&ER&
United States
Environmental Protection
Agency
Environmental Monitoring and
Support Laboratory
Research Triangle Park NC 27711
EPA--600/4-80-027
May 1980
Research and Development
Ambient Air
Monitoring foi
Benzene
24-Hour Integrated
Sampling in Six Cities
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
AMBIENT AIR MONITORING FOR BENZENE
24-HOUR INTEGRATED SAMPLING IN SIX CITIES
by
Barry E. Martin
Environmental Monitoring Branch
Thomas Clark
Quality Assurance Branch
Joseph Bumgarner
Analytical Chemistry Branch
Gary F. Evans
Analysis and Reports Branch
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring Systems
Laboratory, U.S. Environmental Protection Agency, and approved for publi-
cation. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
FOREWORD
Measurement and monitoring research efforts are designed to anticipate
potential environmental problems, to support regulatory actions by developing
an in-depth understanding of the nature and processes that impact health and
the ecology, to provide innovative means of monitoring compliance with regu-
lations and to evaluate the effectiveness of health and environmental pro-
tection efforts through the monitoring of long-term trends. The Environmental
Monitoring Systems Laboratory. Research Triangle Park, North Carol ir,a5 has
the responsibility for: assessment of environmental monitoring technology
and systems; implementation of agency-wide quality assurance programs for
air pollution measurement systems; and supplying technical support to other
groups in the Agency including the Office of Air, Noise and Radiation5 the
Office of Toxic Substances and the Office of Enforcement.
The declaration of benzene as a hazardous material provided the impetus
for monitoring benzene in the atmosphere. Since benzene is one of the pol-
lutants associated with internal combustion engines, monitoring sites for
the present study were selected in metropolitan areas with high traffic
density. Existing air quality monitoring sites were utilized to provide
quick response. Since a method of monitoring benzene has not been standar^-
ized, the best available technology was employed. As deficiencies in the
method became apparent, short-term research was applied to improve it
quickly. Exact data were not obtained because high, variable concentrations
of residual benzene on the adsorbent masked the sample concentrations.
However, sufficient data were collected to assess the potential exposure of
the population to benzene.
Thomas R. Hauser, Ph.D.
Director
Environmental Monitoring Systems
Laboratory
Research Triangle Park, N.C.
-------
ABSTRACT
This monitoring program was initiated to assess the population exposure
to ambient levels of benzene in conjunction with the announcement of benzene
as a hazardous material. The program was conducted during the summer and
fall of 1977.
Average benzene concentrations of 5, 18, and 19 yg/m , respectively,
were measured in the widely separated cities of Dallas, Chicago, and Los
Angeles. The precision of these data is estimated to be ±26 percent of the
measured values.
The improvement of benzene monitoring and analysis technology was an
important secondary objective of this program. However, further research is
needed to identify a more favorable collection medium than the Tenax that
was used in this study.
-------
CONTENTS
Forward ................................. iii
Abstract ................................ iv
Figures ................................. vi
Tables ................................. vi
Acknowledgment. ............................. vii
1. Introduction ......................... 1
2. Summary and Conclusions .................... 3
3, Experimental Method „ .................... 4
Sample Collection .................... 4
Benzene Diffusion in Tenax ............. 4
Sampling Rate .................... 5
Quality Control Samples ............... 6
Sample Analysis ..................... 6
4, Quality Control Procedures .................. 8
External Quality Assurance Procedures .......... 8
External Quality Assurance Samples ............ 8
External Quality Assurance Standards for Benzene ..... 9
Internal Quality Control Procedures ........... 9
Internal Quality Control Samples ............. 9
5. Study Results ......................... 12
Benzene Concentrations .................. 12
Quality Control Results ................. 12
Accuracy ........... 12
Precision 13
6. Problems and Recommendations ......... ... 15
Appendix A. A Method of Determining Benzene In the Atmosphere ..... 17
References ..... . ............ 30
-------
FIGURES
Number Page
1 Benzene generating system 7
2 Two-stage benzene dilution system 11
A-l Sampling cartridge 19
A-2 Single sampler for benzene 20
A-3 Duplicate sampler for benzene 21
A-4 Tandem sampler for benzene 22
A-5 Apparatus for sample recovery and analysis 27
A-6 400-W modified desorption heater 28
TABLES
Number Page
1 Residual Benzene in Unexposed Tenax
from Sample Cartridge (No. 1053) 5
2 Ambient Benzene Concentrations in Chicago, Dallas,
and Los Angeles during Summer and Autumn of 1977 (yg/m ) ... 14
VI
-------
ACKNOWLEDGMENT
The authors wish to acknowledge the substantial contribution to the
method development, quality assurance, and analysis of field samples for
benzene which was provided by the Research Triangle Institute, operating
under EPA Contract No. 68-02-2725.
Vll
-------
SECTION T
INTRODUCTION
As a result of the designation of benzene as a hazardous air pollutant
under Section 112 of the Clean Air Act, monitoring data were required to
assess general population exposures and to evaluate the population risk.
Preliminary monitoring data were desired by the middle of August 1977.
Although a standard method for monitoring ambient benzene did not
exist, data were to be collected immediately using the best available
technique, rather than by developing a method through an experimental
program. As problems became apparent, they were resolved by short-term
research. Hence, the method described in Appendix A represents the most
recent techniques .used, but these were not necessarily throughout the
benzene monitoring program.
