c/EPA
United States
Environmental Protection
Agency
Environmental Monitoring and Support EPA-600/4-79-059
Laboratory September 1 979
Research Triangle Park NC 27711
Research and Development
Monitoring System
for Collection and
Analyses of
Ambient Ethylene
Dichloride (EDC)
Levels in the Urban
Atmosphere
-------
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.
-------
MONITORING SYSTEM FOR COLLECTION AND ANALYSES OF
AMBIENT ETHYLENE DICHLORIDE (EDC) LEVELS
IN THE URBAN ATMOSPHERE
by
L. Elfers, G. Fusaro, and A. Khalifa
PEDCo Environmental, Inc.
Cincinnati, Ohio 45246
Contract No. 68-02-2722
Seymour Hochheiser
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL MONITORING AND SUPPORT 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 and Sup-
port Laboratory, U.S. Environmental Protection Agency, and approved for pub-
lication. Approval does not signify that the contents necessarily reflect
the views and policies of the U. S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
11
-------
ABSTRACT
A procedure was developed and tested to measure ambient levels of ethy-
lene dichloride (EDC). An activated charcoal tube was employed to collect a
24-hour integrated sample which was subsequently desorbed of ethylene di-
chloride using carbon disulfide. The carbon disulfide solution was then
analyzed for ethylene dichloride by gas chromatography (GC) separation com-
bined with detection by mass spectrometry (MS). Developmental methods in-
cluded the following steps:
• selection of gas chromatographic conditions and a detection
system for separation and quantification of EDC;
• determination of adsorption capacity of the charcoal tube
for EDC;
• evaluation of the desorption of EDC from adsorbents under dry
and wet conditions;
• evaluation of optimum sampling rates;
• determination of total method efficiency; and
• evaluation of the method under field conditions by use of
field study data.
This report was submitted in fulfillment of Contract No. 68-02-2722,
assignments No. 7 and No. 12 by PEDCo Environmental, Inc. under the sponsor-
ship of the Environmental Protection Agency. This report covers the period
August 1, 1978, to March 1, 1979, and work was completed as of May 1, 1979.
111
-------
FOREWORD
Measurement and monitoring research efforts are designed to anticipate
potential environmental problems, to support regulatory actions by develop-
ing an in-depth understanding of the nature and processes that impact health
and the ecology, to provide innovative means of monitoring compliance with
regulations and to evaluate the effectiveness of health and environmental
protection efforts through the monitoring of long-term trends. The Environ-
mental Monitoring Systems Laboratory, Research Triangle Park, North Carolina
has 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 Radiation, the
Office of Toxic Substances and the Office of Enforcement.
A system for measurement of ethylene dichloride in ambient air was
developed and evaluated for use in fence!ine monitoring. Field studies were
conducted and data was reported by use of this measurement system. Data on
ethylene dichloride in ambient air was requested by the Office of Air, Noise
and Radiation of EPA for their use in regulatory decision-making. This
report documents the measurement method used in these field studies.
/•?
Thomas
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgment . .viii
1. Introduction 1
2. Conclusions 3
3. Method Evaluation 4
4. Evaluation of EDC Method Under Field Conditions 20
References 32
Appendix
A. Tentative method for the determination of ethylene dichloride
in the atmosphere by 24-hoUr integrated sampling 33
-------
FIGURES
Number Page
1 Diagram of a 150-mg charcoal tube 1
2 Apparatus for adsorption tests at high humidity 9
3 Apparatus for adsorption tests at moderate humidity 16
4 Diagram of tandem charcoal tubes (with broken ends)
for EDC sampling 19
5 Standard deviation and arithmetic mean concentration
of EDC for collocated samples 23
6 Standard deviation and mean concentration of EDC for
duplicate analyses of selected samples ....24
7 Apparatus for generating and sampling gas mixtures of
EDC on charcoal tubes 30
A-l Diagram of tandem charcoal tubes for EDC sampling 35
A-2 Sketch of 24-hour integrated sampler for EDC
monitoring 39
VI
-------
TABLES
Number Page
1 Desorption Efficiency of EDC From Dry Charcoal 7
2 Desorption Efficiency of EDC From Humidified Charcoal 8
3 Evaluation of Charcoal and XE-340 11
4 Evaluation at a Sampling Rate of 500 cm3/min and 139
pg/m3 EDC at High Humidity 12
5 Evaluation at a Sampling Rate of 150 cm3/min and 126
yg/m3 EDC at High Humidity 12
6 Evaluation at a Sampling Rate of 65 cm3/min and 126
yg/m3 EDC at High Humidity 14
7 Evaluation of a Sampling Rate of 65 cm3/min and 14.6
yg/m3 EDC at High Humidity . 14
8 Evaluation at a Sampling Rate of 65 cm3/min and 348
yg/m3 EDC at High Humidity 15
9 Evaluation at a Sampling Rate of 65 cm /min and 2.5
yg/m3 EDC at High Humidity 15
10 Evaluation at a Sampling Rate of 65 cm /min and 2.7
yg/m3 EDC at Moderate Humidity 17
11 Evaluation at a Sampling Rate of 65 cm3/min and 36
yg/m3 EDC at Moderate Humidity . .17
12 Analyses of Unknown External Standard Solutions of
EDC 26
13 Quantity of EDC Found in EPA Quality Assurance
Samples ..... .28
14 Standard Deviation of Replicate Analyses for EDC found
in EPA Quality Assurance Samples. . .29
15 Quantity of EDC in PEDCo Quality Assurance Samples. ........ .29
16 Standard Deviation of PEDCo Quality Assurance Analysis. ...... .29
VII
-------
ACKNOWLEDGMENT
This report was prepared for the U.S. Environmental Protection Agency by
PEDCo Environmental, Inc., Cincinnati, Ohio. Mr. Lawrence Elfers was the PEDCo
Project Director.
Mr. Seymour Hochheiser was the Project Officer for the U.S. Environmental
Protection Agency. The authors appreciate the many contributions made to this
study by Messrs. R. Baumgardner, J. Bumgarner, G. Evans, P. Finkelstein, K.
Greer, T. Hartlage, J. Knoll, B. Martin, H. Sauls, J. Smith, L. Truppi, P.
Youngblood, and G. Wahl of the U. S. Environmental Protection Agency, Research
Triangle Park, North Carolina.
Vlll
-------
SECTION 1
INTRODUCTION
In 1964, Otterson and Guy described a technique for monitoring workplace
air for toxic organic vapors (1). They sampled the air with "homemade" tubes
filled with activated charcoal; removed the organics with a solvent; and anal-
yzed the resulting solution by gas chromatography. Since passage of the Oc-
cupational Safety and Health Act of 1970, researchers have studied the possi-
bility of using many solid sorbents in the sampling of toxic vapors. Work by
the National Institute for Occupational Safety and Health (NIOSH), in conjunc-
tion with private researchers and commercial companies, has led to the commer-
cial availability of charcoal tubes exhibiting only minor variability and which
are simple and inexpensive to use (Figure 1). The standard 7-cm tube contains
a 100-mg primary section and a 50-mg backup section to detect breakthrough.
~T
6mm
BACKUP
SECTION
20-40 MESH ACTIVATED
COCONUT CHARCOAL
Figure 1. Diagram of a 150-mg charcoal tube.
Both the U.S. Environmental Protection Agency (EPA) and the Occupational
Safety and Health Administration (OSHA) have recently emphasized the carcino-
genic potential of exposure to ethylene dichloride (EDC). As a result, consid-
erable work has been done on EDC sampling with charcoal tubes. Documentation
-------
of OSHA methods for EDC sampling (2,3) includes information on the behavior of
EDC on charcoal. Samples were generated to determine the breakthrough volume
and desorption efficiencies for EDC at various concentrations (0.5, 1, and 2
times the threshold limit of 100 ppm). Tests run at a sampling rate of 187
cm3/min with an EDC concentration of 821 mg/m3 (-200 ppm) for 2.5 h showed a
5 percent breakthrough of EDC to the backup section. NIOSH reports that the
desorption efficiency with carbon disulfide is approximately 95 percent. Like-
wise, the migration of sample through the tube is essentially zero.
Recent work by Saalwechter, et al. (4) demonstrated that sampling effi-
ciency on charcoal is reduced in the presence of high humidity because water
molecules take up available receptor sites.
Other techniques employing large stainless steel tubes packed with gas
chromatographic substrates such as Tenax have been used for collection of organic
materials. Tenax, however, has some shortcomings with respect to the collec-
tion of low-molecular-weight compounds and also with respect to water adsorption.
Tenax has been widely used in qualitative analyses in which thermal desorption
is followed by cryogenic concentration; gas chromatographic separation; and mass
spectrometer identification (15).
