associated
with these, were calculated based on total identified HCs (INMHC). The percentage of C2 to C5
HCs and the associated o- were calculated based on both identified HCs as well as total HCs
(identified and unidentified HCs). C2 to C5 HCs, which are easily measured, may provide a reli-
able index of total HCs and their distributions. In addition, total NMHC as well as the total
identified NMHC are reported in Tables A-1, A-2, and A-3. The RTI data (Table A-1) did not
report unknown HCs.
Tables A-1 through'A-3 clearly show that total NMHC concentrations in the range of 0.1 to 3
ppm C were observed. Continuous NMHC monitors that seem to operate well above 0.5 to 1 ppm C
would probably produce some unreliable data in the HC concentration range encountered. The need
for HC species measurements is therefore clearly justified. Tables A-2 and A-3 further show that 10
to 35 percent of the total NMHC fraction was unidentified in Houston and SCAB, a significant
fraction of the total NMHC burden. This unidentified fraction could be a major source of error in the.
exercise of an air quality model that requires as input details other than total NMHCs.
Table A-1 shows that in Houston in 1977 the NMHCs varied significantly at each site
(oVmean — 75 percent) as well as from site to site; this was not the case with the CrC5 HC
fraction, however. At each site (Table A-1) the CfCs fraction.of the NMHC burden had a varia-
bility of about 15 to 20 percent Therefore, in Houston the CrCs HC fraction was a reasonable
indicator of NMHCs. As can be expected, the chemical group concentration varies somewhat
more. The largest fraction is alkanes and this varies less than 10 percent The somewhat large
48
-------
Htmpttutd
UJ
O
»
Q
-80
-40
-20 0
DISTANCE — km
20
40
Figure A-1. Locations of selected monitoring sites in the Houston area.
49
-------
Figure A-2" Southern California sampling tites for 1975.
TABLE A-1. HOUSTON AMBIENT HC COMPOSITION BASED ON DATA
COLLECTED BY RTI IN AUGUST-SEPTEMBER 1977
Hydrocarbon Category
Alkanes (%)
Alkenes(%)
CjH, (%)
Aromttics (%)
Aldehydei (%)
SCj-Cj/jINMHC (%)t
2INMHC (ppb C)
2UINMHC (ppb C)T
^NMHC (ppb C)
Site
Aldin Mall
73.6 ± 5.8*
4.7115
4.8 ±1.3
163 ±5.5
0.0 ± 0.0
63.0 ± 9:5
589 ±448
—
—
Clinton Drive
75.9 ± 5.8
4.4 t 2.0
3.2 1 1.8
16.6 ± 4.8
0.0 ± 0.0
59.8 ± 11.6
7382301
—
—
Fuqua
73.4 ± 10.6
5.6 ± 3.8
3.6 ± 1.8
17.4 ± 7.5
0.0 ± 0.0
68.7 ±11 .2
574 ±413
—
—
Crawford
72.9 ± 5.4
4.3 ± 0.6
5.9 ± 1.9
16.7 ±6.2
0.0 ± 0.0
55.1 ± 9.3
890 ±356
—
—
Standard deviation
'I - Identified; Ul - Unidentified
50
-------
TABLE A-2. HOUSTON AMBIENT HC COMPOSITION BASED ON DATA
COLLECTED BY EPA IN SEPTEMBER 1978
Hydrocarbon Category
Alkanes (%)
Alkenes (%)
f* LJ IV.\
**2""2 \nl
Aromatics (%)
Aldehydes (%)
2C2-C5/2;INMHCt(%l
ZCj-Cg/jrNMHC (%)
SlNMHC (ppb C)
ZUINMHC(ppbC)T
£NMHC (ppb C)
Site
May Drive
70.6 ±8.2*
10.4 ±2.8
1.8*0.9
16.0 ±5.0
1.3*1.3
615*7.6
455*12.9
985*612
344*409
1329*808
Crawford
61.3*3.2
11.3*1.2
3.1 ±0.5
23.4 ± 2.4
0.9 ±0.9
51. 4 ±3.7
41. 4 ±6.3
1060 ±472
308*428