In July 1977 the objective was to take ambient air samples for 10 days
in six major metropolitan areas. Monitoring was extended into October 1977
for three cities, however, because the results of these early measurements
indicated that improvements in the monitoring method were needed.
Monitoring was planned in metropolitan areas with high traffic densities
because the major sources of benzene are automobile emissions and gasoline
evaporation. Since CO is also an automotive pollutant, siting criteria
already published for CO were useful in selecting the benzene monitoring
sites.
Five sampling locations were chosen within each of the six cities: New
York, St. Louis, Los Angeles, Dallas, Boston, and Chicago. A range of urban
characteristics was represented by the sampling sites, including street
canyon, peak neighborhood, corridor, urban and background area.
A quality control program was implemented to minimize the introduction
of erroneous data. Due to the time restraints and nature of the study, the
quality control program was somewhat limited, but it did include controls
over benzene standards used for calibration, preparation of sampling cartridges,
sampling, and analysis of samples.
To complete the monitoring program in the allocated time using the best
available techniques, various tasks were assigned to several branches of the
Environmental Monitoring Systems Laboratory (EMSL), as follows.
-------
The Environmental Monitoring Branch (EMB) was required to visit cities
and select sites; procure, calibrate, transport, and install sampling and
support equipment; perform sampling; and implement a quality control program
for sampling.
The Quality Assurance Branch (QAB) provided technical expertise in
methods development and quality control procedures; provided approximately
50 quality control samples for random insertion among field samples; and
performed sample liaison (receipt, logging, and distribution of quality
control and field samples).
The Analytical Chemistry Branch (ACB)was assigned to prepare approxi-
mately 600 sampling cartridges (500 for field sampling and 100 for QAB
work); develop analytical methodology; and analyze field and quality assur-
ance samples. Much of the method development and analytical work was actu-
ally performed under contract by the Research Triangle Institute (RTI).
The Statistical and Technical Analysis Branch (STAB) statistically
analyzed the data; assisted with the design of sampling and quality control
procedures; and coordinated the preparation of a final project report.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
The average ambient concentrations of benzene in Dallas, Chicago, and Los
Angeles were estimated at 5, 18, and 19 ug/m , respectively. The positive
bias associated with the observed values is estimated at 44 percent of the
true value as determined by control samples. The coefficient of variation
for an individual analysis made by the ambient benzene measurement
method was approximately 26 percent.
The major reason for the considerable measurement error in the analyses
was the large and variable concentration of residual benzene in the Tenax
used for sample adsorption. Removal of impurities from the Tenax polymer
resin was difficult and time consuming, but after cleanliness was apparently
achieved (as verified by analysis) 9 benzene was likely to reappear for at
least two reasons: (1) the Tenax was most probably contaminated with benzene
during the manufacturing process, and its presence was detected later due to
diffusion to the surface; (2) the polymer resin material might alter to
benzene as a result of chemical reaction.
The benzene monitoring program contributed to improvements in the
system design used to desorb the collected benzene from the Tenax and to the
development of a system to consistently and accurately prepare quality
control samples.
Before a benzene measurement method used can be standardized, further
research is needed to find a better adsorbent material for benzene and to.
develop a sampling cartridge that can be reliably sealed and conveniently
used. The adsorbent material should offer a consistently low benzene back-
ground, and not react chemically with benzene or other compounds to form
benzene.
-------
SECTION 3
EXPERIMENTAL METHOD
The sample collection and analysis techniques used in the study are
presented in this section. Short-term research that was conducted to
validate a technique or to resolve a specific problem is also discussed.
SAMPLE COLLECTION
Benzene was adsorbed onto Tenax resin beads packed in a sampling
cartridge. The cartridge was prepared from a cleaned stainless steel tube
of 7.9 mm inside diameter (i.d.) and 300 mm length. The Tenax was cleaned
and packed in the center section (250 mm) of the tube, and a glass wool
plug (25 mm length) was placed in each end to hold the Tenax in position.
Both ends were sealed with compression fittings to complete the assembly of
a sample cartridge. In field use, the fittings were removed and one end was
connected to an ambient air sampler for the collection of a 24-hour benzene
sample. After sampling, the compression fittings were reinstalled and the
cartridges sent to the analysis laboratory.
Details of the sampling procedure are presented in Appendix A. As the
appendix indicates, the samples were refrigerated at -18°C (except during
sampling) to reduce the diffusion of benzene from the Tenax.
In support of the monitoring study, research was conducted to assess
the diffusion of benzene in Tenax resin, to determine the maximum sampling
rate without benzene breakthrough, to verify the uniformity of quality con-
trol samples, and to assess the adequacy of the cartridge sealing procedure.
This research is described below.
Benzene Diffusion in Tenax
One of the prepared, unexposed sampling cartridges was first analyzed
for benzene, then stored at room temperature for 3 days, and later analyzed
twice in successive runs on the same day. The sample cartridge again was
stored at -18°C for 3 days and analyzed, then stored 1 day at -18°C and
analyzed, and finally stored another 13 days at -18°C and analyzed. After
the original analysis, the Tenax was exposed to hydrocarbon-free (HCF) air
in an attempt to cleanse the benzene from the resin, but subsequent analyses
revealed that this exposure did not affect the quantity of benzene measured
later. Table 1 summarizes the above procedure and the results of the analysis
-------
An unexpected result occurred in the second run on the second day of
analysis (10/21/77): no measurable benzene was detected in the Tenax resin
in which residual benzene had just been measured. A possible explanation
for this effect is that benzene, which apparently was dispersed throughout
the Tenax resin, had been desorbed from the surface of the Tenax beads
during the day's first analysis, but was not replaced by benzene diffusing
from the interior of the beads before the second analysis. Consequently, no
benzene was measured during the second analysis.