The purpose of this study was to develop from existing technology a method
for quantitating ambient EDC levels over a 24-h integrated sampling period.
-------
SECTION 2
CONCLUSIONS
A literature search resulted in the development of a monitoring system
which was subsequently subjected to a laboratory examination and a field moni-
toring program. Data from the latter study were then evaluated to determine
the applicability of the method.
The procedure was based on adsorbing EDC on charcoal at a sampling rate
o
of 65 cm°/min for a 24-h period. Analysis consisted of desorption of EDC from
charcoal, and its separation and detection by gas chromatography and mass spec-
trometry (GC-MS).
Laboratory evaluations indicated an overall method efficiency (accuracy)
of approximately 90 percent when tested within an EDC concentration range of
2.5 to 126 yg/m (0.6 to 31 ppb). The precision of the laboratory tests, based
on statistical analyses (relative standard deviation) of replicate standard
solutions, was 3 percent. No estimate can be made of its accuracy under field
conditions. Analyses of "blind" check samples of EDC on charcoal tubes indica-
ted a total recovery of 70 ±20 percent. Precision of the method under field
conditions based on statistical analyses of duplicate (collocated) samples was
6 percent.
-------
SECTION 3
METHOD EVALUATION
SELECTION OF EDC DETECTION METHOD
Gas chromatography was selected for quantitation of EDC because of its
selectivity, reproducibility, and relatively short analysis time requirement.
Various detectors were examined and compared to determine which would give the
best blend of selectivity, sensitivity, and reproducibility. Flame ionization
was quickly discarded because its lower limit of reproducible detection was
approximately 20 ng (injected on column). Electron capture detection offered
a tremendous increase in sensitivity (down to 50 pg injected), but interferences
from the desorbing solvent (carbon disulfide) precluded its use as the detector.
In evaluating mass spectrometry, EDC was detected by monitoring selected
masses that, although not unique to EDC, were unlikely to be affected by inter-
ference from other compounds. In monitoring masses at 62, 98, 100, and 102 m/e,
which are all characteristic of EDC, quantitation was performed with the cer-
tainty that this compound was EDC and not some other coeluting contaminant. In
addition, the multiple ion detection (MID) technique gave a sensitivity to
approximately 3 ng for column injection. This method presented the best blend
of sensitivity and selectivity and was therefore selected for the EDC analyses.
GC-MS Methodology
Selective ion monitoring (SIM) or multiple ion detection (MID) are two
names given the mass spectrometric technique employed to quantitate EDC in car-
bon disulfide solutions. It is based on the ionization of a compound to frag-
ments of certain masses unique to that compound. The abundance of these fragment
-------
masses is directly proportional to the amount of compound injected.
For EDC, four masses at 62, 98, 100, and 102 m/e were selected. The abun-
dance of mass 62 quantitates the amount of EDC present; the other three masses
are used to confirm that the peak measured is EDC. The mass spectrometer plots
a curve of the integration units obtained from mass 62 against time. The area
under this curve at the retention time expected for EDC is used in the quantita-
tion.
Analysis of several standards on our GC-MS (Hewlett-Packard model 5992-A)
indicated that responses to identical concentrations of injected EDC varied
from hour to hour, and more substantially from day to day. To alleviate this
problem, an internal standard was added to each sample before GC-MS analysis.
The compound 1-bromohexane was selected because of its favorable retention time
and its unlikely appearance in ambient air samples. An important fragment mass
of 1-bromohexane, 57 m/e, was monitored for each sample and the area was deter-
mined. The area obtained for EDC was divided by the area obtained for 1-bromo-
hexane to yield a response ratio. (Although the responses to EDC and 1-bromo-
hexane may change over time, the ratio is expected to remain constant.) The
response ratio obtained from standards was then used to determine the amount of
EDC in unknowns by use of the following equation:
= cv
EDC Rstd
where W = weight EDC, ng
EDC
_ area EDC peak
spl area 1-bromohexane peak (from sample)
area EDC peak
"" area 1-bromohexane peak (from blank)
_ area EDC peak
Rstd ~ area 1-bromohexane peak (from standard)
C = concentration of standard, ng/yl
V = volume of carbon disulfide and 1-bromohexane desorbing solution
used to desorb the charcoal tube; normally constant at 750 yl
-------
Gas Chromatographic Conditions
Separation of EDC was done on a nickel column 6.1 m long with an inner
diameter (i.d.) of 2.0 mm containing 10 percent SP 1000 on 80/100 mesh Supelco-
port (wt/wt). Operating conditions were selected so that EDC was separated
from carbon disulfide solvent before injection of the sample stream into the
mass spectrometer. Typical operating conditions for analyses of EDC by SIM
on a Hewlett-Packard 5992-A analyzer were as follows:
• helium — 0.45 atm
• injector temperature — 170°C
• oven temperature — 120°C isothermal
• solvent elution -MTim — 5.3 min
• run time — 12.5 min
• electron multiplier voltage — indicated by autotune
• ion masses for EDC — 62, 49, 98, 100, and 102 m/e
» ion masses for 1-bromohexane — 57, 85 m/e
• dwell times —750.0 ms for 62, 49, 57, 85 m/e; 500.0 ms for 98, 100,
and 102 m/e
• selective ion monitoring window sizes — 0.10 m/e
• amount of carbon disulfide injected — 4 yl
• retention time of EDC — 6.4 min
• retention tirae of 1-bromohexane — 9.0 min
INITIAL DETERMINATION OF ADSORPTION CAPACITY OF EDC ON CHARCOAL
NIOSH has demonstrated that the 100-mg front portion of the charcoal tube
will hold more than 23 mg of EDC with minimal breakthrough (less than 5%) (3).
Tests were conducted with an EDC concentration of approximately 200 ppm over a
period of 2.5 h. Because of this high capacity of charcoal for EDC, additional
confirmation tests were omitted.
-------
DETERMINATION OF EDC DESORPTION EFFICIENCY FROM CHARCOAL
Desorption of EDC from Dry Charcoal
Charcoal tubes from a supplier were injected with 41.3 yg EDC via a micro-
liter syringe and then refrigerated at 0°C for 24 h before desorption and anal-
ysis. Desorption consisted of removing the front and back sections of the tube
and placing them in separate 3-ml reaction vessels. The vials were sealed with
Teflon-coated serum liners and screw caps,- 0.5 ml of cold carbon disulfide was
injected into each vial; and the samples were placed in an ultrasonic ice bath
for 30 min. After desorption, samples were analyzed for EDC by GC-MS. Data
from this test series are presented in Table 1. Average desorption efficiency
was 94 ± a 6 percent. EDC was not detected in the back section of the tubes,
an indication that migration to the back did not occur.
TABLE 1. DESORPTION EFFICIENCY OF EDC FROM DRY CHARCOAL
EDC
injected, yg
41.
41.
41.
41.
41.
41.
41.
0
3
3
3
3
3
3
3
EDC
recovered, yg
35.
34.
40.
41.
38.
41.
39.
N.D
8
8
5
3
5
0
8
. *
Desorption
efficiency, %
86.
84.
98.
100.
93.
99.
96.
7
3
1
0
2
4
3
*
N.D. - nondetectable
Desorption of EDC from Humidified Charcoal
Charcoal tubes were humidified with an air stream of 99+ percent relative
humidity at a rate of 200 cm3/min for 24 h. After humidification the tubes
were injected with 41.3 yg EDC and desorbed. The results of this test (Table
-------
2) demonstrated that humidification of the charcoal slightly reduced the de-
sorption efficiency, which averaged 84 ± a 14 percent. From these desorption
efficiency tests, adsorption of EDC under dynamic conditions was evaluated.
TABLE 2. DESORPTION EFFICIENCY OF EDC FROM HUMIDIFIED CHARCOAL
EDC
injected, yg
41.3
41.3
41.3
41.3
41.3
41.3
41:3
41.3
0
EDC
recovered, yg
35.7
24.8
33.5
33.1
46.0
31.7
36,9
35.4
N.D.*
Desorption
efficiency, %
86.5
60.0
81.0
80.0
111.3
76.6
89.3
85.7
*
N.D. - nondetectable
EDC ADSORPTION EFFICIENCY TESTS
A device was fabricated to generate test atmospheres for the dynamic ad-
sorption tests (Figure 2). The system consisted of a regulated clean air
source (room air); bubblers to humidify the air stream; and apparatus to control
the temperature and moisture content of the gas stream. EDC under pressure (in
cylinders) was used, at concentrations certified by the supplier to ±2 percent.
EDC was then diluted with humidified air to achieve the concentration ranges
anticipated in ambient atmospheres. A water-jacketed manifold was employed
to distribute the gas mixture to the charcoal tubes. Sampling rates through
the tubes were controlled by means of critical-flow orifices and a vacuum
source activated by a timer. Actual test periods were recorded with a time
meter. Testing apparatus was operated in a small laboratory module held 2
to 3°C above the temperature of each test, to eliminate possible condensation
on the inlet to the water-jacketed manifold. The temperature of the test mix-
ture was measured with a total immersion thermometer placed inside the manifold.
