United States
Environmental Protection
Agency
Atmospheric Research and
Exposure Assessment Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S3-89/016Aug. 1989
&ERA Project Summary
Analytical Method
Evaluation for Measuring
Ethylene Oxide Emissions from
Commercial Dilute-Acid
Hydrolytic Control Units
Joette Steger
The Source Branch of the U.S.
Environmental Protection Agency's
(EPA) Environmental Monitoring
Systems Laboratory (EMSL) at
Research Triangle Park, North
Carolina, has a program to develop
stationary source test methods of
known precision and accuracy for
determining compliance with EPA
standards. While participating in this
program, Radian Corporation
performed a field evaluation of a
method for sampling and analyzing
ethylene oxide (EO) In the vent
stream from a dilute acid scrubber of
a commercial sterilizer. The purposes
of the test were to evaluate the ability
of the analytical method for
quantitatlng EO in the scrubber
emissions, to determine EO stability
in sample containers, and to measure
the EO concentration profile and EO-
to-oxygen and EO-to-
dlchlorodifluoromethane (CFC-12)
ratios during the initial sterilizer
evacuation.
The test method acquired three-
second grab samples that were
analyzed by gas chromatography with
flame lonization detection (GC/FID).
Samples were acquired at one-,
three-, four- or six-minute intervals.
All samples were analyzed in the field
and a portion of the samples were
reanalyzed in the laboratory.
The analytical method was
evaluated on three 1/8-inch (32
millimeters) OD stainless steel
columns packed with a liquid phase
coated on 60/80 mesh Carbopack B.
The 5% Fluorcol column was 10 feet
(3 meters) long; the 3% SP-1000
column was 8 feet long (2.4 m); and
the 1% SP-1000 column was 6 feet
(1.8 m) long. Evaluations were based
on column efficiency, resolution,
linearity, retention time stability, and
limits of detection and quantitatlon.
Sample stability was evaluated in
one-liter Tedlar bags, 5-mL gas-tight
syringes, and Vacu-Samplers. The EO
concentration profile was measured
by taking samples at one-minute
intervals during Test 1 Evacuation 1.
This test used 12 wt% EO and 88 wt%
CFC-12 as the sterilizing gas.
The full report contains
conclusions and recommendations
based on the field test results, a
description of the field test site, the
grap sampling and analytical method
used, results of the field evaluation,
test data, and references used to
prepare the report
This Project Summary was
developed by EPA's Atmospheric
Research and Exposure Assessment
Laboratory, Research Triangle Park,
NC, to announce key findings of the
research project that is fully
documented in a separate report of
-------
the same title (see Project Report
ordering information at back).
Introduction
Because EPA has listed EO as a
possible hazardous air pollutant, a
standardized analytical method is
needed to measure EO levels in a vent
stream and for determining control
equipment performance. The method
tested acquired three-second grab
samples that were analyzed by GC/FID.
The purposes of the field evaluation were
to evaluate the analytical method using
samples from an operating commercial
sterilizer with a dilute acid scrubber, to
determine EO stability in sample
containers as a function of storage time,
and to measure the EO concentration
profile and the EO-to-oxygen and
EO-to-CFC-12 ratios during the initial
sterilizer evacuation.
Procedures
Facility Description
The field evaluation was conducted at
Burron Medical, a medical supply
sterilization facility, located in Allentown,
Pennsylvania. The facility has three
1000-ft3 (28-m3) sterilizers, one uses
a 12/88 (wt%/wt%) EO and CFC-12 gas
mixture and two use 100% EO.
The exhaust from the sterilizers is
controlled by a DEOXX™" system. The
DEOXX system is a dilute acid scrubber
manufactured by Chemrox, that
hydrolyzes the EO to ethylene glycol. At
the time of the test the scrubber
contained dilute sulfuric acid and
approximately 10 wt% of ethylene glycol.
Each chamber is equipped with a total
recirculating oil-sealed pump. The
pumps are equipped with gas/liquid
separators that emit the gas to the
DEOXX system and recirculate the liquid
to the pump inlet.
The two different sterilization programs
used for testing are shown in Figures 1
and 2. Three tests were performed, one
with product in the chamber using 100%
EO and two without product using 12/88
gas. Data from all of these tests were
reduced and used to prepare this report.