Whether diffusion occurs only at elevated temperatures (20°C and above)
is unknown. Therefore the sample was stored at -18°C for 3 days and analyzed.
Since a relatively high benzene concentration was found, the cartridge was
refrigerated for a shorter period of time to assess the effects of time. In
this case, a significantly lesser quantity of benzene was measured, so the
sample was refrigerated for a substantially longer period (13 days). The
concentration found from this time period was approximately the same as that
of the 3-day storage, which indicates a maximum value is attained within about
3 days. The overall results, however, are somewhat inconsistent because
higher benzene concentrations were recorded after 3 days of refrigerated
storage than had been measured at room temperature.
TABLE 1. RESIDUAL BENZENE IN UNEXPOSED TENAX FROM SAMPLE CARTRIDGE (NO. 1053)
Date
10/18/77
10/21/77
Tenax Conditions
Blank as received from EPA
HCF air loaded. Tube had been stored
Benzene
39
39
(ng)
.2
.0
10/21/77
for 3 days at room temperature
Analyzed immediately after previous
run (no connections were broken)
0.0
10/24/77
10/25/77
11/8/77
Storage
Storage
Storage
at
at
at
-18°
-18°
-18°
C
C
C
since
since
since
previous
previous
previous
run
run
run
79
'35
85
o
.0
.0
Sampling Rate
A prime consideration in selecting the appropriate sampling rate for
Tenax was ensuring the availability of sufficient pollutant quantity for
good analytical resolution, without losing any of the sample (i.e., without
breakthrough occurring). To detect breakthrough, two cartridges were placed
in series. Analysis of the second (downstream) cartridge indicated whether
or not breakthrough in the first (upstream) cartridge had occurred.
-------
In many experiments with various sampling flow rates, breakthrough
generally occurred after 17 liter of ambient air had passed through 2.2 g of
Tenax. For the present monitoring study, a safety factor of approximately
2.5 seemed advisable, so the total sampling volume was set at approximately
7 liter. During the 24-hour sampling a flow rate of 5 cm /min was maintained.
Quality Control Samples
Quality control samples were prepared with known concentrations of
benzene adsorbed on Tenax. To generate the benzene, a permeation tube with
nitrogen carrier gas was pulled through a sample cartridge, as indicated
in Figure 1. At the beginning of the program, six quality control sample
cartridges were prepared simultaneously using a manifold. The total flow
rate was measured and was assumed to be uniformly distributed among the
cartridges. However, analyses of some of these cartridges later revealed
nonuniform concentrations. Presumably the Tenax was packed nonuniformly in
the six tubes, resulting in a variable pressure loss and flow rate through
the cartridges. After the nonuniformity was determined and rectified by
individual preparation of samples, good comparisons among cartridges were
attained.
SAMPLE ANALYSIS
To analyze a sample, benzene was measured by (1) recovery with heat
desorption, (2) cryogenic trapping on silanized glass beads in liquid
nitrogen, (3) separation in a specially prepared column, and (4) analysis
with a flame ionization detector (FID). Specifications of the apparatus
and of the operating conditions are presented in Appendix A.
Each sample was desorbed at a temperature of 250°C for 10 minutes
utilizing a 400-W heater. Developmental work was required in order to dis-
tribute the heat uniformly over the exterior surface of the sample, and to
raise the sample temperature rapidly. The samples were rapidly placed in
the heater and connected to the sample recovery apparatus while the heater
was operating. Heat was generated by two diametrically opposed coils placed
along the length of the cartridge. An aluminum enclosure both reduced heat
loss and, through high thermal conductivity, assisted with circumferential
heat distribution. Rollers on the heater assembly and quick-connect fas-
teners on the sample cartridge assisted with attaching the hot assembly to
the sample recovery tubing.
-------
BENZENE
PERMEATION
TUBE .
IITROGEN
WATE:R
^
WATER-JACKETED
CONDENSER
•TEMPERATURE
MEASUREMENT
8ENZENE-
NITROGEN
FLOW
MIXING
FLASK
SAMPLING
CARTRIDGE
CONTROL
VALVE
—ROTAMETER
Figure 1. Benzene generating system
-------
SECTION 4
QUALITY CONTROL PROCEDURES
EXTERNAL QUALITY ASSURANCE PROCEDURES
An external quality assurance plan was initiated by QAB/EMSL to supple-
ment the quality control procedures developed by the contractor for this
study, for the analysis of the collected field samples. This plan assisted
in determining the performance of analysis procedures used by the .contractor,
and provided a method of qualifying the sample data.
The external controls consisted of randomly inserting among each day's
collected field samples both unexposed (blank) sample cartridges and exposed
samples of known benzene concentrations. These quality control samples were
analyzed as if they were field samples; moreover, because of uncertainties
in the analysis procedures for benzene, a large number of these samples was
utilized with respect to field samples. Two blank cartridges and two
prepared samples were included with every four field samples undergoing
analysis. This large ratio of quality control samples to field samples was
dictated by the high and variable blank sample values that were obtained
using the Tenax material, and the limitation of being able to perform only
one analysis per field sample.