8
-------
TIMER AND
TIME METER
VACUUM MANIFOLD
WITH CRITICAL ORIFICES
VALVE MASS FLOW
METER
0-50 cm3/min
STANDARD
GAS
IN
NITROGEN
PRESSURE
GAUGE
FLOW/PRESSURE
ADJUSTMENT
VALVE
ROOM
AIR -
INLET
ACTIVATED
CHARCOAL
COLUMNS
CONSTANT
TEMPERATURE
WATER BATH
30±I°C
CHARCOAL
TUBES
-THERMOMETER
. GAS
EXHAUST
WATER-JACKET
MANIFOLD
MASS
FLOW
METER
0-20 hter/min
GLASS
FIBER
FILTER
PREHUMIDIFIERS
MODIFIED STEM
SMITH GREENBURG
IMPINGERS
COARSE-
GLASS FRIT
BUBBLER
STEMS
-SMITH GREENBURG
IMPINGER
Figure 2. Apparatus for adsorption tests at high humidity.
-------
Preliminary tests were conducted to determine the degree of relative humi-
dity that the apparatus could generate. Gas was sampled from the manifold
through a known weight of magnesium perchlorate, and the weight of water col-
lected was determined. The relative humidity was calculated as 99+ percent.
Relative humidity tests were not routinely conducted during test runs, provided
that the standard compressed EDC gas mixture (prepared in dry nitrogen) was
maintained at less than 1 percent of the total flow through the system.
Evaluation of EDC on Charcoal and XE-340
Four separate tests were executed to determine whether carbonaceous sor-
bent XE-340, manufactured by Rohm and Haas, possessed better sorbent character-
istics than those of charcoal. Four sorption tubes were used in each test.
Dry Charcoal—
Charcoal tubes were charged with 41.3 ug EDC. After injection, tubes were
subjected to humidified air at a rate of 500 cm3/m±n for 24 h at a relative
humidity of 99+ percent. The average EDC recovery was 9.8 percent.
Dry XE-340—
The same procedure used on dry charcoal was performed on dry resin XE-340.
The average EDC recovery was 25.8 percent.
Wet Charcoal—
Charcoal tubes were humidified for 24 h at 99+ percent relative humidity
and a flow rate of 500 cm3/min. The tubes were injected with 41.3 pg EDC and
then subjected to humidified air at a rate of 500 cm3/min for 24 h. The aver-
age EDC recovery was 7.7 percent.
Wet XE-340—
The humidification test procedure was applied to prehumidified XE-340 tubes.
The average EDC recovery was 23.4 percent.
The data are summarized in Table 3. It was concluded that carbonaceous
sorbent XE-340 offers more desirable qualities as a sorption medium than charcoal.
10
-------
Tests were discontinued with XE-340 because the material was not commercially
available in sufficient quantities to continue the evaluation.
TABLE 3. EVALUATION OF CHARCOAL AND XE-340
Test condition
Dry
Wet
Dry
Wet
charcoal
charcoal
XE-340
XE-340
EDC
injected,
yg
41.
41.
41.
41.
3
3
3
3
EDC found,
Front
3.
2.
9.
7.
3 ±
1 ±
7 ±
6 ±
half
0.4
1.0
1.7
0.5
ug
Back half
1
1
1
2
.5 ±
.0 ±
.0 ±
.0 ±
0.7
0.1
0.5
0.1
EDC
recovered,
9
7
25
23
%
.8
.7
.8
.4
METHOD EVALUATION TESTS
Tests at High Humidity and Temperature
Several test series were conducted with the apparatus shown in Figure 2.
The sampling period was 24 h, relative humidity was 99+ percent, and gas tem-
perature was 30°C. Sampling rates and EDC concentrations were varied in order
to obtain data indicating the highest EDC recovery without sacrificing sensi-
tivity.
Evaluation of a 500 cm3/min Sampling Rate with 139 yg/m (34 ppb) EDC—
Charcoal tubes injected with EDC at a concentration of 139 yg/m3 were
placed on the apparatus, and sampling was conducted for 24 h. Samples were de-
sorbed and analyzed by GC-MS. The results presented in Table 4 show that ap-
proximately 84 percent of the EDC passed through both halves of the charcoal
tube. The average recovery was 16 percent; 6 percent was found on the back
half of the tube. Stemming from these results, lower sampling rates were in-
vestigated in later studies.
Evaluation of a 150-cm3/min Sampling Rate with 126 yg/m3 (30 ppb) EDC—
Tests were conducted at a sampling rate of 150 cm3/min and an EDC concen-
tration of 126 yg/m3. Average EDC recovery was 58 percent (Table 5). EDC was
detected in the back and front halves of the tube.
11
-------
TABLE 4. EVALUATION AT A SAMPLING RATE OF 500 cm3/min
AND 139 yg/m3 EDC AT HIGH HUMIDITY
EDC
sampled, yg
115.3
99.4
113.8
118.1
118.1
115.3
99.4
0
EDC
Front half
13.7
10.2
12.8
9.5
11.7
J.0.9
11.7
N.D.*
f ound , yg
Back half
7.1
6.2
6.5
4.8
6.3
7.1
7.2
N.D.
EDC
•recovered, %
18.0
17.0
16.9
12.0
15.2
15.6
19.0
*
N.D. - nondetectable
TABLE 5.
EVALUATION AT A SAMPLING RATE OF 150
AND 126 yg/m3 EDC AT HIGH HUMIDITY
cm3/min
EDC
sampled, yg
29.09
29.09
28.67
28.62
0
EDC
Front half
12.5
10.8
12.4
12.0
N.D.*
found, yg
Back half
4.5
4.8
4.8
5.1
N.D.
EDC
recovered, %
58.3
53.3
60.1
59.7
N.D. - nondetectable
12
-------
Evaluation of a 65 cm3/min Sampling Rate with 126 yg/m3 (30 ppb) EDC—
Additional tests were run at a sampling rate of 65 cm3/min and an EDC con-
centration of 126 yg/m3. Average EDC recovery was 92 percent (Table 6). Less
than 5 percent of the EDC was found in the back half of the charcoal tube.
Evaluation of a 65 cm3/min Sampling Rate with 14.6 yg/m3 (3.65 ppb) EDC—
In tests at a sampling rate of 65 cm3/min and an EDC concentration of 14.6
yg/m-', the average EDC recovery was 91 percent (Table 7). EDC was undetected
in the back half of the charcoal tube.
Evaluation of a 65 cm3/min Sampling Rate With 348 yg/m3 (186 ppb) EDC—
At a sampling rate of 65 cm3/min and an EDC concentration of 348 yg/m3
the average EDC recovery was 90 percent (Table 8). Approximately 8 percent of
EDC was detected on the back half of the charcoal tube. This breakthrough of
EDC could have resulted in losses beyond the back portion of the tube.
Evaluation of a 65 cm3/min Sampling Rate with 2.5 yg/m3 (0.6 ppb) EDC--
For sampling rates of 65 cm /min and EDC concentrations of 2.5 yg/m , the
average EDC recovery was 80 percent (Table 9). EDC was undetected on the back
half of the tube because any EDC present would have been below the GC-MS detec-
tion limits.
Tests at Moderate Humidity and Temperature
For tests at moderate humidity, a test atmosphere was generated in which
impingers previously employed to saturate the air with water vapor were re-
placed with a temperature equilibration coil of copper tubing (Figure 3). These
tests were performed at 25°C and a relative humidity of 64 percent (the prevail-
ing relative humidity of the air within the laboratory facility).
0 o
Evaluation of 65 cirr/min Sampling Rate with 2.7 yg/m° (0.6 ppb) EDC—
A test series was conducted for 24 h at a temperature of 25°C and relative
humidity of 64 percent. The front portion of the charcoal tube contained 79
percent of the EDC, the average recovery value. The back portion was analyzed,
but any EDC present would have been below the GC-MS detection limit. The data
are given in Table 10.
13
-------
TABLE 6. EVALUATION AT A SAMPLING RATE OF 65 cm3/min
AND 126 yg/m3 EDC AT HIGH HUMIDITY
EDC
sampled, yg
11.54
12.20
13.33
12.30
0
EDC
Front half
10.0
10.0
12.2
11.2
N.D.*
found, yg
Back half
0.5
N.D.
0.6
0.5
N.D.