1368 * 851
Parkhurst
66.6 * 5.6
9.1 ±1.9
2.5 ±0.8
20.2 ±2.9
1.6 ±1.7
54.4 ±5.4
42.7 ±9.4
479 ± 246
163 ±262
642 ±405
Lang
57.9 ± 8.9
11.8 ±3.3
4.6 ± 4.8
24.6 ± 6J3
1.2*1.1
53.3 * 3.4
43.8 ± 6.9
736 * 303
188*219
924 * 384
Standard deviation
Tl - Identified; Ul - Unidentified
TABLE A-3. SOUTH COAST AIR BASIN (SCAB) AMBIENT HC COMPOSITION BASED
ON DATA COLLECTED BY ARB IN AUGUST-SEPTEMBER 1975
Hydrocarbon Category
Alkanes (%)
Alkenes (%)
C2H2 (%)
Aromatic* (%)
Aldehydes (%)
ZCj-Cg/SlNMHC (%)'
ZCj-Cg/ZNMHC (%)
IINMHC(ppbC)
EUlNMHCfppbC)1
INMHC (ppb C)
Site
Upland
61. 9 ±6.7*
63 ± 2.0
4.1 ± 1.2
27.7 * 7.1
0.0 ± 0.0
45.0 ±12.6
38.6 ±125
1072 * 887
244*445
1316* 1240
Long Beach
73.4 ± 7.1
5.4 ± 1.0
3.0 * 0.7
18.3*7.5
0.0 ± 0.0
60.4 * 7.4
51.0*7.0
1066*885
160*257
1066*885
El Monte
66.4 ± 5.7
7.4 * 2.0
4.0 ± 0.6
22.2 ± 6.7
0.0 * 0.0
52.8*4.1
43.2 ± 5.9
958*999
229 * 273
1187*610
Azusa
60.7 * 4.0
5.9 * 0.7
3.5 * 0.4
29.8 ± 4.3
0.0 * 0.0
42.2 * 3.6
30.0 ± 6.3
1189*595
614 ± 907
1803± 1179
Standard deviation
Tl - Identified; Ul - Unidentified
51
-------
variability of the alkefie fraction ^attributable-to the lack of-ethylene data. Which was
'estimated based on Washington State University data (Westberg, et al., 1978) [C2H4(p0b
-C)/C2H2(ppb C) = 0.74].
Table A-2 from EPA data of 1978 shows very similar behavior*'Once again, CrC5 HCs
can be used to represent total identified HCs with an uncertainty of about 10 percent, total
NMHCs, however, are also estimated with an uncertainty of about 20 percent. At each site, the
variability of the alkane, alkene, and aromatic fractions is not large. The largest identified frac-
tion is alkanes, which make up about 60-70 percent of the identified HC fraction.
The results from SCAB are nearly the same. Table A-3 shows that the CrCs fraction can
be used to estimate NMHCs with an uncertainty of about 25 percent. At each site the
variability of individual chemical groups is quite small and hence it is feasible that these distri-
butions can also be estimated from CrC5 HC fractions.
Because of the great expense involved in the analysis of Ce-C12 HCs, we find it highly
useful that fora given site CVCs HCs can be used to estimate total NMHC concentrations with
an uncertainty of less than ±30 percent These varlablilities could be further reduced when
meteorological details are incorporated (e.g., when days with identical wind directions are con-
sidered). These options should be studied in more detail. In practice, a few cases can be
analyzed for detailed HCs and, In the remainder of'cases, CrC5 HCs used to estimate NMHC
and the contribution of individual chemical groups. An uncertainty of ±30 percent is not large
when one considers that 15-35 percent of the HC burden is unidentified and measurement
errors of the order of 10-15 percent are typical, especially in the analysis of C6-C12 HCs.
Preliminary results of our analysis appear encouraging. -More detailed analysis of avail-
able data should be conducted to further verify and refine these simplification procedures.