Field sample cartridges were stored under refrigeration until they were
analyzed by the contractor. Both the blank and benzene-exposed control
samples were then randomly introduced to the sample set.
EXTERNAL QUALITY ASSURANCE SAMPLES
Benzene concentrations from 50 to 1000 ng per cartridge were necessary
to provide an adequate quality control check of the contractor's analysis
procedure. To obtain these concentrations, nitrogen was passed over the
benzene permeation tube, through the sample cartridge; the gas was measured
by a rotameter and controlled with a flow control valve (see Figure 1). A
constant flow rate of 57 cm /min ±2 percent was maintained through the
sample cartridge.
The weight loss of the benzene permeation tube was determined gravimetri-
cally and a QC analysis confirmed that benzene was the only compound permeated
Sample concentration was calculated from the gravimetric permeation rate.
The obtained value was validated independently by three laboratories within
Research Triangle Park (RTP), NC, using gas chromatographs (GCs) with FIDs.
-------
Quality control samples were prepared each analysis day and were
presented to the contractor as part of a complete sample set, which consisted
of four field samples, two blank samples, and two benzene-exposed quality
control samples. Routinely, one quality control sample containing approx-
imately 100 ng benzene, and one containing approximately 50 ng benzene, were
sent with the sample set. Periodically, samples with benzene concentrations
as high as 800 ng per cartridge were included.
EXTERNAL QUALITY ASSURANCE VERIFICATION FOR BENZENE
Since standard calibration gases were not available in the concentration
range of interest (0-10 ppsn)9 the concentrations provided by the permeation
tube used to prepare the quality control samples were verified by three RTF
laboratories. At the beginning of the study, the samples were prepared with
HCF air as a diluent. But concentrations determined by analysis and by
weight loss calculations of the permeation tube using this diluent did not
agree within 20 percent„ The discrepancy was traced to hydrocarbon impuri-
ties in the permeation tube, as detected by the GC.
For those quality control samples prepared with HCF air, the average
of the three analysis values from the RTF laboratories was used in the
study. In subsequently prepared quality control samples5 nitrogen was used
as the dilution gas, and benzene permeation tubes were used that were known
to be free of hydrocarbons. As presented later in the quality control
results, strong agreement was obtained Between the permeation rate of the
tubes and GC analyses for the second set of quality control samples.
Toward the end of the monitoring study, a more convenient and equally
accurate method was used to prepare the quality control samples. Cylinders
containing 10 pprn benzene in nitrogen were purchased. This gas mixture was
further diluted with nitrogen and pulled through the sample cartridges. The
calculated sample concentration from the benzene cylinder agreed with GC
analyses of sample cartridges.
INTERNAL QUALITY CONTROL PROCEDURES
An in-house quality control program was performed in order to insure
the proper operation of the analytical system. The program consisted of
having the contractor load known concentrations of benzene onto Tenax in
blank cartridges (supplied by EPA) and then analyzing these cartridges as if
they were field samples. Blank cartridges were also analyzed. Clean air
and nitrogen used in the preparation of quality control samples were intermit-
tenly analyzed throughout the program to ensure no benzene contamination.
INTERNAL QUALITY CONTROL SAMPLES
To evaluate the performance of the analytical system, sample cartridges
were spiked with benzene by the contractor. As mentioned previously, an air
dilution system was one of two methods used to produce quality control samples
-------
This system is illustrated in Figure 2 and consists of two stages of mixing,
both of which received approximately equal injections of HCF air. The flow
from a 9.27 ppm benzene cylinder was diluted with HCF air so that the concen^
tration entering a cartridge was 0.04 ppm. The time of exposure determined
the total weight of benzene collected on the Tenax. Typically, this weight
was 125 ng per cartridge, which is a concentration of 125 yg/m .
10
-------
9-27 ppm
BENZENE
SAMPLING
CARTRIDGE
ORI
FICE-I
LARGE PUMP
Figure 2. Two-stage benzene dilution system
11
-------
SECTION 5
STUDY RESULTS
The monitoring study produced a set of values for ambient benzene
concentrations obtained from sampling sites within each of three cities.
Quality control data were used to estimate the accuracy and precision of
these concentration values.
BENZENE CONCENTRATIONS
Table 2 presents the ambient benzene concentrations observed in Chicago,
Dallas, and Los Angeles. The average values in Chicago and Los Angeles
were approximately equal, but that in Dallas was only about one^fourth as
high. Each value in Table 2 was obtained by subtracting 9 yg/m from the
measured value (the constant 9 is the average correction determined for the
residual benzene in the Tenax).
The amount of residual benzene in the Tenax was equal to a large
percentage of the ambient level. Also, the residual value varied among
unexposed samples. The average quantity of benzene in unexposed (blank)
quality control samples was 48.2 ng with a standard deviation of 30.7 ng.
The lower and upper 95 percent confidence limits were 36.1 and 60,4 ng,
respectively, which translates to concentrations of 7 and 11 yg/m benzene.
Hence, the average value of 9 yg/m was subtracted from each data point, as
stated.