EDC
recovered, %
90.9
86.0
96.0
95.0
*
N.D. - nondetectable
TABLE 7. EVALUATION OF A SAMPLING RATE OF 65 cm3/min
AND 14.6 yg/m3 EDC AT HIGH HUMIDITY
EDC
sampled, yg
1.56
1.35
1.38
1.45
1.60
1.39
0
EDC
found, yg
1.54
1.17
1.32
1.37
1.23
1.27
N.D.*
EDC
recovered, %
98.7
86.7
95.6
94.7
77.0
91.4
N.D. - nondetectable
14
-------
TABLE 8. EVALUATION AT A SAMPLING RATE OF 65 cm3/min
AND 348 yg/m3 EDC AT HIGH HUMIDITY
EDC EDC found, yg
sampled, yg Front half Back half
37.70 30.58 2.31
32.20 20.63 1.77
32.00 26.91 1.22
32.46 28.61 1.02
33.71 28.08 2.42
37.21 31.40 3.74
33.21 27.85 2.99
33.71 26.25 3.51
0 N.D.* N.D.
EDC
recovery, %
87
91
88
91
90
94
93
88
*
N.D. - nondetectable
TABLE 9. EVALUATION AT A SAMPLING RATE OF
AND 2.5 yg/m3 EDC AT HIGH HUMIDITY
65 cm3/min
EDC EDC
sampled, yg found, yg
0.24 0.19
0.23 0.20
0.25 0.21
0.23 0.17
0.23 0.18
0.24 0.21
0.24 0.19
0 N.D.*
EDC
recovered, %
79
84
82
75
76
84
80
*
N.D. - nondetectable
15
-------
CTi
FLOW/PRESSURE
ADJUSTMENT
VALVE
VACUUM MANIFOLD
WITH CRITICAL ORIFICES
•ROOM
AIR
INLET
NEEDLE
VALVE
MASS FLOW
METER
0-50cm3/min
STANDARD
GAS
IN
NITROGEN
TEMPERATURE
CONDITIONING
PRESSURE
GAUGE
MASS FLOW
METER
0-20 liter/mln
CONSTANT
TEMPERATURE
WATER BATH
25±
GLASS
FIBER
FILTER
CHARCOALTUBES
THERMOMETER
GAS
" EXHAUST
WATER-JACKETED
MANIFOLD
CARBON VANE
PUMP WITH
BYPASS
PRESSURE VALVE
ACTIVATED
CHARCOAL
COLUMNS
Figure 3. Apparatus for adsorption tests at moderate humidity.
-------
TABLE 10. EVALUATION AT A SAMPLING RATE OF 65 cm3/min
AND 2.7 yg/m3 EDC AT MODERATE HUMIDITY
EDC
sampled, yg
0.24
0.25
0.26
0.24
0.27
0.25
0.24
0
EDC EDC
found, yg recovery, %
0.18 77
0.22 85
0.22 82
0.18 73
0.20 75
0.20 81
0.20 80
N.D.*
*
N.D. - nondetectable
TABLE 11. EVALUATION
AND 36 yg/m3
AT A SAMPLING RATE OF 65 cm3/min
EDC AT MODERATE HUMIDITY
EDC
sampled, yg
3.37
3.47
3.63
3.27
3.21
3.32
3.37
0
EDC EDC
found, yg recovered, %
2.97 88
2.98 86
3.01 83
2.94 90
2.95 92
2.95 89
2.97 88
N.D. *
*
N.D. - nondetectable
17
-------
Evaluation of 65 cm3/min Sampling Rate with 36 yg/m3 (9 ppb) EDC—
At a sampling rate of 65 cm3/min, an EDC concentration of 36 yg/m3, and
64 percent humidity, the average EDC recovery was 88 percent (Table 11). No
EDC was detected on the back half of the tube.
SUMMARY OF LABORATORY EVALUATIONS
Dry charcoal was found to have a high affinity for EDC. However, the
presence of water markedly decreased this quality. Although the carbonaceous
sorbent XE-340 was found superior to charcoal for EDC collection, the limited
supply of this material terminated any further evaluation of this method.
Dynamic tests at various EDC levels under conditions of high temperatures
and humidity (30°C, 99+ relative humidity) for 24-h sampling periods revealed
significant EDC breakthroughs in the charcoal tubes for sampling rates above
65 cm3/min. These experiments yielded an overall average recovery above 90
percent for EDC levels ranging from 2.5 to 126 yg/m3 (0,6 and 31 ppb). For EDC
concentrations of 348 yg/m3 (86 ppb) , a breakthrough of approximately 10 per-
cent was observed.
As a result of these studies, a configuration employing two tubes was
designed to ensure against the EDC breakthrough which occurred during single
tube sampling. This tandem tube configuration (Figure 4) was incorporated in-
to the EDC Method described in Appendix A, and was used in all field monitor-
ing studies. The first tube was used as the front portion, and the second tube
was used as the back portion.
18
-------
FRONT TUBE
GLASSTUBE-\ FIBERGLASS^ ^
4tr 67m C 'Sfe^^SS-^l
\ \
BROKEN END \ \
OF Ti IRF PRIMARY \ BACKUP
ur ludc. SECTION \SECTIOh
\\
?n- an N/IF^H arTi\/ATFn— i
BACK TUBE
^NE G! ASS TURF^
M A FIBERGLASS^
\ /
3 C Wti^l
\ PRIMARY /
/ SECTION /
L TEFLON TUBING /
AVG. NO. 3 /
1— 0
J
n
URETHANE
^ — FOAM
fc\""'iM ~~^\ T0 VACUUM
?^*:"^_J) " SOURCE
/ BROKEN END
1 BACKUP OF TIJRF
SECTION Uh IUbt
n /lrl^fl^cu ArTix/A-rcrn
COCONUT CHARCOAL
COCONUT CHARCOAL
Figure 4. Diagram of tandem charcoal tubes (with broken ends) for EDC sampling.
-------
SECTION 4
EVALUATION OF EDC METHOD UNDER FIELD CONDITIONS
A study was conducted near the facilities of EDC producers and users to
determine ambient EDC levels (6). Sampling was performed over 24-h sampling
periods at 3 sites 10 times. Twelve samplers were used at each site. The
sampling was performed in duplicate at each site employing the tandem charcoal
tube configuration. All field samples were stored in the dark at 0°C until
analyses were performed. Samples were hand carried to the PEDCo laboratory,
where 50 to 70 percent of the samples collected daily were analyzed. Selection
of samples, including those from downwind and background samplers, was based
on the prevailing wind conditions during a 24-h sampling period. Ten percent
of the duplicate samples from the selected sites were analyzed. In addition,
replicate analyses were performed on 10 percent of all desorbed samples. In-
ternal standard solutions, check samples prepared on charcoal tubes, and blanks
were also analyzed. For this particulate study, the EDC method was followed
as described in Appendix A.
MATERIALS AND STANDARDS
Purity Cheek of Carbon Disulfide Solvent
The solvent used for desorption of EDC from the charcoal tubes was Fisher
Reagent Quality carbon disulfide (Lot No. C184). The solvent was analyzed by
GC-MS more than 30 times (several analyses of each opened bottle), and no EDC
was detected.
20
-------
Purity Check of Charcoal Tubes
More than 30 unused 150-mg charcoal tubes manufactured by the SKC Corpora-
tion (Pittsburgh, Pennsylvania: Lot No. 107) were analyzed by GC-MS during field
sample evaluation. No EDC was detected.
Purity Check of EDC Standard Reference Material
Ethylene dichloride standard solutions were prepared with EDC supplied by
Chem Services -(West Chester, Pennsylvania; Lot No. 0-642). The manufacturer's
quoted purity was 99+ percent. A GC-MS analysis of this material under the
identical conditions used during field sample analysis showed no contamina-
tion from other compounds. The manufacturer's assay of 99+ percent was there-
fore assumed accurate. No corrections were made with respect to the purity
of this reagent when used in preparing external standard solutions.
EDC BREAKTHROUGH IN SAMPLING OF HIGH LEVELS OF EDC
Although laboratory studies excluded breakthrough tests at extremely high
EDC levels, high levels were observed during one of the field studies. To de-
tect a breakthrough, PEDCo analyzed the second (rear) 150-mg charcoal tube for
the presence of EDC. Several sets of tubes were selected for analysis, and EDC
was detected in the backup tubes twice. Breakthrough occurred only in the EDC
range of 200 to 600 yg/m3 (60 to 180 ppb) and amounted to about 1 percent. Be-
cause breakthrough was undetected in all samples analyzed in that range, it
could have resulted from nonhomogeneous packing of the charcoal within some of
the tubes.
ADSORPTION OF EDC FROM THE AMBIENT ATMOSPHERE ON CHARCOAL TUBES UNDER STATIC
CONDITIONS
During the three field studies, about 60 charcoal tubes were exposed with
one end open to the ambient air for a period of 24 h. This test was to deter-
mine whether EDC migrated into a tube before or after dynamic sampling began.