Sampling Location
Samples were acquired after the
scrubber, at a point as close to the end
of the plant's stack as was feasible. The
stack was modified by installing
"Mention of trade names or commercial products
does not constitute endorsement or recom-
mendation for use
additional pipe to remove exhaust gases
from the sampling areas and orifice
plates to provide a small amount of back
pressure to facilitate sampling and
eliminate the possibility of diluting the
sample with ambient air.
Sampling Procedures
Grab sampling was used for
determining EO, CFC-12, and oxygen
concentrations. For EO and CFC-12
measurements, sample was withdrawn
into a heated, 1/4-inch (64-mm),
Teflon® line using a Teflon-lined
diaphragm pump. A stainless steel, 1/4-
mch (64-mm) tee was used after the
pump to provide a continuous purge of
the vent gas through the sampling
system. After the tee, on one branch, the
sample stream was routed through
Tygon tubing and exhausted downwind
from the sampling area. There was a
pinch clamp on the Tygon tubing to
block the exhaust line during sample
collection. On the other branch of the
tee, a stainless steel, 1/4-inch (64-
mm), toggle-operated shut-off valve
was placed prior to the sample container.
The EO and CFC-12 sampling line
was continually flushed with sample
throughout the test. Prior to attaching the
sample container, the line was flushed
with sample by closing the pinch clamp
on the Tygon exhaust line and opening
the toggle valve on the sample line. The
same procedure was used to fill the
sample container.
BACHARACH Fyrite oxygen indicator
was used to determine percent levels of
oxygen in the sterilizer exhaust at the
sampling location. The oxygen was
chemically absorbed from the sample by
chromous chloride. Accuracy of analysis
was ± 0.5 percent. Percent levels of
oxygen were measured every time a bag
sample was taken to indicate the dead
volume of the scrubber system and to
determine EO-to-oxygen ratios. An
aspirator bulb removed sample from the
stack downstream of the EO sampling
port.
Exit gas temperatures were measured
using a bimetallic temperature probe and
a pyrometer. Stack temperatures were
digitized by a calibrated pyrometer and
recorded every time a bag sample was
taken.
Analytical Procedures
The analytical method used for
measuring EO and CFC-12 was gas
chromatography with flame lonization
detection (GC/FID). A dual FID Varian
3400 GC was equipped with a 10-ft (3-
m) by 1/8-inch (32-mm) OD stainless
steel column containing 5% Fluorcol (
60/80 Carbopack B and a 6-ft (1.8-n
by 1/8-inch (32-mm) OD stainles
steel column containing 1% SP-1000 c
60/80 mesh Carbopack B. On July 9, tr
1% SP-1000 column was replaced wi
an 8-ft (2.4-m) by 1/8-inch (32-mn
OD stainless steel column containing 3'
SP-1000 on 60/80 mesh Carbopack I
The FID electrometers were connected I
Shimadzu CRI-A integrators.
The GC column oven was operate
isothermally at 45°, 55° and 65QC c
June 21, 22 and 23, respective!'
Laboratory reanalyses were performe
with an isothermal column temperature i
65°C except on July 9 when the 3'
SP-1000 column was used at 55°C. Tr
injector oven was heated to I75°C, ar
the detector oven to 200°C. Nitroge
carrier gas flow rates were 30 mL/mi
The FID support gas flow rate
recommended by the GC manufactun
were used. During Test 1 Evacuation
when both EO and CFC-12 wer
quantitated, the FID electrometers wei
programmed so that the range woul
shift from 10'10 or 10'11 amperes 1
10~9 amperes. One-milliliter sample
were injected using a 5-mL gas-tigl
syringe with an on/off valve.
Gas Chromatograph Calibrator
Both channels of the chromatograp
were calibrated for EO at the beginnin
and end of the day. On the first test da'
the GC/FID was also calibrated for CFC
12 to determine the EO-to-CFC-1
ratios. At least one standard was als
analyzed after every twentieth sampl(
Calibration curves consisted of
minimum of three standards the
bracketed the sample concentrations.