Comparison of HC species data for days with similar meteorological conditions would be
revealing.
'Here it is not our intent to compare EPA 1978 Houston data with RTI 1977 Houston data. Internal discrepancies are
mostly attributable to measurement problems rather than a significant change In'the air mix As an example, the RTI
benzene/toluene ratio was highly unusual. In addition, ethylene data from RTI were not available and had to be
estimated from a mean CjHjCjHj ratio.
52
-------
REFERENCES
Denyszyn, R.B., E. Pellizzari, J.M. Harden, and D.L. Hardison (1979), Houston Area Oxidant Study
Detailed Hydrocarbon Analysis, RTI/1456/00-01F, Research Triangle Institute, North Caro-
lina.
Grosjean, D. (1979): Application of a New Method to Measurements of Ambient Levels of
Aldehydes in Urban Air, Private Communication, ERT-lnc., Westlake Village, California.
Harden, J. (1979), private communication, Research Triangle Institute, Research Triangle Park,
North Carlolina.
Holdren, M., S. Humrickhouse, S. Truitt, H. Westberg, and H. Hill (1979), Analytical Techniques
to Establish the Identity and Concentration of Vapor Phase Organic Compounds, unpub-
lished manuscript, Washington State University, Pullman, Washington.
Joshi, S.B. (1978), Houston Field Study (1978)—Formaldehyde and Total Aldehyde Monitoring
Program, Document ESC-TR-79-22, Northrop Services, Inc., North Carolina.
Lonneman, W.A., S.L Kopczynski, P.E. Durley, and F.D. Sutterfield (1974), Hydrocarbon Compo-
sition of Urban Air Pollution, Env. Sci. and Technol.. Vol. 8, pp. 229-235.
Lonneman, W.A., and J.J. Bufalini (1979), Correspondence, Env. Sci. and Technol., Vol. 13, pp.
236-239.
Lonneman, W.A. (1979), private communication. Environmental Sciences Research Labora-
tories, EPA, Research Triangle Park, North Carolina.
Ludwig, F.L. and E. Shelar (1978), Site Selection for the Monitoring of Photochemical Air Pollu-
tants, EPA 450/3-78-013. SRI International, Menlo Park, California.
Mayrsohn, H., M. Kuramoto, J.H. Crabtree, R.D. Sotherns, and S.H. Mano (1976), Atmospheric
Hydrocarbon Concentrations—June-September 1975, State of California Air Resources
Board, California.
Nitz, K. and J.R. Martinez (1980), Data Base for the 1978 Houston Oxidant Modelling Study,
Draft Report Project SRI 7938, SRI International, Menlo Park, California.
Seila, R.L., W.A. Lonneman, and S.A. Maks (1976), Evaluation of Polyvinyl Fluoride as a Con-
tainer Material for Air Pollution Samples. J. Env. Sci. and HealthlEnv. Sci. Eng., Vol. A11, No.
2, pp. 121-130.
Singh. H.B., et al. (1979), Atmospheric Distributions, Sources, and Sinks of Selected Halocar-
bons. Hydrocarbons, SF8 and A/2O, Final Report Project 4487, SRI International. Menlo
Park. California.
53
-------
Spicer, C.W. (1979), private communication, Battelle Columbus Laboratories, Columbus, Ohio.
U.S. Public Health Service (1965), Selected Methods for the Measurement of Air Pollutants.
Publication No. 999-AP-11, Cincinnati, Ohio.
Westberg, H. (1979), private communication, Washington State University, Pullman, Washington.
Westberg, H., R.A. Rasmussen, and M. Holdren (1974), Gas Chromatographic Analysis of
Ambient Air for Light Hydrocarbons Using a Chemically Bonded Stationary Phase, Anal
Chem.,46, 1852-1854.
Westberg, H., K. Allwine, and E. Robinson (1978), Measurement of Light Hydrocarbons and Stu-
dies of Oxidant Transport Beyond Urban Areas - Houston Study 1976, Contract No. 68-
02-2298, Washington State University, Pullman, Washington.
54
------- |