Over 20 tubes selected at random from the 3 study sites were analyzed by GC-MS.
No EDC was detected.
21
-------
PRECISION UNDER FIELD CONDITIONS
Analyses of Duplicate (Collocated) Samples
All sampling was performed in duplicate. In about 10 percent of the cases,
both collocated samples were analyzed and the resultant data subjected to fur-
ther statistical analysis to establish the total sampling and analytical error.
Statistical analyses demonstrated that the standard deviation of EDC con-
centrations increased as EDC increased. The relationship between standard devia-
tion and mean EDC concentration is depicted in Figure 5.
Because of the apparent relationship between the standard deviation and
the mean, variability among concentrations (based on collocated sampling) was
expressed in terms of relative standard deviation (i.e., s/x [100]). Regression
analysis supported a relative standard deviation of about 6 percent. This vari-
ability measured the precision of both the sampling and analytical procedures.
Duplicate Analysis of Desorbed Field Samples
The relative standard deviation of 6 percent indicated the combined pre-
cision of sampling and analysis. To derive an isolated estimate of the preci-
sion of the analytical method, duplicate analyses were run on 10 percent of all
desorbed samples. The relationship between the mean and standard deviation is
depicted in Figure 6. As with the analysis of collocated samples, the standard
deviation of the duplicates was proportional to the average concentration. The
relative standard deviation (precision) of the analytical method was approximate-
ly 3 percent.
Summary of Precision Under Field Conditions
The precision of an individual EDC measurement is approximately 6 percent.
This estimate, based on analysis of ambient EDC measurements at the three study
sites, includes the precision of both the sampling technique and the analytical
method. Precision of the analytical method alone, derived from duplicate ana-
lyses of desorbed samples, is approximately 3 percent.
22
-------
y = 0.353 + 0.063x
r = 0.82
50 100 150
ARITHMETIC MEAN EDC CONCENTRATION, jug/m
200 250 300 350
3
400 450
500
Figure 5. Standard deviation and arithmetic mean concentration of EDC for collocated samples.
-------
20
ro
E
CT 15
O
10
Q
(T LiJ
< ^
Q <
Q
0
y = 0.085 -I- 0.030*
r =0,85
0 50 100 150
ARITHMETIC MEAN -DUPLICATE ANALYSES, jjg/m
200
3
250
Figure 6. Standard deviation and mean concentration of EDC for duplicate analyses of selected samples.
-------
DETERMINATION OF THE ACCURACY UNDER FIELD CONDITIONS
Analysis of Unknown External Solutions for Quality Control
The GC-MS analyses of field study samples were conducted by several chem-
ists over a period of 2 months. Each chemist prepared standards from a master
stock solution and analyzed a batch of 30 to 40 samples. Another chemist,
working independently, prepared an external check standard from an independent
stock solution of EDC and submitted it to the analyst. The data from this
quality control check on the working standard solutions are presented in Table
12.
To develop a relationship between the amount of EDC in the standard solu-
tion and that measured by the PEDCo analyst, PEDCo analyzed these data by re-
gression analysis. It was assumed that the precision of adding EDC to the so-
lution was substantially better than EDC determination by GC-MS analysis.
The relationship between the quantity of EDC determined by the GC-MS anal-
ysis and that of prepared standard solution can be expressed as:
y = 0.12 + 0.97x (Eq. 2)
where Y = ng/yl EDC determined by the GC-MS
x = ng/yl EDC in the standard solution
Thus, the recovery of EDC from standard solution by the GC-MS procedure
ranged from 109 percent at 1 ng/yl to 98 percent at 12 ng/yl.
Check Samples Prepared by EPA
The EPA provided 48 samples having known amounts of EDC ranging from 1.18
to 8.26 yg. These samples were prepared with a gravimetrically calibrated EDC
permeation tube and a dilution system. Each charcoal tube was charged with
EDC by sampling at a known flow rate of about 65 cm /min from the dilution
system. By varying the sampling time for each tube, the analyst was able to
calculate the amount of EDC adsorbed.
25
-------
TABLE 12. ANALYSES OF UNKNOWN EXTERNAL STANDARD SOLUTIONS OF EDC*
Date analyzed
10-19-78
10-20-78
10-20-78
10-20-78
10-23-78
10-23-78
10-24-78
10-24-78
10-24-78
10-26-78
10-31-78
11-10-78
11-13-78
11-13-78
11-14-78
11-15-78
11-16-78
11-17-78
11-28-78
11-29-78
11-30-78
EDC concentration
Added
1.25
1.43
5.04
3.67
6.25
6.25
1.58
1.58
6.25
3.13
2.60
0.78
0.65
12.43
6.28
3.26
4.96
2.71
6.94
3.49
3.14
, ng/yl
Found
1.50
1.44
5.03
4.40
5.71
5.83
1.69
1.68
6.22
3.26
2.64
0.89
0.50
12.33
6.78
3.21
4.67
2.67
6.65
3.53
3.10
Average 101.3 ± a 9.6
26
-------
Table 13 tabulates PEDCo's analysis (by GC-MS) of the 48 EPA check samples.
The quantities of EDC found by PEDCo were significantly less than those reported
by EPA. For example, in 9 EPA samples reported to contain 1.18 yg EDC, PEDCo
found an average of 0.81 yg; a recovery rate of only 69 percent. In 17 samples
reported to contain 5.90 yg EDC, PEDCo found an average of 4.64 yg; a recovery
rate of 79 percent.
Table 14 shows the standard deviation of replicate analyses for each EDC
level added by EPA to its check samples. It is apparent that the standard
deviation increased with increasing levels of EDC. Over the various EDC levels
added by EPA, the relative standard deviation ranged from 20 to 25 percent.
Check Samples Prepared by PEDCo
PEDCo prepared a set of 16 quality assurance samples, 8 with an EDC level
of 4.51 yg and 8 with 9.02 yg. The samples were prepared by using a standard
gas mixture of EDC certified by the manufacturer to ±2 percent accuracy and a
dilution system. Figure 7 is a diagram of this system, where the two series of
check samples were simultaneously prepared. Critical-flow orifices were select-
ed to provide a sampling rate of 65 ±1 cm /rain. An excess flow of EDC gas mix-
ture was established through the eight-port glass sampling manifold. Eight
tubes were connected to critical-flow orifices. Sampling was conducted for
15.0 min to produce low-level check samples and for 30.0 min to produce high-
level samples. The amount of EDC absorbed on each tube was then calculated
from the EDC concentration as assayed by the supplier and as derived from the
sampling rate and time period.
Table 15 presents PEDCo's analysis by GC-MS of these quality assurance
samples. For the eight samples prepared at an EDC level of 4.51 yg, the average
amount recovered was 4.45 yg a recovery rate of 99 percent. For the 8 samples
prepared at 9.02 yg, the average amount recovered was 8.61 yg a recovery rate
of 95 percent. Standard deviations (Table 16) were 0.30 yg and 0.27 yg for
4.51 yg and 9.02 yg EDC levels, respectively. There was no significant dif-
ference in the standard deviation for either level of EDC.
27
-------
TABLE 13. QUANTITY OF EDC FOUND IN EPA QUALITY ASSURANCE SAMPLES
1.18
Sample EDC
No. found, yg
C 0.53
G 0.66
J 0.89
22 0.89
25 0.64
26 1.05
27 0.95
31 0.99
33 0.65
Sample
No.
D
H
I
N
R
T
X
10
13
14
15
17
20
AMOUNT OF EDC
2.36 3.
EDC Sample
found, yg No.
1.37 A
1.27
1.50
1.13
2.07
1.31
1.68
0.94
1.76
1.11
2.14
1.95
1.84
ADDED BY EPA, yg
54
EDC Sample
found, yg No.
2.62 E
L
M
P
V
W
11
12
16
18
19
21
23
24
28
29
30
32
5.90 8.26
EDC Sample EDC
found, yg No. found, yg
5.45 B 4.47
4.67 F 6.75
4.84 K 6.07
3.56 0 7.49
7.47 Q 0.45*
3.89 U 6.64
4.14 Y 4.81
2.72
1.27*
6.83
4.90
2.56
4.86
4.85
4.13
4.70
4.51
4.79
Deleted - gross error
-------
TABLE 14. STANDARD DEVIATION OF REPLICATE ANALYSES FOR
EDC FOUND IN EPA QUALITY ASSURANCE SAMPLES
AMOUNT OF EDC ADDED BY EPA, yg
Mean
Std. dev.