Test Results and Conclusions
Linearity
Based on a one mL injection volum<
the FID response using these thre
columns was linear between 31.25 an
100 vol% for CFC-12. Below 31.2
vol% CFC-12 linearity was nc
evaluated. Using a one-milliliter injectio
volume, the FID response when using th
3% SP-1000 column had the large:
linear range (0.5 to 100 vol% and 2.28
to 1,254 ppmv) for EO at the percei
level and at the part-per-million (pprr
level. At the part-per-thousand (pp
level, the FID response using the 1°
SP-1000 column had the largest lines
range (22.81 to 5.016 ppmv) for EC
Therefore, the FID response using th
1% SP-1000 column provides the be:
linear range for determining th
-------
5
4)
I
Chamber Program
Test Two— 12/88 Gas
40
80
Figure 1.
120 160 200
Elapsed time (minutes)
Chamber pressure during Test 2 using 12/88.
240
280
320
maximum EO concentration emitted from
a dilute-acid hydrolytic control unit.
Efficiency
Of the three columns evaluated, the
5% Fluorcol column operated at 65°C
provided the best efficiency for analyzing
percent levels of EO. For ppm levels of
EO the two columns were comparable.
The 3% SP-1000 column was clearly
the least efficient column at all tested
temperatures and concentrations.
Therefore, the Fluorcol column provides
the best efficiency for percent level EO
analyses and the 1% SP-1000 column
provides the best efficiency for ppm level
EO analyses.
Resolution
The 1% SP-1000 and the 5%
Fluorcol columns provided comparable
resolution of the EO from the CFC-12 at
the percent and ppt levels At column
temperatures of 45° and 55°C baseline
resolution of EO from CFC-12 was
obtained on both columns Both of these
columns provided significantly better
resolution of the two compounds than did
he 3% SP-1000 column. Therefore, the
J% SP-1000 column is not
recommended for analyzing samples of
12/88 gas. Resolution was not evaluated
at the ppm level.
Retention Time Stability
The EO retention times were most
stable on the 1% SP-1000 column
where they varied from 1.38 to 1.46
minutes (± 6%) at 55°C and were least
stable on the 5 Fluorcol column where
they varied from 2.21 to 5.43 minutes (±
150%) at 55°C. In general the EO
retention times varied least at column
temperatures of 65°C The CFC-12
retention times were most stable on the
3% SP-1000 column where they varied
from 1.46 to 1.56 minutes (± 7%) at
65°C and were least stable on the 5%
Fluorcol column where they varied from
4.72 to 557 minutes (± 18%) at 45°C.
Therefore, the Fluorcol column where
they varied from 4 72 to 5 57 minutes (±
18%) at 45°C. Therefore, the Fluorcol
column should only be used to analyze
for EO at the percent levels in situations
where the concentration is not expected
to change by more than two orders of
magnitude.
Limits of Detection and
Quantitation
The 3% SP-1000 column provided
the lowest detection and quantitation
limits, therefore, it would be the best
column for low-level and trace analyses.
The limits of detection and quantitation
were comparable for the 1% SP-1000
and 5% Fluorcol columns; therefore,
these two columns may not be
distinguished by these criteria.
Ethylene Oxide Stability
Evaluations were made of the stability
of EO in samples collected in 1-L Tedlar
bags, 5-mL gas-tight syringes, and
Vacu-Samplers. After 4 days in an
unpunctured Vacu-Sampler four out of
seven duplicate samples or 57% were
within ± 8% of the EO concentration in
the original samples. The seven duplicate
samples in the unpunctured Vacu-
Samplers lost an average of 15% of the
EO. Therefore, EO samples in
unpunctured Vacu-Samplers remained
stable for up to 4 days, with stability
defined as over 50% of the samples
being within ± 10% of the original
concentrations and the average of all EO
-------
Chamber Program
Test Three — 100% EO
Figure 2.
200
Elapsed time (minutes)
Chamber pressure during Test 3 using 100% EO.
400
concentration changes being within ± 15
percent.
After 4 or 5 days m gas-tight
syringes, only 2 out of 12 duplicate
samples or 17% were within ± 9% of
the EO concentration in the original
samples. The samples in the syringes
lost an average of 20% of the EO.
Therefore, samples are not stable in
syringes for 4 or 5 days.
Stability results are listed in Tables 1
and 2 for bag samples containing
exhausts from 12/88 and 100% EO tests,
respectively. In leak-free bags the EO
in 12/88 exhaust at EO concentrations
above 1,000 ppmv varied by an
averaged relative difference of + 9%
after 48 hours. After 5 to 6 days the bags
with concentrations of EO greater than
1,000 ppmv showed an average EO loss
of 22 percent. In the same time period,
bags with concentrations of EO less than
500 ppmv showed an average EO loss of
36 percent. After 17 to 18 days the bags
with EO concentrations greater than
1,000 ppmv showed an average EO loss
of 41 percent. Therefore, samples of
12/88 exhaust containing EO at
concentrations above 1,000 ppmv are
stable in Tedlar bags for up to 48 hours.