QUALITY CONTROL RESULTS
Accuracy
The accuracy of the analysis procedure was determined from quality
control samples spiked with known quantities of benzene. Quality control
data were expressed as precentages of concentration differences, defined as
follows:
cd = ioo(ca - cs)/cs (i)
where C, = concentration difference between analyzed and spiked values,
percent
3
C = analyzed value of prepared quality control sample, yg/m
cl
C = concentration determined at the time that a quality control
s sample was spiked, yg/m .
12
-------
The value of C,, which is the bias estimate for individual field
samples, is +44 percent.
Precision
Precision was evaluated using replicate analyses of samples
prepared by the contractor. The coefficient of variation for a single
analysis was 26 percent. The value of this coefficient did not appear
to vary as a function of benzene concentration, but wide scatter in
the data precluded regression analysis of precision as a function of
concentration.
13
-------
TABLE 2. AMBIENT BENZENE CONCENTRATIONS IN CHICAGO, DALLAS, AND
LOS ANGELES DURING SUMMER AND AUTUMN OF 1977 (yg/m )
Chicago
26
23
13
9
8
9
14
30
7
14
12
12
22
11
43
29
26
20
22
27
7
9
11
Dallas
4
4
4
8
6
4
4
5
1
1
0
11
3
3
0
5
2
22
Los Angeles
14
11
26
12
15
23
18
20
20
13
17
27
24
17
18
19
23
30
Average 18 5 19
14
-------
SECTION 6
PROBLEMS AND RECOMMENDATIONS
Because of the urgent need for ambient benzene data at the time of this
study, the full development of satisfactory sampling and analytical techniques
was not possible prior to commencement of the monitoring effort. Thus, as
technical problems arose, short-term methods research was performed in
conjunction with the ongoing monitoring program.
Research activities addressed the extent and effects of residual benzene
in the adsorbent material (Tenax) and the development of sampling and analyt-
ical equipment.
The major problems encountered with the use of Tenax as an adsorbent
were that (1) the residual benzene concentration in the Tenax was large
compared to that of a field sample, (2} this concentration was highly variable,
and (3) it could not be completely removed from the Tenax. Since these
effects are believed to be related to the stability and reactivity of the
material, materials other than Tenax should be considered in future benzene
monitoring programs unless a routine procedure can be developed to satisfac-
torily clean the Tenax prior to field exposure. Either an alternative
material or an alternative collection method should be sought. Furthermore,
the material sought should be compatible with solvent desorption of benzene
rather than thermal desorption, for the following reasons:
1. With thermal desorption, only one,analytical result can be obtained for
each sample, thus precluding developmental methods comparisons, analy-
tical accuracy evaluations or replicate statistical analyses for
precision.
2. The extensive tubing lines, fittings, and valves used with thermal
desorption provide opportunities for leaks.
3. External standards must be used for calibration with thermal desorption,
due to the inherent nature of the method, while internal standards such
as those used with solvent desorption are more accurate.
4. Solvent desorption is less costly, less labor intensive, and less time
consuming.
5. With thermal desorption, constant and reproducible temperatures and gas
flow rates are difficult to maintain.
15
-------
Future research is needed to develop a solvent desorplion method,
assuming that an adsorption method is used to collect the sample.
Leakage was also a problem with the sample cartridges. When the caps
were tightened sufficiently to prevent it, the cartridge tubes became so
bulged that subsequent removal of the caps was difficult. After the samples
were collected, some of the tubes required grinding before they could be
resealed. Alternative apparatus should be developed to collect
samples.
16
-------
APPENDIX A
A METHOD OF DETERMINING BENZENE IN THE ATMOSPHERE
The method described evolved during the benzene monitoring program. As
apparatus and procedures were improved, the changes were incorporated in
subsequent analyses. Areas that were investigated were the desorption
heater design, the preparation of uniform quality control samples, sealing
of the sampling cartridges, and the temperature effect on the diffusion of
benzene in Tenax resin.
PRINCIPLE
Benzene was adsorbed from ambient air onto Tenax polymer resin. Subse-
quently the benzene was desorbed by heating the resin, and, through use of a
carrier gas (helium), the concentration was measured chromatographically
with a dual FID.
APPLICATION
The method described is applicable to the measurement of benzene in
ambient air using a 24-hour sampling period. CAUTION: Benzene is a hazardous
air pollutant and care must be exercised to protect operators from breathing
its fumes.
SENSITIVITY AND RANGE
The limit of detection is approximately 0.1 yg/m , and the maximum of
the range is 32,000 yg/m . The range can be increased by extending the
calibration range or by diluting samples in which the concentration exceeds
the maximum limit.
PRECISION
The relative standard deviation of replicate gas chromatographic
analyses of standard gas mixtures is within 26 percent.
17
-------
ACCURACY
The accuracy is approximately 44 percent.
APPARATUS
The apparatus used during sampling, sample recovery, and sample analysis
are described below and illustrated in Figures A-l through A-4.
Apparatus for Sampling
Sampling cartridge - Tenax in 9.5 mm (0.375 in) outside diameter (o.d.)
by 7.9 mm (0.313 in) inside diameter (i.d.) by 300 mm (12 in) long
316 stainless steel tubes, glass wool plug on each end of Tenax, tube
sealed with Swagelok compression fitting on each end (Figure A-l).
Sampler - with auxiliary equipment, including:
Pump - capable of maintaining a minimum pressure ratio of 0.53 across
the critical-flow orifice at a flow rate of 5 cm /min
Critical-flow orifice (Lodge, et al, 1966) - capillary used to control
air flow rate at 5 cm /min
Air-flow meter - bubblemeter used for calibrating critical-flow orifice;
Indicating silica gel - to detect excess moisture flowing between flow-
meter and sample cartridge
Timers - 7-day and elapsed time to measure interval between runs and
time of sampling
Air filters - to provide clean air through critical-flow orifice to
prevent plugging and erroneous flow rate
Tubing - miscellaneous lengths of 316 stainless steel tubing, connectors,
and tees to adapt the sampler for single, duplicate, or tandem
sampling
Bags - plastic, for shipping and storing sample cartridges when not in
use (Ziploc)
Barometer - for measuring atmospheric pressure
Thermometer - for measuring ambient air temperature
Hygrometer for measuring relative humidity of ambient air
Freezer - capable of -18°C operation for storing samples
Apparatus for Sample Recovery
Desorption heater - 400-W, with temperature controller, designed to heat a
sample cartridge to 250°C in 10 minutes
Needle valve - 1.588 mm (0.0625 in) diameter, 316 stainless steel
Four-port valve - 1.5888 mm (0.0625 in) diameter, 316 stainless steel, high
temperature (150°C) type (Valco)
Tubing - various lengths of 1.588 mm (0.0625 in) o.d. by 1.016 mm (0.040 in)
i.d., stainless steel tubing
-------
SWAGELOK END CAP
9.5 mm
7.9 mm
300 mm
STEEL TUBING
O.D. x
I . D . x
LONG STAINLESS
— 25 mm GLASS WOOL
—250 mm TENAX
S
'25 mm GLASS WOOL
rrn
Figure A-1. Sampling cartridge
19
-------
FLOW
SAMPLING CARTRIDGE
HOUSING
FILTER
-CRITICAL-FLOW
ORIFICE
VACUUM
GAUGE
--7-DAY TIMER
VENT-
ELAPSED-TIME
TIMER
Figure A-2. Single sampler for benzene.
20
-------
FLOW
\
FILTER
-CRITICAL-FLOW
ORI F I CE —-"
£_
HOUSING
VACUUM
UAGE VENT
VACUUM
PUMP
JS r"i
•SAMPLING
CARTRIDGES
--7-DAY TIMER
ELAPSED-TIME
TIMER
Figure A-}. Duplicate sampler for benzene
21
-------
FLOW
FILTER-tJ
CRITICAL-
FLOW M.
'OR i F i cr
HOUSING
VACUUM
GAUGE
*\
I VACUUM
J PUMP
SAMPLING
CARTRIDGES
HS Hh
—(-7-DAY TIMER
i
I
ELAPSED-TIME
TIMER
Figure A-A. Tandem sampler for benzene.
22
-------
Apparatus for Sample Analysis
Gas chromatograph - with dual flame ionization detector (Perkin-Elmer 3920)
and with strip chart recorder (Linear, Model 385)
Data acquisition system - Hewlett-Packard model 3352, for analyzing the
output of GC and determining the benzene concentration
Six-port valve - 1.5888 mm (0.065 in) diameter, 316 stainless steel, high
temperature (150°C) type (Valco)
Tubing - various lengths of 1.5888 mm (0.0625 in) o.d. by 1.016 mm (0.040 in)
i .d. nickel tubing
Miscellaneous fittings - Swagelok compression fittings; quick connect
fasteners; and minimum-internal-volume unions of same nominal diameter
and type of material as the tubing
Chromatographic column - 0.5 mm (0.02 in) i.d. by 61 m (200 ft) stainless
steel tubing, MBMA m-bis with a stationary phase of (m-pheno-xyphenoxy)
benzene plus Apiezon L support-coated open tubular (SCOT), supplied
with temperature controller-programmer (Perkin-Elmer)
Trap - 1.588 mm (0.0625 in) o.d. by 1.016 mm (0.040 in) i.d. by 300 mm
long nickel tubing packed with 60/80 mesh silanized glass beads
(Liquid nitrogen surrounded the tubing during trapping. Oil at 190°C
was used to vaporize benzene from the trap.)
Flow controller --capacity to 10 cm /min (6 cm /min used)
Flow meter - 6 cm /min midrange capacity
Gas sampling loop -l.lm, gravimetrically calibrated, for cross-checking
benzene standards used to calibrate the GC
REAGENTS
All reagents used in this porcedure should be of a chromatographic
grade that conforms to the specifications established by the Committee
of Analytical Reagents, American Chemical Society. For a chemical not
covered by these specifications, the best available grade should be
used.
The following is a summary of the reagents used during sampling and
sample analysis.
Reagents for Sampling
Tenax - polymer resin adsorber for benzene
Glass wool - borosilicate
Silica gel - moisture indicator in flow meter during sampling
Reagents for Sample Analysis
Helium - zero-grade gas, for chromatographic carrier gas
Combustion air - containing less than 1.3 mg/m hydrocarbons (2 ppm
methane), for operating FID Benzene calibration standards - purchased
Benzene calibration standards - purchased cylinders (Scott Environmental
Research) containing 13, 72, and 350 ppm benzene (The first two
cylinders were used directly. The latter was used to prepare 2,
5, 41, and 200 ppm standards.)
23
-------
PROCEDURES
The following procedures were employed during sampling, and sample
recovery and analysis.