Recovery, %
Relative std.
dev. , %
1.18
0.18
0.19
69.00
23.00
2.36
1.54
0.39
65.00
25.00
3.54 5.90
2.62 4.64
0 1.22
74.00 79.00
0 26.00
8.26
6.04
1.18
73.00
20.00
TABLE 15. QUANTITY OF EDC IN PEDCo QUALITY ASSURANCE SAMPLES
AMOUNT OF EDC ADDED BY
4.51
EDC found, yg
4.67
4.27
4.12
4.63
4.86
4.71
4.31
4.05
PEDCo, yg
9.02
EDC found, yg
8.29
8.83
8.31
8.72
8.88
8.65
8.28
8.91
TABLE 16. STANDARD DEVIATION OF PEDCo
QUALITY ASSURANCE ANALYSIS
AMOUNT OF EDC
4.51
Mean 4.45
Std. dev. 0.30
Recovery, % 99.00
Relative std. 6.7
ADDED BY PEDCo, yg
9.02
8.61
0.27
95.00
3.1
dev., %
29
-------
TIMER
AND
TIME METER
:^V? ^
0:
^^k. NEEDLE
C ) VALVE M/
r~T71
^
VACUUM
GAUGE VACUUM MANIFOLD
/^ Q) WITH CRITICAL ORIFICES
^UM ^TTTTYWT
PUMP
21 in Hg
<"'
N.
'-/
\SS FLO
METER
CHARCOAL
TUBES s
^~>i
] [
.] t
o c
-] c
]C
•3 C
<] [
^ ^ GAS MIXTURE
W GAS MIXING DISTRIBUTION
EXHAUST
AIR
t] 0-6
"-^ liter/min
EXCESS AIR F
] ^
/X^MX \y~^ ROTAMETER
i i [ ()-|o Mtpr/min
BULB MANIFOLD X^K
0-5 liter/min L ^
TANDARD
GAS
IN
NITROGEN
VACUUM
PUMP
i
!
i
ACTIVATED
CHARCOAL
COLUMN
Figure 7. Apparatus for generating and sampling gas mixtures of EDC on charcoal tubes.
-------
Summary of Accuracy Based on Analysis of Check Samples
Analytical results of the two sets of quality assurance samples differed
considerably. EDC recovery rates from prepared EPA samples were between 65
and 79 percent. Contrastingly, the rates from samples prepared by PEDCo were
95 and 99 percent. Furthermore, the standard deviation for EPA samples increas-
ed as the EDC level increased: from 0.19 yg at the 1.18 yg level to 1.18 ug
at the 8.26 yg level. Because the analytical methods were identical, this
discrepancy cannot be explained without further investigation.
Since the accuracy of the method cannot be precisely defined, one can only
state that the recovery of spiked samples, which is an indication of accuracy,
is in the range of 70 ±20 percent.
31
-------
REFERENCES
1. Otterson, E. J., and C. F. Guy. Trans. 26th Annual Meeting of the Amer-
ican Conference of Governmental Industrial Hygienists, Philadelphia,
Pennsylvania, 1964. p. 37.
2. National Institute for Occupational Safety and Health. NIOSH Method S
for the Determination of Ethylene Dichloride. NIOSH Manual of Analytical
Methods, Part II, 2nd Edition. DHEW Pub. No. 77-157-B. Department of
Health, Education, and Welfare, Cincinnati, Ohio, 1977.
3. National Institute for Occupational Safety and Health. Documentation of
the Niosh Validation Tests. NIOSH Pub. No. 77-185. Cincinnati, Ohio,
1977. p. S-122-1.
4. Saalwaechter, A. T. , C. S. McCammon, C. P. Roper, and K. S. Carlberg.
Performance Testing of the NIOSH Charcoal Tube Technique for the Deter-
mination of Air Concentrations or Organic Vapors. J. American Industrial
Hygiene Assoc., 38(9):476-486, 1977.
5. Pellizzari, E. D., J. E. Bunch, B. H. Carpenter, and E. Sawicki. Col-
lection and Analysis of Trace Organic Vapor Pollutants in Ambient Atmos-
pheres. Environmental Science and Technology, 9(6):556-560, 1975.
6. PEDCo Environmental, Inc. Monitoring of Ambient Levels of Ethylene Di-
chloride (EDC) in the Vicinity of EDC Production and User Facilities.
Submitted to EPA under Contract No. 68-02-2722, Task No. 8. Cincinnati,
Ohio, 1979.
32
-------
APPENDIX A
TENTATIVE METHOD FOR THE DETERMINATION OF
ETHYLENE DICHLORIDE IN THE ATMOSPHERE BY 24-HOUR INTEGRATED SAMPLING
This method has been drafted from available information, as presented in
the bibliography, and from laboratory and limited field evaluation. It is
still under investigation and is subject to revision.
PRINCIPLES OF THE METHOD
A known volume of air is drawn through a charcoal tube for a period of
24 h to trap the ethylene dichloride (EDC) vapors present. The charcoal in
the tube is tranferred to a small, stoppered sample container, where it is de-
sorbed with a solvent mixture of carbon disulfide and an internal standard com-
pound (1-bromohexane).* Subsequently, an aliquot of the desorbed sample is in-
jected into a gas chromatograph using a mass spectrometer as the detector. The
amount of EDC in the sample is then ascertained by determining the ratio of the
peak area for EDC to the peak area for 1-bromohexane and comparing that with
the ratio of peak areas obtained from a standard.
RANGE AND SENSITIVITY
The limit of detection is approximately 0.5 yg/m3 (0.13 ppb). The maxi-
>f the range is approximately 50C
diluting the sample after extraction.
mum of the range is approximately 500 ug/m3 (125 ppb); it may be increased by
* Warning: Because EDC is a suspected carcinogen, care must be taken to pro-
tect operators from breathing fumes. Carbon disulfide is toxic, and its
vapors form explosive mixtures with air; therefore, this material should be
handled in a we11-ventilated room with a fume hood.
33
-------
This method was evaluated over an EDC range of 2.5 to 348 yg/m3 (0.6 to
86 ppb), at temperatures of 25° and 30°C and relative humidities of 64 and 99+
percent, respectively. The sampling rate was 65 cm3/min at 760 mm Hg for 24 h.
The charcoal adsorption tube (Figure A-l) consisted of two sections of activated
charcoal (totalling 150 mg) separated by a section of urethane foam. Each tube
was backed with a second tube to determine breakthrough. The laboratory eval-
uations employed a single charcoal tube and yielded an average total method
efficiency (adsorption and desorption) of approximately 90 percent for EDC
levels of 2.5 to 126 yg/m3 (0.6 to 31 ppb). At concentrations of 348 yg/m3
(86 ppb), approximately 10 percent of the EDC was found on the back portion of
the tube. At an EDC concentration of approximately 2.5 yg/m3 (0.6 ppb), over-
all method efficiency was approximately 80 percent. The backup tube was anal-
yzed, but its EDC content remained unknown due to GC-MS detection limits.
Although no effect from humidity was observed during the laboratory eval-
uations at the 65 cm /min sampling rate, water condensation within the charcoal
tube during field sampling could have affected the adsorption efficiency. Sam-
pling for 24 h at rates in excess of 65 cm3/min affected the collection effi-
ciency. At an EDC concentration equivalent to 139 yg/m3 (34 ppb), tests using
a flow rate of 500 cm /min and a relative humidity of 99 percent indicated an
overall method efficiency of less than 20 percent. For atmospheres suspected
of possessing higher levels of EDC and humidity, sampling time should be short-
ened to less than 24 h.
INTERFERENCES
It is highly unlikely that any common pollutants are present in the ambi-
ent atmosphere in concentrations which would interfere with EDC measurements.
Several criteria must be met for identification and quantification. Retention
time and the specific mass ions of 49, 62, 98, and 100 m/e are simultaneously
monitored. These ions should be represented by the identical peak retention
time as EDC in order to provide positive EDC quantification. The data system
of the spectrometer uses the response of mass ion 62 for quantification.
34
-------
FRONT TUBE
BACK TUBE
^J 1-. ~f
GLASS TUBE-A URETHANE GLASS TUBE^
Y FIBERGLASS -7 /^-FOAM \ FIBERGLASS7
4mm 6mm (^ BJS^IS^
••a )) cc
BROKEN END \ \ /
OF TURF PRIMARY \ BACKUP / PRIMARY
ui- lUbt SECTION \SECTION / SECTION
^J
• i
URETHANE
^— FOAM
«^P-\v*^i
m-£t"£"a
/BACKUP
/ SECTION
— <5s IU VACUUM
_J) *" SOURCE
BROKEN END
OF TUBE
20-40 MESH ACTIVATED-^
COCONUT CHARCOAL
•TEFLON TUBING
AVG. NO. 3
-20-40 MESH ACTIVATED
COCONUT CHARCOAL
Figure A-l. Diagram of tandem charcoal tubes for EDC sampling.