Leak-free bags containing 100% EO
exhaust at EO concentrations above
1,000 ppmv varied by an averaged
relative difference of + 3% after four
days. After 8 days these bags averaged
a 20% loss and after 12 days an average
of 33% of the EO was lost. In leak-free
bags of 100% EO exhaust at
concentrations below 500 ppmv an
average of approximately 20% of the EO
was lost after 4 days. No additional loss
was observed in these bags between
Day 4 and 12. Therefore, samples of
100% EO exhaust containing EO at
concentrations above 1,000 ppmv are
stable for at least 4 days.
Since the EO loss in the bags is
concentration dependent and more EO is
lost at lower concentrations, the EO loss
is expected to be caused by adsorption
of the EO onto the bag walls. Initially, an
equilibrium probably occurs between the
bag surface and the gas within the bag
accounting for the 2- to 4-day period
of stability observed As the gas remains
m the bag longer, the EO may be
penetrating further into the bag wall,
accounting for the additional EO loss
The sample stability was decreased by
the high surface-to-volume ratio m th
1-L bags.
Ethylene Oxide Concentration
Profile
During the first evacuation of the fir
test using 12/88 gas, the E
concentration ranged from 400 ppmv I
1,500 ppmv and reached a maximui
plateau between 13 and 20 minute
Concentrations of EO during I2/8
exhausts were approximately 1,200 ppm
during the second exhaust, 600 pprr
during the third exhaust, 300 ppm
during the fourth exhaust, 100 pprr
during the fifth exhaust, and 80 pprr
during the sixth exhaust. During the fir
evacuation of the third test using 100'
EO, the EO concentration ranged froi
500 ppmv to 3,000 ppmv Concentratior
of EO during 100% EO exhausts wei
approximately 600 ppmv during th
second exhaust, and 300 ppmv dunr
the third exhaust. Therefore, to measui
the maximum EO concentration emitte
samples should be acquired between 1
and 20 minutes after the start of the fir
chamber evacuation when 12'88 is used
Ratios of the EO-to-CFC-1
concentration were calculated in 12/f
-------
exhaust and a plot of the ratios versus
time was similar to a plot of EO
concentration versus time due to the high
and relatively constant CFC-12
concentration. Oxygen concentrations in
the 12/88 exhaust were monitored using
Fyrite and indicated that after five
minutes the exhaust contained minimal
air. Therefore, EO-to-CFC-12 ratios
could be established for determining the
maximum EO concentration emitted from
chambers using 12/88. The EO-to-
CFC-12 ratios would only be useful
during the first evacuation when the
expected CFC-12 concentration is 99.9
percent.
Recommendations
Based on the test results, Radian
makes the following 7 recommendations.
1 A 5% Fluorcol column at 65°C is
recommended for measuring per-
cent level EO and CFC-12 concen-
trations, levels expected at the
scrubber inlet or m the exhaust of
uncontrolled sterilizers. The FID
response when using a 5% Fluorcol
column is linear, as indicated by an
F-test for linearity from 0.5 to 50
vol% EO. The 5% Fluorcol column is
the most efficient column tested for
percent levels of EO, and provides
baseline resolution of the CFC-12
and EO at column temperatures of
45 and 55°C and near baseline
resolution at 65°C. Also, the
retention time instability on the
Fluorcol is less pronounced at
percent concentrations.
2 A 1% SP-1000 column at 45°C is
recommended for measuring ppm
levels of EO, levels expected at the
scrubber outlet. The FID response
when using a 1% SP-1000 column
is linear from 22.81 to 5,016 ppmv
EO. The 1% SP-1000 column is
the most efficient column tested for
ppm levels of EO, and provides
baseline resolution of the CFC-12
and EO at 450C.
3. It is recommended that a 3% SP-
1000 column at 45°C be investigated
for quantitating sub-ppm levels of
EO, levels expected in ambient air at
sterilization facilities. The 3% SP-
1000 column provided limits of
detection and quantitation almost
three times lower than the other two
tested columns.