Sampling Procedure
The sampling procedure involved preparation of the sample cartridges,
special handling precautions, and sample collection procedures.
Sample Cartridge Preparation--
A stainless steel tube as previously described previously was cleaned
and packed atone end with glass wool to a depth of 2.5 cm. Cleaned Tenax
was placed on top of the glass wool to a depth of approximately 25 cm. The
top 2.5 cm of a tube was filled with glass wool to hold the Tenax in position.
Both ends of a tube were capped with Swage!ok fittings to complete the
assembly of a sample cartridge. Approximately 600 cartridges were prepared
and each was stamped with a unique serial number for identification.
The nuts on the compression fittings were tightened manually, then
turned 1.25 revolutions further with a wrench.
An assembled sample cartridge is illustrated in Figure A-l.
Special Handling Precautions--
The caps on the sample cartridges were maintained in place to prevent
contamination of the Tenax. During sampling the nuts were stored in Ziploc
plastic bags to decrease the possibility of contaminating the cap assembly
with ambient benzene and other pollutants.
Because Tenax is thermally unstable, and will decompose into other
compounds (including benzene), the cartridges were kept in cold storage
after preparation. The cartridges were stored at a temperature of -18°C in
freezers at the sites and at the analysis laboratory. During transport to
and from the sites, the cartridges were placed in Trans Temp Cool Paks,
insulated shipping carriers containing an eutectic mixture. Field operators
supervised shipment to the sites to ensure quality handling. One-day service
by Federal Express was used to carry the samples from the sites to the
analysis laboratory. Shipments occurred only on Mondays through Thursdays,
so that the time the samples were removed from the freezers was minimized.
To prevent condensation of water vapor, each cartridge was permitted to
equilibrate to room temperature for 8 hours.
Sample Collection Procedure
Three sampling systems were used, distinguished by the number of sample
cartridges and the method of connecting the cartridges to the samplers. The
systems, using either single, duplicate, or tandem samplers, are illustrated
in Figures A-2 through A-4.
24
-------
The majority of samples was obtained over a 24-hour period with the
single-cartridge sampler. The duplicate-cartridge sampler, with the car-
tridges connected in parallel, provided quality control data for assessing
the accuracy of the sampling data. The tandem cartridge sampler, with the
two cartridges connected in series, verified that benzene breakthrough did
not occur in the upstream cartridge.
Sampling occurred over a period of 10 consecutive days, and the daily
time period was delayed two hours between samples to prepare and clean the
equipment. Sampling occurred from 7 a.m. to 7 p.m. the first day, 9 a.m. to
9 p.m. the second day, and continued in a similar manner for the duration of
sampling. The duplicate quality-control samples were obtained during the
second, fifth, and eighth days from one site in each of the six cities that
were sampled.
The following data were recorded with each day's sample:
1. Location - city name and site number
2. Time - date sampling started and ended, time of day sampling started and
ended, elapsed time from timer
3. Meteorology - temperature (daily maximum and minimum), relative humidity,
barometric pressure, sky conditions (cloudy, clear, scattered showers,
heavy rain, snow, smog), wind (calm, steady, gusty)
4. Type sampler - single, duplicate, tandem
5. Serial number of each sample cartridge
6. Flow rates - three measurements were obtained at each of three times -
beginning, middle, and end of sampling (A total of nine measurements
was obtained for each cartridge. The average flow rates determined for
the three times were averaged to determine the average sample flow
rate.)
7. Total volume sampled - product of average sample flow rate and elapsed
time
8. Pump operating vacuum
Typical data recorded once for each site included:
1. Site location - building name, street address, city and state, general
location (central business district, central core, outer core, old
residential, inner suburb, outer suburb, rural)
2. Operating agency - name, street address, city and state, phone number,
name of operator :
3. Date - start and stop dates of sampling
4. Probe location - distance from curb, height above ground level, compass
direction of exposure, distance below top of building (if applicable)
5. Street type - freeway, highway, commercial, or residential street, and
number of lanes in each direction
6. Traffic condition - very heavy, heavy, medium, light, very light
7. Sampler data - manufacturer, model and serial numbers, type of pollu-
tant monitored, and probe size and type of material.
25
-------
Analytical Procedure
A sample was recovered by desorbing the benzene from the Tenax
resin, liquifying it in a cryogenic trap, and separating it from miscellaneous
organic compounds with a capillary column from which it was subsequently
analyzed. Figure A-5 illustrates the sample flow during recovery and analysis
A sample cartridge was removed from the storage freezer and placed in
th^ desorption heater (presented in Figure A-6). Quick-connect fasteners
enabled prompt fastening of the cartridge to valve ports (No. 3 and 4 of the
four-port valve) with the 400-W heater being preheated to a temperature of
250°C. With the two valves positioned as indicated for sample recovery,
helium carrier gas was permitted to flow at 30 cm /min through the cartridge
and into the trap containing 60/80 mesh silanized glass beads, where benzene
and other inorganic impurities were condensed. The sample was collected for
4 minutes. Also Curing this time the MBMA column was purged with helium at a
flow rate of 6 cm°/min. Immediately following recovery, the liquid nitrogen
surrounding the trap was removed and replaced by an oil bath at a temperature
of 190°C, which flash vaporized the benzene. After the trap had been in the
bath for 1 minute, the two valves were positioned as indicated in the sample-
analysis view (Figure A-5). The 6 cm /min flow of both helium and heat from
the oil bath caused benzene gas to flow from the MBMA column 8 minutes after
the valves were repositioned.