-------
PRECISION AND ACCURACY
Replicate GC-MS analyses of standard liquid mixtures and sample aliquots
must not deviate by more than ±5 percent. Information is presently unavailable
on the accuracy of the total method. The precision of the analytical method,
based on statistical analysis (relative standard deviation) of replicate stan-
dard solutions, was determined as 3 percent. The precision of both the analy-
tical and sampling methods, based on statistical analyses of replicate samples,
was determined as 6 percent. Standard check samples prepared by EPA and by
PEDCo determined the accuracy of the analytical method. The EPA check samples
yielded an average recovery of 72 percent; the PEDCo samples, an average re-
covery of 97 percent. In an attempt to explain this discrepancy, the possibil-
ity of EDC decay on the tube between preparation and analysis was statistically
evaluated. No biases were observed. Since an origin for the discrepancy was
not ascertained, accuracy of the analytical method can only be estimated as be-
tween 72 and 97 percent.
ADVANTAGES AND DISADVANTAGES OF THE METHOD
The sampling device is small, portable, and uses no liquids. Interfer-
ences are minimal, and most are eliminated by altering chromatographic condi-
tions. The amount of sample that can be taken is limited by the sampling rate
and by tube capacity. When the sample value obtained for the backup tube ex-
ceeds 10 percent of that found on the first tube, the possibility of sample
loss exists.
A GC-MS system for separation and detection is more costly than a flame
ionization detection (FID) system. However, the GC-MS is the only readily
available system with the required sensitivity which also provides the maximum
criteria that must be met prior to integrating a chromatographic peak. Attempts
to use other detection systems (FID, electrolytic conductivity, and electron
capture) presented unavoidable problems.
36
-------
APPARATUS AND MATERIALS
Sample Collection Materials
pump — capable of maintaining an air pressure differential
greater than 0.5 atm at the desired flow rate
critical orifice — 30-gauge, 0.5-in hypodermic needle to
control flow rate at approximately 65 cm /min
filter cartridge — Disposable, 47 mm diameter, 0.45 ym filter
porosity (Millipore Filter Corp., Bedford, Massachusetts)
charcoal adsorption tubes — 150-mg, standard NIOSH type
connected in tandem (Figure A-l)
vacuum gauge — 0-to 760-mm Hg
airflow meter — rotameter type, 0 to 120 cm3/min, calibrated
against an NBS traceable bubble meter
Sample Recovery Materials
• muffle furnace — for operation at 250°C
• syringe — 0 to 1 ml, gastight with Teflon plunger
• vials — 2-ml capacity
• caps — screw type, with septum hole for 2-ml vials
• serum cap liners — Teflon-coated rubber, for sealing
vials and caps
• ultrasonic cleaner — 0.5- to 1-gal capacity (Bronson
Cleaning Equipment, Sheldon, Connecticut)
Analytical Equipment
gas chromatograph with mass spectrometer — Hewlett-Packard
5992-A or equivalent, with glass jet separator and data
system
chromatographic column — nickel, 6.1 m x 2 mm i.d., containing
10 percent SP 1000 on 80/100 Supelcoport
37
-------
• micro syringes — 0 to 10, 0 to 100, and 0 to 500 yl range
• vials (sample) — 5 ml and 50 ml capacity, with screw caps
and Teflon-lined serum cap liners
e pipettes — volumetric Class A, 50 ml, 5 ml, and 1 ml
Reagents
• chromatographic quality carbon disulfide
• 1,2-dichloroethane, reagent grade
• 1-bromohexane, reagent grade
• purified helium
PROCEDURE
Cleaning of Equipment
All glassware employed for laboratory analysis should be washed with de-
tergent, thoroughly rinsed with tapwater and distilled water, dried, and placed
in a muffle furnace at 350°C for 30 min to remove trace organic compounds.
Collection of Samples
A tandem arrangement consisting of two 150-mg charcoal tubes is used for
sampling; one is identified as the front tube and the other as the backup tube.
A 30-gauge, 0.5-in hypodermic needle is installed for critical-orifice flow
control. The ends from each tube are removed by breaking off the glass bead.
The two tubes are connected with Teflon tubing (5-mm i.d.), directly exposing
the large (100-mg) section of charcoal to the sampled air. Tubes are then con-
nected to the sampler,- and the pump is activated. Figure A-2 is a schematic
diagram of the sampler.
The operator records the initial vacuum, starting time, adsorption tube
number, and site location. The sampling rate is measured by connecting a rota-
meter (calibrated with an NBS traceable bubble meter) to the inlet of a tandem
38
-------
RAIN
SHIELD
FOUR SAMPLE
TUBES
HIGH-DENSITY
POLYPROPYLENE
TUBING-
CRITICAL FLOW
ORIFICES
5ft
THREE-WAY
SOLENOID
VALVE
Figure A-2. Sketch of 24-hour integrated sampler for EDC monitoring.
-------
charcoal sampling tube; initial flow is then recorded. A shield is placed over
the adsorption tubes to protect them from light and rain, and sampling is con-
tinued for 24 ±0.25 h. At the completion of the sampling period, the operator
records the time, final vacuum, and flow rate. Samples are then removed from
the sampler; plastic caps are placed on the adsorption tubes; the tubes are
wrapped in aluminum foil and stored in a freezer at < 0°C until laboratory anal-
yses are performed. One out of every 20 tubes used for field monitoring is re-
tained and returned to the laboratory as a blank.
Sample Recovery
Contents of the 150-mg charcoal tube are transferred into a clean 2-ml
vial. A Teflon serum cap liner is placed on the vial and secured with a screw
cap. By use of a 1-ml gastight syringe, 0.750 ml of cold (0°C) carbon disulfide
and 1-bromohexane mixture is injected through the septum into the vial contain-
ing charcoal. The vial is then placed in an ultrasonic cleaner containing ice
and water for 30 min. The sample containing charcoal in contact with carbon
disulfide is subsequently stored at < 0°C until GC separation and analyses are
performed.
Analyses
Column preconditioning—
Before initial use, the chromatographic column is heat-treated to remove
impurities. A flow of 20 to 30 ml/min of pure helium is established through
the column, and the temperature is raised at 2°C/min from ambient to 200°C.
This temperature is maintained for 40 h.
GC-MS conditions—
Typical operating conditions for the analyses of EDC with specific ion
monitoring on a Hewlett-Packard 5992-A analyzer are as follows:
• helium — 0.45 atm
» injector temperature — 170°C
40
-------
• oven temperature — 120°C isothermal
• solvent elution time — 5.3 min
• run time — 12.5 min
• electron multiplier voltage — indicated by autotune
• ion masses for EDC — 62, 49, 98, 100, and 102 m/e
• ion masses for 1-bromohexane — 57, 85 m/e
• dwell times — 750.0 ms for 62, 49, 57, 85 m/e; 500.0 ms for 98, 100,
and 102 m/e
• selective ion monitoring window sizes — 0.10 m/e
• amount of carbon disulfide injected — 4 yl
• retention time of EDC — 6.4 min
• retention time of 1-bromohexane — 9.0 min
Sample injection—
The first step in the analysis is injection of the sample into the gas
chromatograph. To eliminate difficulties arising from blowback or distillation
within the syringe needle, the solvent flush injection technique should be used.
A 10—yl syringe is flushed with solvent several times to moisten the barrel and
plunger, then 1 yl of pure CS£ is drawn into the syringe. The needle is re-
moved from the solvent, and the plunger is pulled back approximately 0.5 yl to
separate the solvent flush from the sample; a pocket of air is used as the
marker. The needle is then immersed in the sample, and a 4-yl aliquot is with-
drawn. The volume of the needle must be carefully selected, because the sample
will be completely injected. After the needle is removed from the sample and
before it is injected into the gas chromatograph, the plunger is pulled back
1 yl to minimize sample evaporation from the tip of the needle. The analyst
should observe that the sample occupies 3.9 to 4.0 yl in the barrel of the
syringe. When duplicate injections of a solution are made, no more than a 3
percent difference in area can be expected when the same syringe is used.
41
-------
Measurement of area—
The areas of the EDC and 1-bromohexane peaks are determined by an elec-
tronic integration system capable of determining the area of all ions monitored
at the same retention time. A printout of the area expressed in integration
units is obtained. The most predominant ion or base peak in the mass spectrum
of EDC, 62 m/e, is selected for quantitation. For 1-bromohexane, 85 m/e is
selected for quantitation; this ion is not the base peak, but it is less sub-
ject to interference from other compounds than the base peak.
STANDARDS CALIBRATION AND ANALYSES
Preparation of Desorbing Solution and Standards
Preparation of desorbing solution (weekly)—
The analyst adds 5.0 pi of 1-bromohexane to a 250-ml volumetric flask con-
taining 240 ml CS2 and brings the flask to volume with CS2- Concentration of
the 1-bromohexane solution is 23.5 ng/yl.