4 It is recommended that grab
samples collected m gas-tight
syringes be analyzed within several
hours of collection, samples col-
lected in bags within 48 hours of
collection and samples collected in
Vacu-Samplers within 4 days of
collection. Syringe samples lost an
average of 20% of the EO after 4
days. In leak-free bags, EO
remained stable for at least 48 hours
and may be stable for up to 4 days.
After 5 or 6 days, bag samples lost
an average of one-third of the EO.
In unpunctured Vacu-Samplers,
57% of the samples were within ±
8% of the expected concentration
after 4 days and the average EO
loss for all samples was 15 percent.
5. Sampling at least three times
between 13 and 20 minutes after the
first indication of flow through the
stack is recommended for
determining maximum EO
concentration emitted during the first
evacuation. The EO concentration in
the exhaust increased linearly
between 5 and 13 minutes, pla-
teaued between 13 and 20 minutes,
and dropped off after 20 minutes.
This time frame corresponds to
chamber pressures between 16 and
10 pounds per square inch actual
and may need to be adjusted when
using a different sterilization
program.
6. If the second and following
evacuations are monitored, sampling
between 5 and 10 minutes during
these evacuations is recommended.
During the first 5 minutes, the gas
coming out of the stack is a mixture
of gas from the chamber and
residual gas from the scrubber so
samples taken during this time may
be biased high or low.
7 To determine EO-to-CFC-12
ratios during the first evacuation, a
0.1-mL sample volume is recom-
mended for CFC-12 quantitation. A
1-mL sample of percent levels of
CFC-12 overloads the analytical
column, decreasing the column per-
formance.
The full report was submitted in partial
fulfillment of EPA Contract No. 68-02-
4119 by Radian Corporation under the
sponsorship of the U.S. Environmental
Protection Agency. The report covers the
period from March to August of 1988.
-------
Table 1.
Sample
Bag No.
13
20
22
25
30
32
36
37
39
41
44
45
48
50
55
56
57
Ethylene
June 21
1230
1340
1200
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Oxide Stability in Tedlar Bags Containing 12/88
Concentration (ppmv)
June 22
1370
NA
1360
976
374
133
28.5
36.7
29.8
1170
NA
1240
NA
305
NA
NA
NA
June 23
1320
NA
1320
NA
NA
96.5
NA
NA
27.2
NA
NA
1380
NA
340
80.6
101
83.5
June 27
995
1010
956
734
NA
34.2
NA
NA
1.08
NA
1120
958
462
214
57.4
53.6
51.8
July 1
824
830
828
562
NA
34.4
61.8
NA
11.9
636
954
865
370
195
75.2
70.5
65.8
Exhaust
Julys
716
1420
790
962
166
NA
48.8
124
NA
546
803
690
556
173
71.0
70.6
48.4
July 9s
733
985
760
536
NA
NA
65.6
NA
NA
538
810
683
NA
NA
NA
NA
NA
a Samples analyzed on a 3% SP-1000 column
NA = Not available
-------
Table 2. Ethylene Oxide Stability in Tedlar Bags Containing 100% EO Exhaust
Concentration (ppmv)
Sample
Bag No.
58
60
63
64b
65
66
67
68
69
70
71
72
June 23
343
2,840
544
342
344
334
430
426
387
1,530
1,440
1.200
June 27
NA
2,410
470
170
238
230
370
352
282
1,760
1,490
1.270
July 1
270
2,030
426
13.7
258
268
322
307
270
1,400
1,080
974
JulyS
248
1,660
430
NA
242
244
303
516
226
1.200
902
756
July 9s
2128
1,408.0
364.8
NA
220.8
225.6
2840
NA
NA
NA
NA
NA
a Samples analyzed on a 3% SP-1000 column
b Bag leaked
NA = Not available
-------
Joette Steger is with Radian Corporation, Research Triangle Park, NC 27709
John H. Margeson is the EPA Project Officer (see below).
The complete report, entitled "Analytical Method Evaluation for Measuring
Ethylene Oxide Emissions from Commercial Dilute-Acid Hydrolytic Control
Units," (Order No. PB 89-155 253/AS; Cost: $21.95, subject to change) will
be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Atmospheric Research and Exposure Assessment Laboratory,
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S3-89/016
.,_,
------- |