The benzene concentration was detected, measured, and recorded with a
FID, GC, a data acquisition system, and strip chart recorder.
The operating conditions of the various components and reagents are
summarized as follows:
3
Helium carrier gas - 6 cm /min, 50 gsig
Air - catalytically cleaned, 450 cm /min, 50 psig
Hydrogen - 54 cm /min, 20 psig
Temperatures: Oven - 70°C for 8 min, followed by 120°C
Detector - 200°C
Valves - 150°C
Transfer lines - 150°C
Oil bath - 190°C
Desorption chamber - 250°C
Liquid nitrogen - (-200°C)
CALIBRATION AND STANDARDS
Calibration mixtures were purchased from Scott Environmental Research
and used as working standards. The cylinder concentrations were verified
by preparing a series of benzene gas standards in gravimetrically calibrated
glass bulbs. These standards were injected into a gravimetrically cali-
brated gas sampling loop (1.1 ml). The glass bulb standard results showed
that the Scott working standards were well within experimental error. Some
of the tanks were also analyzed by other laboratories and the reported
concentrations reported well within experimental error (better than 10%).
26
-------
HELIUM
CARRIER
GAS
NEEDLE
VALVE
30 cmVmin
1 .588 mm O.D. x 1.016 mr I . D
STAINLESS STEEL TUBING
,6 cm'/min.
FLOW \2
CONTROLLERS^" US 3
DESORPTION HEATER 250°C
1 . 588 mm
LOW DEAD
VOLUME
UNION
SAMPLE CARTRIDGE
VALVES IN SAMPLE RECOVERY POSITION
61 m M. 6 . M . A .
S.C.O.T.
VALVES IN SAMPLE ANALYSIS POSITION
Figure A-5- Apparatus for sample recovery and analysis
27
-------
ALUMINUM SHROUD
200 W STRIP HEATER
SAMPLING CARTRIDGE
200 W STRIP HEATER
50 mm
O
Figure A-6. ^00~W modified desorption heater.
-------
The detector response was checked by diluting portions of a 350 ppm tank to
£, 5, 41, and 200 ppm. A calibration curve was plotted using the results of
these analyses; moreover, purchased mixtures of 13 and 72 ppm also were used
to check the calibration curve.
The stability of working standards was inferred by studies of a 9.37 ppm
calibration standard that was analyzed 36 times over a 2-month period.
Maximum and minimum concentrations from the mean value were 20.5 and 9.7
percent, respectively, and the standard deviation was 8.7 percent. To
obtain analyses accurate within 10 percent, a daily calibration was required.
CALCULATIONS
The volume of the air sample was not corrected to S.T.P. because of
uncertainty in the 24-hour average temperature and pressure values. The
air sample volume used in the analysis was determined as follows:
F, + F9 + F., ,.
V = -! - - - 2. x T x 10~6 (A-l)
m 3
where V = volume of gas sampled (uncorrected) , m
3
F, = measured flow rate before sampling, cm /min
o
F? = measured flow rate after sampling, cm /min
F- = measured flow rate during sampling, cm /min
T = sampling time, min
The GC and data acquisition system were calibrated in units.,of ppm,
which may be converted (Federal Register, 1971) to units of yg/m as follows:
yg/m3 = molecular weight x 106 (ppm) (A_2)
24,470
(ppm)
24,470
= 3192 (ppm)
where the molecular weight of benzene is 78.1 g/mole and air is dry at 25°C
and 760 mm Hg.
29
-------
REFERENCES
1. Federal Register, 1971. 35:8186-8201
2. Lodge, J. P.. J. B, Pate, B. E. Ammons, and G. A. Swanson. The Use of
Hypodermic Needles as Critical Orifice in Air Sampling. J. Air Pol
Control A, 16(4):197-200, 1966.
30
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA 600/4-80-027
3. RECIPIENT'S ACCESSION«NO.
4. TITLE AND SUBTITLE
Ambient Air Monitoring for Benzene 24-Hour Integrated
Sampling in Six Cities
5. REPORT DATF
May 1980
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
B. Martin, T. Clark, J. Bumgarner, Gary Evans
8. PERFORMING ORGANIZATION REPORT NO.
A2ALID (Toxics)
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Monitoring Systems Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Final (6/1/77-4/1/80)
14. SPONSORING AGENCY CODE
EPA 600/08
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This monitoring program was initiated to assess the population exposure to
ambient levels of benzene in conjunction with the announcement of benzene as a
hazardous material. The program was conducted during the summer and fall of 1977.
3
Average benzene concentrations of 5, 18, and 19 yg/m , respectively, were
measured in the widely separated cities of Dallas, Chicago, and Los Angeles. The
precision of these data is estimated to be ±26 percent of the measured values.
The improvement of benzene monitoring and analysis technology was an important
secondary objective of this program. However, further research is needed to identify
a more favorable collection medium than the Tenax that was used in this study.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Hazardous material
Air Quality Monitoring
Mobile Sources
Benzene
Tenax, GC/FID
Chicago
Dallas
Los Angeles
43F
68A
8. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (This Report)
Unclassifi ed
20. SECURITY CLASS (Thispage)
21. NO. OF PAGES
31
22. PRICE
EPA Form 2220-1 (9-73)
31
------- |