Preparation of stock standard solution (weekly)—
The analyst adds 5.0 yl of pure EDC to a 10-ml volumetric flask containing
9.8 ml of the desorbing solution and brings the flask to volume with the de-
sorbing solution. This yields an EDC concentration of 0.628 yg/yl.
Preparation of working standard procedure (daily)—
The analyst places 20 yl of the stock standard solution in a 5.0-ml vol-
umetric flask containing 4.8 ml of the desorbing solution and brings the flask
to volume with CS2/l-bromohexane desorbing solution. This produces an EDC con-
centration of 2.512 ng/yl.
Calibration of the GC-MS System
The GC-MS system is set up according to the conditions previously described.
Before use, the instrument is autotuned in accordance with the manufacturer's
operations manual. Four yl of the working standard are injected onto the GC
column and analyzed. A response ratio, based on the area obtained from the EDC
42
-------
peak (ion 62) and divided by the area obtained from the 1-bromohexane peak
(ion 85), is calculated and used to find the concentration of EDC in real sam-
ples. The response ratio is obtained by use of the following equation:
Al
response ratio = — = R (Eq. A-l)
A2
where A = area of the EDC ion in integration units
A = area of the 1-bromohexane ion in integration units
Since R is a ratio instead of an absolute area, variations between injections
in the amount of sample or the detector response will not affect its value.
The average response ratio is determined; no single response ratio must deviate
from the average by more than ±5 percent. If a response ratio exceeds this
limit, an error has occurred either during sample injection, in preparation
of the working standard, or from instrument malfunction.
Analyses of Samples
Four yl are withdrawn from the desorbed sample solution consisting of the
charcoal tube contents plus 750 yl of CS2/l-bromohexane eluting reagent. The
areas of 62 m/e ions from EDC and 85 m/e ions from 1-bromohexane are recorded.
An injection of 4 yl of a standard in the range of the samples analyzed is in-
serted into every tenth sample. The response ratio is calculated and compared
with the original calibration. An identically prepared blank tube is analyzed
in the same manner. Any area resulting from the specific ion of 62 m/e is re-
corded.
CALCULATIONS
Uncorrected Sample Volume
The volume of air sample is left uncorrected for standard temperature and
pressure because of the uncertainty associated with changes in average tempera-
ture and atmospheric pressure during 24-h sampling. The air sample volume taken
for analysis is determined as follows:
43
-------
Fl + Fp
V = -^=-r - - X T X 10 6 (Eq. A-2)
m 2.
n
where V = volume of gas sampled (uncorrected) , m
m
F = measured flow rate before sampling, ml/min
F = measured flow rate after sampling, ml/min
T = sampling time, min
Ethylene Dichloride Concentration
Calculation of the EDC collected on the adsorption tube —
From the integrated areas for ion masses 62 and 85, the EDC content col-
lected on the charcoal tube and corrected for the blank is calculated as fol-
lows :
Vc
std
where W = weight EDC, ng
EDC
area EDC ... . .
R , = - : — ; - - - (from sample)
spl area 1-bromohexane
area EDC ...
R,, = - : — : - : - (from blank)
blk area 1-bromohexane
area EDC ... ,
R J_, = - ; — : - : - (from standard)
std area 1-bromohexane
C = concentration of standard, ng/yl
V = volume of CS2 and 1-bromohexane desorbing solution mixture
used to desorb the charcoal tube (750 yl under normal condi
tions)
Calculation of EDC concentration*—
The concentration of EDC as yg/m3 in the sampled ambient air is calculated
as follows:
*The overall method efficiency was determined as 90 ±10 percent. A correction
factor of 1.1 may be used if desired. cpnr is multiplied by 1.1 to obtain the
concentration corrected for method efficiency. For levels in excess of 10 yg/
m , the backup adsorption tube is analyzed and the concentration added to the
results from the first adsorption tube. In this case, no correction factor is
recommended.
44
-------
"pric"* — ^
x 10 (Eq. A-4)
EDC Vm
where C = concentration of EDC in the ambient air sampled, yg/m3
EDC
W = weight of EDC, corrected for the blank, ng
EDC
Q
Vm = volume of air sampled under sampling conditions, in m3
The concentration of EDC may be calculated as parts per billion EDC:
ppb = yg/m3 x 0.247. (Eq. A-5)
EFFECTS OF STORAGE
Although little is known about the effects of storing charcoal tubes con-
taining adsorbed EDC, some evidence indicates that compounds with chemical char-
acteristics similar to EDC are adversely affected by strong sunlight and heat.
Tubes should be stored in a dark place and at a low temperature. Carbon disul-
fide/1-bromohexane solutions of EDC in the yg/m3 range are stable for at least
1 month if they are refrigerated in a sealed serum bottle with minimum head
space. Information is unavailable on the effects of storage of samples con-
taining other active substances commonly found in ambient air.
45
-------
BIBLIOGRAPHY
APPENDIX A
Gulf Science and Technology Company. 1977. Method for the Determination of
Benzene in Ambient Air (Integrated Sampling). American Petroleum Insti-
tute, Pittsburgh, Pennsylvania.
Levadie, B., and S. MacAskill. 1976. Analysis of Organic Solvents Taken
on Charcoal. Tube Samples by a Simplified Technique. Analytical Chem.
48(1):76.
Lodge, J. P., J. B. Pate, B. E. Ammons, and G. A. Swanson. 1966. The Use
of Hypodermic Needles as Critical Orifices in Air Sampling. J. Air
Pollution Control Assoc. 16(4):197-200.
National Institute for Occupational Safety and Health. 1977. Niosh Method
S for the Determination of Ethylene Dichloride. Niosh Manual of Analytical
Methods, Part II, 2nd Edition. DHEW Pub. No. 77-157B. Department of
Health, Education, and Welfare, Cincinnati, Ohio.
PEDCo Environmental, Inc. 1979. Monitoring System for the Collection and
Analyses of Ambient Levels of Benzene in Urban Atmosphere. Submitted
to EPA under Contract. No. 68-02-2722, Task No. 7. Cincinnati, Ohio.
PEDCo Environmental, Inc. 1978. Sampling and Analyses for Ethylene Dichlo-
ride in Ambient Air. Submitted to EPA under Contract No. 68-02-2722,
Task No. 8. Cincinnati, Ohio.
PEDCo Environmental, Inc. 1974. Vinyl Chloride Monitoring Near the B. F.
Goodrich Chemical Company in Louisville, Kentucky. Region IV, U. S.
Environmental Protection Agency, Sureillance and Analysis Division,
Athens, Georgia. Submitted to EPA under Contract. No. 68-02-1375,
Task No. 20. Cincinnati, Ohio.
Yasuda, S. K., and E. D. Lounghron. 1977. Air Sampling Methods for Tetra-
chloroethane and Other Related Chlorinated Hydrocarbons. J. Chromato-
graphy. 137 (2) :283-292.
46
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 . REPORT NO.
EPA 600/4-79-059
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
MONITORING SYSTEM FOR COLLECTION AND ANALYSES
OF AMBIENT ETHYLENE DICHLORIDE (EDC) LEVELS IN
THE URBAN ATMOSPHERE
September 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
L. Elfers, G. Fusaro, and A. Khalifa
PEDCo Environmenta, Inc., Cincinnati, Ohio
8. PERFORMING ORGANIZATION REPORT NO
PN 3320-L
9. PER
ERFQRMING QRGANIZATIQN NAME AND ADDRESS
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2722
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Environmental Monitoring and Support Lab
MD-75, Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA 600/08
15. SUPPLEMENTARY NOTES
Project Officer: Seymour Hochheiser
16. ABSTRACT
A method for the measurement of ambient levels of ethylene dichloride (EDC)
was developed and field tested. A 24-hour integrated sample is taken with an
activated charcoal tube, followed by desorption of the EDC with carbon disulfide.
The carbon disulfide solution is then analyzed for EDC by separation on a gas
chromatograph and detection with a mass spectrometer. Development of the method
included the following area steps:
- Selection of gas chromatographic conditions and a detection system
for separation and quantification of EDC.
- Determination of adsorption capacity of the charcoal tube for EDC.
- Evaluation of the desorption of EDC from adsorbents under dry and
wet conditions.
- Evaluation of optimum sampling rates.
- Determination of total method efficiency.
- Evaluation of the method under field conditions by use of data
from field studies.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Ethylene dichloride
Air pollution
Gas chromatography
Mass spectroscopy
Chloroethanes
Ambient monitoring
Chemical analysis, test
methods
43F
68A
19. SECURITY CLASS (This Report)
Unclassified
13. DISTRIBUTION STATEMENT
21. NO. OF PAGES
51
Unlimited
20. SECURITY CLASS (This pagej
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |