EPA/600/R-00/066
September, 2000
INTEGRITY OF VGA-VIAL SEALS
Brian A. Schumacher
U.S. Environmental Protection Agency
National Exposure Research Laboratory
Characterization and Monitoring Branch
P.O. Box 93478
Las Vegas, NV 89193-3478
and
Martha M. Minnich, John H. Zimmerman, and J. Blasdell
Lockheed Martin Environmental Services
980 Kelly Johnson Drive
Las Vegas, Nevada 89119
Contract Number 68-C5-0091
NATIONAL EXPOSURE RESEARCH LABORATORY
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193-3478
-------
NOTICE
The U.S. Environmental Protection Agency through its Office of Research and
Development funded and managed the research described here under contract 68-C5-0091 to
Lockheed-Martin Environmental Services. It has been subjected to the Agency's peer and
administrative review and has been approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute endorsement or recommendation for use.
-11-
-------
Abstract
Preservation of soil samples for the analysis of volatile organic compounds (VOCs)
requires both the inhibition of VOC degradation and the restriction of vapor movement in or out
of the sample container. The control of VOC vapor movement is generally assured through the
use of screw caps with polytetrafiuoroethylene (PTFE) faced silicone septa that can be tightened
to form a vapor lock within the volatile organic analysis (VGA) vial. The U.S. Environmental
Protection Agency's (EPA) Region 7 laboratory expressed concern that visual imperfections in
the glass lips and threads of VGA vials might allow VOCs to escape during storage. The
objectives of this study were to determine if these imperfections lead to VOC losses and to
identify an inexpensive screening test that could be used to distinguish between "defective" and
competent vials.
Clear, 40-mL glass VOA vials manufactured by the four major U.S. glass manufacturers
were tested for seal integrity. The vials were purchased precleaned, with the manufacturer's or
distributor's choice of PTFE faced silicone septa. All 216 vials from each manufacturer (864
vials, total) were visually inspected and then vacuum-tested for seal integrity by evacuating the
vials, glass lips down against a polished aluminum plate.
Visual inspection revealed a variety of imperfections ranging from small indentations,
bumps, and scratches on vial threads or lips, through obvious defects, such as large indentations
or grooves in the vial lips and chipped or broken glass. Imperfections were found on the lips or
threads of 4% to 15% of the vials depending upon manufacturer.
The aluminum plate vacuum test proved to be unreliable in identifying potentially leaky
vials. No vials formed a complete seal regardless of the presence of visual imperfections.
However, from each set of manufacturer's vials, the ten vials with the highest vacuum readings
and the ten vials with the lowest vacuum readings (80 vials, total) while the pump was running
were selected for two more tests, a septa-sealed vacuum test and a VOC-loss test.
The septa-seal vacuum test was conducted twice on the 80 selected vials. No clear
conclusion could be drawn about whether the flexibility of the silicone septa is sufficient enough
to form a complete seal against VOC losses. For one manufacturer, the test was capable of
identifying vials that would leak while for the other manufacturers, the test failed. The septa-seal
vacuum test appears to be subject to a noticeable rate of false positives.
Mean VOC concentrations after 14 days storage generally were within ± 20% of the
known concentration with a majority of the concentrations within ± 15% of their known values.
There were no statistically significant differences in VOC concentrations between vials in the
potentially leaky and control group for any of the manufacturers. Only 1 vial lost VOCs and that
was due to a large chip in the vial's lip and neck. These findings indicate that the silicone septa
are flexible enough to overcome most vial imperfections and form a complete seal against VOC
loss. A careful inspection of the VOA vials prior to use to remove any vials with large and
obvious imperfections should be sufficient to screen out vials that are subject to VOC losses.
-111-
-------
Contents
Page
Notice ii
Abstract iii
List of Figures v
List of Tables vi
Acronyms vii
Acknowledgments viii
1. Introduction 1
2. Materials 2
2.1 Vial Selection 2
2.2 Analytical Instrumentation 2
3. Procedures 3
3.1 Visual Inspection 3
3.2 Plate Vacuum Test 3
3.3 Septa-Sealed Vacuum Test 5
3.4 VOC Loss Test 5
3.5 VOC Gain Test 6
4. Results 6
4.1 Visual Inspection 6
4.2 Plate Vacuum Test 6
4.3 Septa-Sealed Vacuum Test 11
4.4 VOC Loss Test 14
4.5 VOC Gain Test 14
5. Summary and Conclusions 15
References 18
Appendix A - Visual Inspection Log and Plate Vacuum Test Results by Vial A-l
Appendix B - Individual VOA Vial Analytical Results by Manufacturer B-l
Appendix C - Quality Assurance/Quality Control Report for the Vacuum Studies C-l
Appendix D - Quality Assurance/Quality Control Report for the VOC Loss Test D-l
-IV-
-------
List of Figures
Figure 1. Aluminum plate vacuum test apparatus 4
-v-
-------
List of Tables
Table 1. Results of Plate Vacuum Test for Selected Vials 7
Table 2. Septa-Sealed Vacuum Test Results 12
Table 3. Summary of VOC-Loss Test Results 15
-VI-
-------
Acronyms
EPA United States Environmental Protection Agency
HP Hewlett Packard
PTFE polytetrafluoroethylene
RSD relative standard deviation
VOA volatile organic analysis
VOC volatile organic compound
-Vll-
-------
Acknowledgments
The authors would like to thank Mr. Don Miller of the EPA's Region 7 laboratory for his
input and guidance throughout the initiation and running of this Regional Methods Initiative
project.
-Vlll-
-------
Introduction
Preservation of soil samples for the analysis of volatile organic compounds (VOCs)
requires both the inhibition of VOC degradation and the restriction of vapor movement in or out
of the sample container. VOC degradation is generally controlled by chilling the sample to 4° C
(or in special cases, -20° C) and/or adding a preservative to the sample. For soils with VOC
concentrations expected to be between 0.5 and 200 i-ig/kg, the addition of sodium bisulfate to
reduce the sample's pH to < 2 is required (EPA, 1996). However, the sampler is cautioned that
for soils containing carbonate minerals, the addition of sodium bisulfate is inappropriate due to
sample effervescence effectively purging the sample prior to analysis. In contrast, for soils with
high VOC concentrations (VOC concentrations > 200 |_ig/kg), the addition of methanol is
required.
Where the use of preservatives in the field is impractical or undesirable, a soil sample
may be collected and temporarily stored in hermetically sealed samplers, such as the EnCore™
or SoilCore™ discrete samplers. Once the soil is sealed in the sampler, the sampler is chilled to
at least 4° C for transport to the analytical laboratory. The collected soil sample is then
transferred to a volatile organic analysis (VOA) vial and preserved (as appropriate) as soon as
possible or, at least, analyzed within 48 hours of collection.
The control of VOC vapor movement is generally assured through the use of
polypropylene screw caps with polytetrafiuoroethylene (PTFE) faced silicone septa that can be
tightened sufficiently enough to form a vapor lock within the VOA vial. Leaky VOA vial seals
allow volatile contaminants to either escape or enter a sample; thereby, resulting in erroneous
data. Sample integrity can be compromised by vapor losses or cross contamination that can
occur during sample collection, transport, and storage. The current EPA method for the
preparation of VOC samples, SW 846 Method 5035 (EPA, 1996), requires that VOA vials
remain hermetically sealed until analysis by a purge-and-trap instrument capable of purging the
sample by puncturing the septa of the open-top 40-mL VOA vial. If the samples remain
hermetically sealed, the potential routes of sample vapor loss or contamination are limited to
leaky seals or diffusion of VOCs through the septa.
Poor vial sealing can occur due to: (a) the presence of sand grains (or other particles) on
the lip or threads of the VOA vials; (b) septa expansion and contraction during temperature
fluctuations associated with sample cooling/freezing or sample warming and cooling cycles
during shipment from the field to the laboratory; or (c) puncturing the septa. The single-most
important factor in sealing soil samples for VOC analysis is to insure that the lip and threads of
the vial are clean before sealing. Traces of soil or grit on the glass lip will compromise the seals.
Hewitt et al. (1995) intentionally left soil on the lip of three 135-mL (4 oz) bottles and thoroughly
cleaned three others. Water (125 mL), spiked with a methanol solution containing four VOCs,
was added to each of the six bottles and analyzed after 5 days storage at 4° C. Samples stored
with dirty closures had an average of 41% to 46% lower VOC concentrations than the samples
stored in the bottles where the threads and lips were cleaned prior to sealing.
-------
Another potential cause of poor vial sealing is suspected to be slight aberrations in the lip
and threads of the glass VOA vials. Quick preliminary visual examination of numerous VOA
vials revealed that imperfections (e.g., indentations, grooves, and bumps) in the vial lips exist
and may be large enough to provide a pathway for VOC loss. However, it is unclear whether the
flexibility of the septa is sufficient enough to control the losses potentially associated with the
imperfections in the VOA vials. Therefore, the study objectives were to: (1) identify the type
and extent of imperfections in the lips or threads of VOA vials made by different glass
manufacturers, (2) determine whether the imperfections lead to a loss of VOCs or if the septa
creates a vapor tight VOA vial seal, and (3) determine if an inexpensive screening test could be
used to distinguish between "defective" and competent vials.
Materials and Methods
Vial Selection
Multiple VOA vial distributors (e.g. Fisher Scientific, Cole Farmer, I-Chem, Qorpak,
Wheaton, etc.) were contacted to determine if they manufactured their vials or if they assembled
the pieces purchased from different manufacturers. Most distributors indicated that they perform
the following before marketing their products: (1) select a vial manufacturer and screen the glass
vials to meet minimum quality control specifications, (2) select septa from any of a large number
of laminators, (3) select a screw cap for the vial, and (4) perform precleaning steps, as necessary.
As a result of the calls, it appears that four glass manufacturers supply nearly all the U.S.
distributors with the 40-mL glass VOA vials. These manufacturers are: Chase Scientific Glass
(Rockwood, TN), Comar Glass (Baltimore, MD), Kimble Glass Inc (Vineland, NJ), and Wheaton
(Millville, NJ).
The study was originally designed to include vials obtained directly from the major glass
manufacturers. However, only one manufacturer offered precleaned vials. Therefore, the
remaining three manufacturer's glass vials were ordered precleaned from distributors (i.e., Eagle-
Pitcher, I-Chem, and Qorpak). A total of 864 vials (216 vials from each of the four major glass
manufacturers) were examined. All vials were clear borosilicate glass and came with the
manufacturer's or distributor's selection of open-top screw caps, and 3.05 mm silicone, 0.127
mm PTFE faced septa.
Analytical Instrumentation
VOCs were introduced by closed-system purge-and-trap (SW-846 Method 5035) into a
gas chromatograph with mass spectrometer detector (MSB) following SW-846 Method 8260A
except calibration was only for the analytes of interest (EPA, 1996). The Varian Archon™
purge-and-trap autosampler in conjunction with a Tekmar 3000 sample concentrator and a
Vocarb 3000 trap was used to extract the samples. All samples were analyzed on a Hewlett-
Packard (HP) Model 5890 Series II gas chromatograph with a HP Model 5970 mass spectrometer
and a 60-m, 0.25-mm i.d. RTX Volatilization/RTX 5022 fused silica capillary column. The
-------
MSB was scanned from 35 to 300 m/z at 70 eV in the El mode. The spectroscopic signal was
analyzed using HP Chemstation software.
Procedures
Three types of tests, visual inspection, vacuum tests, and VOC measurements were used
to assess the integrity of the vial seals. If a correlation between the visual or vacuum tests and
the VOC measurements was observed, then one of the simple visual or vacuum tests may be
offered as a screening method for a priori insurance that the vials to be used will seal properly.
Visual Inspection
All 216 vials from each manufacturer were uncapped and the glass lip and threads
inspected for any defects. Slight imperfections consisted of: indentations or, conversely, bumps
on the lip or threads; variability in the smoothness of the lip or threads; or variations in the glass
thickness near the lip or threads. Obvious defects included chipped or broken glass, or clearly
noticeable imperfections of the types described above. Vials were tracked by the serial number
affixed to the vials, or numbered 1 through 216 if the vials did not have manufacturer assigned
serial numbers (i.e., bar codes).
Plate Vacuum Test
A fiat plate vacuum test was conducted on all 216 vials from each of the manufacturers to
determine if the vial lip imperfections prevented the formation of a vacuum tight seal. Vials
were placed such that the vial lip was in direct contact with a polished aluminum plate (Fig. 1).
In the center of the plate, a small hole was drilled and brass swagelok fittings were used to
connect the open hole to a 0.25 horse-power vacuum/pressure pump (Gast Manufacturing
Corporation, Benton Harbor, Ml). The swagelok nut on the top had the hex points ground nearly
smooth to allow the 40-mL VOA vial (21.74 mm neck id) to fit directly over the nut. A grade
AA accuracy vacuum gauge (Marsh PG-73, KW Instruments, Ontario CA) was connected
between the vial and a valve used to isolate the gauge and vial from the pump. No grease or
other sealant was used in this test.
Vials were placed on the plate one at a time. The vacuum pump was turned on and a
maximum vacuum reading was obtained in 30 seconds. While the vacuum pump was turned on,
the readings remained stable. If the vial was isolated from the pump, the vacuum dissipated
within 1 to 5 seconds indicating an incomplete seal between the aluminum plate and VOA vial.
Therefore, the vacuum attained for each vial after 30 seconds was recorded for the test result.
After conducting this test, data were ranked (lowest to highest vacuum obtained) for each
set of vials. The 10 vials with the lowest vacuum readings from the plate vacuum test and 10
vials with the highest vacuum readings were selected as "potentially leaky" and control vials,
respectively, for each manufacturer. These vials were designated by manufacturer (A through D)
-------
Figure 1. Aluminum plate vacuum test apparatus.
-------
and numbered 1 through 10 for the vials with the lowest vacuum readings and 20 through 29 for
the vials with the highest vacuum readings.
Septa-Sealed Vacuum Test
In the second vacuum test, vials were evacuated with the septa and open-top screw caps
fastened in place. The screw caps were hand tightened until a dimple formed in the center of the
septum. Two needles were inserted into the septum, the first needle was connected to the
vacuum gauge and the second needle was clamped to a hose connected to the vacuum pump.
Sturdy 18-gauge needles were used in this test. Smaller gauge needles tended to core the septa
and bend too easily, causing more problems than the heavy gauge needles. The silicone appeared
to seal after puncturing, leaving a visible puncture mark in the PTFE only.
The needles were inserted approximately 2 mm from and on opposite sides of the center
point. Vials were evacuated to the capacity of the vacuum pump (i.e., 18620 ± 380 mm Hg).
The evacuation needle was then removed leaving the vacuum gauge needle in place. The
vacuum was noted and the needle was removed. After 3 days and 7 days, the vacuum remaining
in each of the vials was measured. Puncture points for subsequent vacuum measurements were
located approximately 2 mm off the center point and 90° from the prior insertion points.
VOC Loss Test
To monitor VOC losses from the sealed vials, VOC-spiked water was added to each of
the 80 test vials. A bulk aqueous solution containing 50 ng/mL of each of the following analytes
was prepared: acetone, benzene, chlorobenzene, 1,1-dichloroethene, methylene chloride,
toluene, and trichloroethene. The spiked water was acidified with sulfuric acid to a pH < 2 and
transferred to a Tedlar bag. Approximately 5 mL of VOC-spiked water was added to each of the
test vials by gravity flow through PTFE tubing. Each vial was weighed before and after the
addition of the spiked water to calculate the spike addition to individual vials.
Vials were sealed, stored upright at 4° C, and VOC concentrations were measured after
14 days. Vials from the four manufacturers were prepared on four separate days to allow
adequate time for the analysis. Duplicate samples were prepared by washing the vials after the
first test was completed (water rinse, methanol rinse, and 105° C oven dry for 4 hours) and
repeating the same spiking/storage/analysis procedure using a new septum for each vial. Internal
standards (pentafiuorobenzene, 1,4-difiuorobenzene, chlorobenzene-d5, and 1,4-dichlorobenzene-
d4) and system monitoring compounds (l,2-dichloroethane-d4, toluene-dg, and
bromofiuorobenzene) were added by the Archon™ purge-and-trap autosampler just prior to
analysis.
-------
VOC Gain Test
To monitor VOC gains from external contamination sources through the septa of sealed
vials, 5 mL of deionized water were added to each of the 80 test VOA vials. The vials were
capped with new, unpierced septa and placed in a 1 gallon paint can. An open vial containing 5
mL of methylene chloride was placed in the center of the paint can prior to closure. Samples
were stored at room temperature for 14 days and then analyzed following SW-846 Methods 5035
and 8260A.
Results
Visual Inspection
Imperfections were observed in 15 out of 216 vials from manufacturer A, 8 out of 216
vials from manufacturer B, 32 out of 216 from manufacturer C, and 13 out of 216 vials from
manufacturer D (Appendix A). The imperfections noted for manufacturer A vials were minor
indentations along the lip, which could not be felt while running one's finger around the lip, but
could be seen when the vials were held up to the light and rotated. Two of the eight imperfect
vials from manufacturer B were observed to have somewhat larger indentations on the lip and
one vial had a chip in the glass threads. The remaining five vials from manufacturer B had minor
indentations along the lip. Imperfections in the manufacturer C vials included: seven vials with
dips (i.e., large indentations or grooves) in the glass or uneven lips (i.e., where one side of the lip
was obviously lower than the other side); one vial with a chipped thread; one vial with a small
bump on the lip; and the remainder of the vials had indentations that could be seen when held up
to the light. One vial from manufacturer D had a major defect, a chip in the lip that extended
down the side of the neck. One vial had a crack in its neck while minor chips in the neck were
observed on six other manufacturer D vials. The remaining five vials had chips on their lips.
Plate Vacuum Test
The vacuum system was capable obtaining a vacuum of 18620 ±380 mm of Hg.
Vacuum readings obtained in the vials after 30 seconds with the pump running ranged from 3800
to 19000 mm Hg (Appendix A). However, when the valve to the vacuum pump was turned off,
the vacuum dissipated within 5 seconds. None of the vials sealed completely against the
aluminum plate. Generally, the larger the visual imperfections, the worse the vial performed on
the plate vacuum test. Very small or minor imperfections tended not to effect the results of the
plate vacuum test.
Vacuum readings for the "potentially leaky" vials ranged from 3800 to 13680 mm Hg
(Table 1). In contrast, in the control vials, vacuum readings were consistently greater than 16720
mm Hg and were as high as a full vacuum reading of 18620 mm Hg. Interestingly, while no
visible imperfections were identified for any of the control vials, both imperfect and visually
-------
Table 1. Results of Plate Vacuum Test for Selected Vials.
Vial ID
Al
A2
A3
A4
A5
A6
A7
A8
A9
A10
A20
A21
A22
A23
A24
A25
A26
A27
A28
A29
Serial No.
5979
6031
4523
5971
4530
6060
5929
6047
4564
6013
5988
6025
4515
4567
6057
4471
4536
6061
4563
4568
Visual
defects
Y
Y
Y
Y
Y
Y
Y
N
N
Y
N
N
N
N
N
N
N
N
N
N
Plate Vacuum
(mm Hg)
5624
6840
7068
7220
7448
8208
8740
8740
9576
10336
18620
18620
18468
18468
18392
18316
18316
18316
18316
18240
-------
Table 1 . Results
Vial ID
Bl
B2
B3
B4
B5
B6
B7
B8
B9
BIO
B20
B21
B22
B23
B24
B25
B26
B27
B28
B29
of Plate Vacuum
Serial No.
29200
29214
30547
29232
29163
29205
29228
29223
29189
30584
29166
30530
30568
29220
30630
29196
30586
29235
30525
30508
Test for
Visual
defects
Y
Y
N
N
N
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Selected Vials (cont).
Plate Vacuum
(mm Hg)
5168
7144
10260
10640
10640
10868
11400
11476
11476
11628
18468
18468
18392
18240
18164
18164
17860
17860
17860
17860
-------
Table 1 . Results
Vial ID
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CIO
C20
C21
C22
C23
C24
C25
C26
C27
C28
C29
of Plate Vacuum
Serial No.
3291
5016
3982
4071
4351
4254
4399
3893
4583
3337
5025
4570
4143
4777
4360
3365
4490
3541
4623
4266
Test for
Visual
defects
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
N
N
N
Selected Vials (cont).
Plate Vacuum
(mm Hg)
3800
5320
5320
5548
5700
5776
6688
8208
9576
8576
17708
17480
17480
17480
17328
17328
17328
17328
17328
17328
-------
Table 1 . Results
Vial ID
Dl
D2
D3
D4
D5
D6
D7
D8
D9
D10
D20
D21
D22
D23
D24
D25
D26
D27
D28
D29
of Plate Vacuum
Serial No.
80
117
5
183
213
38
161
16
48
202
212
156
159
165
167
67
17
7
14
32
Test for
Visual
defects
Y
Y
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Selected Vials (cont).
Plate Vacuum
(mm Hg)
5928
9728
12388
12388
12920
13148
13148
13680
13680
13680
17860
17708
17708
17480
17480
17480
17480
17480
17328
17328
10
-------
imperfection-free vials were found in the potentially leaky vials. In general, the imperfection-
free vials in the potentially leaky group held the greater vacuum during the test.
Septa-Sealed Vacuum Test
The septa-sealed vacuum test was conducted twice to confirm the results. A loss of 3800
to 4560 mm Hg occurred each time the vacuum gauge was inserted into a vial caused by the dead
air space in the vacuum gauge and needle. Vacuum readings that dropped by more than 4560
mm Hg by day 3 or more than 9120 mm Hg by day 7 were considered to be leaky. Vials that had
readings of < 7600 mm Hg were considered to have catastrophically failed to seal. Vials that
successfully held their vacuum after 3 and 7 days in at least one test were considered to be
capable of forming a complete seal (i.e., the silicone septa was flexible enough to fill any
potential leaks caused by the vial imperfections).
Two vials from both the potentially leaky (vial numbers 1-10) and control groups (vial
numbers 20 - 29) from manufacturer A were leaky after day 3 during the first test (Table 2).
Three of those vials catastrophically failed to seal (i.e., vials A9, A21, and A24). After 7 days, 2
other vials, A5 and A28 had catastrophic seal failures. Upon retesting the vials, vial A24 again
catastrophically failed and vials A21 and A28 successfully maintained its seal after 3 days but
failed to hold a vacuum after 7 days.
All the potentially leaky vials and 6 control vials made by manufacturer B were leaky,
with one catastrophic leak in vial B3, during the first septa-sealed vacuum test (Table 2). Upon
retesting, all of the potentially leaky vials and 3 of the originally identified leaky vials in the
control group (i.e., vials B20, B21, and B28) remained leaky.
Similar to the vials from manufacturer B, vials from manufacturer C gave very poor
results on the first septa-sealed vacuum test with only one vial not being classified as leaky or
having catastrophically failed after 3 or 7 days. It was noticed that the septa on these vials were
more flexible than the septa on other vials. These septa would flex during puncturing by the
needle gauge, causing leakage if the puncture was near the center. In the second test, all septa
were punctured approximately 2-mm off the center point. During the retest, only 5 vials were
deemed leaky and all those vials were from the potentially leaky group.
The three vials from manufacturer D that had visual defects were leaky or failed
catastrophically during the first septa-sealed vacuum test. During retesting, all three vials again
either failed catastrophically (i.e., vials Dl and D3) or were leaky (i.e., vial D2) after 3 days and
catastrophically failed after 7 days. Vial Dl had a major chip in the lip and was expected to leak.
The glass lip of vial D3 was slightly raised on one side, but this vial had maintained a partial
vacuum during the plate vacuum test, reaching a maximum vacuum of 12388 mm Hg (Table 1).
Only one of the control group vials failed catastrophically during the first test; however, the same
vial (i.e., D26) maintained its vacuum seal during the second septa-sealed vacuum test.
11
-------
Table 2. Septa-sealed vacuum test results
Vacuum (mm Hg)
Vial ID
Al
A2
A3
A4
A5
A6
A7
A8
A9
A10
A20
A21
A22
A23
A24
A25
A26
A27
A28
A29
Bl
B2
B3
B4
B5
B6
B7
B8
B9
BIO
B20
B21
B22
B23
B24
B25
B26
B27
B28
B29
Initial
18430
18430
18430
18430
18468
18468
18468
18620
18620
18620
18620
18620
18620
18620
18620
18544
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
Day 3
14250
14250
14250
14250
14250
13680
13680
11856
5016
14250
14592
6460
14653
14630
<1520
14630
14630
14440
14136
14630
12730
12730
4750
12350
11590
13718
10260
11438
13110
11552
13300
13110
14250
14136
14098
14098
13870
10602
8360
13718
Day 7 Initial-D
1
11590
11438
11400
11590
1900
10260
10070
7790
<1520
11210
11590
4560
11590
11590
<1520
11590
11590
11400
<1520
11590
8930
8930
<1520
8550
7258
10450
5700
7220
9500
7410
9880
9462
11020
11248
10868
11020
10488
5168
4180
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18240
18810
18810
18810
18810
19000
19000
19000
19000
19000
19076
19076
19152
19152
19190
19190
19190
19190
19190
18810
18810
18810
19000
19000
19000
19000
10450 19000
Day 3-D
14440
14440
14630
12540
14440
14630
14820
14440
14630
14440
14820
14630
15010
15010
<1520
14630
14820
14440
14440
6080
11590
11970
12730
2280
5320
14440
11970
11020
13490
13300
13490
8740
12350
8740
13300
13870
14440
14630
12350
14820
Day 7-D
<1520
3230
11400
<1520
11020
11020
11400
11590
9880
11020
<1520
<1520
<1520
11590
<1520
10830
<1520
11020
9310
<1520
7030
7790
8550
<1520
1710
10830
7410
6460
9500
9120
9690
3800
9310
5510
9500
10450
11210
11780
7600
9120
12
-------
Table 2. Septa-sealed vacuum test results (cont.).
Vacuum (mm Hg)
Vial ID
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CIO
C20
C21
C22
C23
C24
C25
C26
C27
C28
C29
Dl
D2
D3
D4
D5
D6
D7
D8
D9
D10
D20
D21
D22
D23
D24
D25
D26
D27
D28
D29
Initial
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
17480
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
18620
Day 3
<1520
<1520
12540
<1520
12008
<1520
<1520
14440
1710
<1520
<1520
9880
11210
2280
10450
<1520
14250
9348
14592
<1520
<1520
12388
<1520
14250
14250
14212
14250
14250
14250
14060
14288
14250
14250
14250
14212
14288
<1520
14250
14288
14288
Day 7 Initial-D
1
<1520
<1520
<1520
<1520
9690
<1520
<1520
4940
<1520
<1520
<1520
<1520
8550
<1520
8056
<1520
6080
4370
11590
<1520
<1520
8550
<1520
11210
11210
11020
11020
11058
11172
10830
11362
11020
11020
11210
11020
11248
<1520
11172
11210
19190
19000
19000
18620
19190
19190
19190
19190
19190
19190
19190
19190
19190
19190
19190
19190
19190
19228
19228
19228
18620
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19380
19190
11210 19190
Day 3-D
14630
13110
8170
15010
15010
14820
12160
14630
13490
11400
14630
14820
15010
15010
15010
15010
15010
15010
15010
15010
<1520
11780
<1520
14630
14630
14630
14630
<1520
14630
14820
14630
14630
14630
<1520
14820
14630
14630
14630
14630
14630
Day 7-D
10830
8930
3420
11780
11780
11400
7600
10450
9500
6650
11020
11590
11780
11780
11780
11780
11780
11590
11780
11780
<1520
7220
<1520
11210
11210
11210
11210
<1520
11210
11210
11210
11210
11210
<1520
11020
11400
11210
11210
11020
11400
Second run for all vials labeled "-Dr
13
-------
VOC Loss Test
Nearly all mean VOC concentrations after 14 days were within ± 20% of the initial VOC
concentrations indicating little to no loss of VOCs during storage except for acetone in vials from
manufacturer C (Table 3). The first analytical runs were performed on vials from manufacturer
C. Acetone appeared to require a number of samples be run before the trap was properly
"conditioned." Therefore, the "noise" in the acetone recoveries is unlikely to be related to the
condition of the vial seals.
The somewhat consistently lower VOC recoveries observed for vials from manufacturer
C can be attributed to instrument instability (Table 3). Vials from each manufacturer were run as
a batch, alternating the potentially leaky and control vials in sequence to even out instrument drift
that tends to occur over the 14 hour batch runs (Appendices B and D). VOC concentrations in
manufacturer C vials tended to be lower near the end of the analytical run, but recovered to the
initial values before the run ended (Appendix B - Tables B2-5 and B2-6, Appendix D). The
influence of this instrument fluctuation on VOC concentrations is difficult to assess; however,
the lower concentrations are more likely due to an artifact of the analytical system rather than
caused by leaky vial seals.
Precision among the vials was acceptable with RSDs of < 20% with one exception (Table
3). A RSD of 26% was found for 1,1-dichloroethene in the control vials from manufacturer C.
Upon examination of individual sample data, only vial Dl was clearly leaky (Appendix
B). The chip in the side of the lip was so large that the vial would not seal. No data are reported
for this vial because the recoveries of internal standards, added just prior to purging, were well
below the QC criteria (Appendix D).
No statistically significant differences were identified between VOC concentrations in the
potentially leaky and control group vials from the same manufacturer (Table 3). The maximum
absolute concentration difference between the two vial groups was 2 |j,g/mL.
VOC Gain Test
Only 3 vials from this test were analyzed due to overwhelming concentrations of
methylene chloride found in each of the tested vials. The resultant chromatographic peaks were
so broad and the concentrations so far above the calibration range that no further analytical
testing was performed. The results of this test indicate that the septa could not prevent sample
cross contamination by methylene chloride. Whether the cross contamination was due to poor
seals or diffusion through the septa is not be positively known although the later explanation is
most probable.
14
-------
Table 3. Summary of VOC-Loss Test Results.
VOC Concentration, (ng/mL)
A1-A10
Acetone
1 , 1 -Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
Mean
36
40
57
44
38
40
38
RSD
8
5
2
5
5
5
5
A20-A29
Acetone
1 , 1 -Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
Mean
37
40
57
44
38
39
37
RSD
5
5
4
5
5
5
5
B1-B10
Mean
37
40
61
41
38
37
37
RSD
11
8
5
7
8
8
8
B20-B29
Mean
37
41
61
42
38
37
37
RSD
11
7
3
5
8
8
8
C1-C10"
Mean
28
35
53
40
33
35
34
RSD
7
17
6
10
9
9
9
C20-C29
Mean
27
34
53
39
32
34
33
RSD
7
26
9
15
19
14
9
Dl-D10f
Mean
39
37
56
41
33
36
34
RSD
10
11
5
7
9
8
9
D20-D29f
Mean
37
38
57
42
34
37
35
RSD
8
5
7
5
3
3
6
* - Mean and relative standard deviations (RSD) often potentially leaky vials (1 through 10)
and ten control vials (20 through 29) for manufacturers A through D.
** - data are mean and RSDs of 8 vials (n=8).
f - data are mean and RSDs of 9 vials (n=9).
Summary and Conclusions
Visual inspection of numerous VOA vials from the four major manufacturers in the
United States found minor imperfections ranging from small indentations, bumps, scratches on
vial threads or lips, through obvious defects, such as large indentations or grooves in the vial lips
and chipped or broken glass. A less obvious imperfection that could affect the ability of a vial to
15
-------
seal completely was uneven rims in which one side of the vial's neck was clearly longer than the
other side. Observed imperfections rates were 7, 4, 15, and 6% for manufacturer A, B, C, and D,
respectively. Vials from manufacturer C had over twice as many imperfections (i.e., 32 out of
216) as vials from the other manufacturers.
An aluminum plate vacuum test was developed as a means to perform a quick and simple
test to determine if a vial will form a complete vacuum seal. Unfortunately, no matter how
smooth the plate's surface was, none of the vials sealed completely against the plate. With the
vacuum pump on, some vials were capable of reaching the pump's maximum pressure; however,
when the pump was turned off, the vacuum quickly dissipated. These results make the value of
the aluminum plate vacuum test in selecting whether or not a vial will obtain a complete seal
against VOC loss highly questionable.
A septa-sealed vacuum test was conducted to determine if the septa were flexible enough
to form a complete seal even in the presence of vial imperfections. The results of this test were
relatively inconsistent with the prior two tests. The vials from manufacturer A with visual
imperfections did not lose their vacuum seal. In stark contrast, the only vials that failed to hold a
vacuum were free from visual defects or were in the vial control group. All vials, with or
without visual imperfections, in the potentially leaky group and 3 vials from the control group of
manufacturer B were leaky. The aluminum plate vacuum test appeared to be a viable screening
option for manufacturer B's vials although some false positives (i.e., vials falsely declared leaky)
did occur. Vials made by manufacturer C had half of the vials in the potentially leaky group fail
to hold their vacuum and the presence of a visual imperfection could not be used to clearly
identify which ones. However, due to greater flexibility in the septa, the analyst had to be sure
that the needle punctured the septa within 2 mm of the center or else the potential for seal
leakage markedly increased. For manufacturer D, the presence of visual imperfections clearly
indicated which vials would fail to hold their vacuum. All 3 vials with imperfections
catastrophically failed to hold a vacuum while the remaining 7 vials identified as potentially
leaky remained sealed after 7 days. No clear conclusion can be drawn based on the septa-sealed
vacuum test about whether the flexibility of the silicone septa is sufficient enough to form a
complete seal against VOC losses during storage. This test was subject to a noticeable rate of
false positives.
Mean VOC concentrations after 14 days storage generally were within ± 20% of the
known concentration with a majority of the concentrations within ± 15% of their known values.
There were no statistically significant differences in VOC concentrations between vials in the
potentially leaky and control groups for any of the manufacturers. Only 1 vial lost VOCs and
that was due to a large chip in the vial's lip and neck. These findings indicate that the silicone
septa are flexible enough to overcome most vial imperfections and form a complete seal against
VOC loss. Further, the use of the other tests (excluding visual inspection) performed in this
study are unnecessary to screen for potentially leaky vials. A careful inspection of the VOA vials
upon receipt (or at least, prior to use) to remove any vials with large and obvious imperfections
16
-------
should be sufficient to screen out vials that are subject to VOC losses through inadequate septum
sealing.
It should be noted that when puncturing the septa, the point of entry of the sparging
needle on the purge-and-trap unit should be in the center (or within a few mm) of the septa. If it
is not in or near the center, differences in septa flexibilities may result in a loss of the vial's vapor
tight seal. A study to evaluate the properties of septa that influence their flexibilities and thus,
their sealing potential, may be warranted.
17
-------
References
EPA. 1996. Test methods for evaluating solid waste, SW-846. 3rd Edition. Office of Solid Waste
and Emergency Response. U.S. Environmental Protection Agency, Washington, DC.
December, 1996.
Hewitt, A.D., T.F. Jenkins, and C.L. Grant. 1995. Collection, handling, and storage: Keys to
improved data quality for volatile organic compounds in soil. Am. Environ. Lab. 2:25-28.
18
-------
APPENDIX A1
Visual Inspection Log and Plate Vacuum Test Results by Vial
1 NOTE: Vacuum readings presented in the following tables are directly from the vacuum
gauge and have the units of inches Hg. To convert from inches Hg to mm Hg: 760 mm Hg = 1
inch Hg.
-------
APPENDIX B
Individual VOA Vial Analytical Results by Manufacturer
-------
Table B2-1. Analytical Results for potentially leaky vials from manufacturer A .
Al A2
A3
A4
AS
A6 A7 A8
A9
A10
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
40 35
40 38
56 55
45 42
39 36
40 37
39 36
System
88 88
91 92
84 84
36
38
56
42
35
37
35
38
36
55
41
35
36
34
38D
39
57D
43
38
39
37
34 31 37
41 41 41
57 57 58
46 44 45
42 39 39
42 40 41
41 38 39
34
41
57
44
40
41
40
33
42
59
46
41
42
41
Monitoring Compound (SMC), % Recovery
90
93
83
89
93
83
87
91
83
- data superscripts indicate flagged data. Flag definitions are presented
Table B2-2. Analytical
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
Results for control vials from
A20 A21 A22
37 35 40
36 37 40
53 55 56
40 41 44
35 35 37
36 37 38
34 35 35
A23
42
39
56
43
36
37
35
manufacturer
A24
VOC
35
39
56
42
36
37
35
A25
88 89 91
92 92 93
83 82 83
in Appendix D.
A*.
A26 A27
85
91
83
A28
85
90
83
A5D
26
36
41
39
32
33
38
76
73R
83
A29
A25D
Concentration, (ng/mL)
38
43
58D
47
42D
42
40
System Monitoring Compound
l,2-Dichloroethane-d4
Toluene-d8
Bromofluorobenzene
89 91 90
93 93 94
85 84 84
89
93
83
89
93
85
95
97
87
35 39
41 41
57 59
45 44
38 40
40 41
38 39
35
40
57
44
39
40
38
38
42
59
45
40
41
40
(SMC), % Recovery
89 91
93 93
83 84
89
92
82
91
92
81
29
36
38
38
30
32
34
78
74R
83
- data superscripts indicate flagged data. Flag definitions are presented in Appendix D.
B- 1
-------
Table B2-3. Analytical Results for potentially leaky vials from manufacturer B*.
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
Bl B2
39C 39C
43 37
61 58
44 40
41 37
40 37
40 37
System
89 87
89 90
87 85
* - data superscripts indicate flagged data. Flag
Table B2-4. Analytical
B3
VOC
37C
33
57
36
31
31
31
B4
B5 B6 B7
B8
B9
BIO
Concentration, (ng/mL)
37C
38
60
40
37
36
36
40C 39C 35C
40 40 42
62 60 61
41 42 43
37 39 39
36 37 38
36 37 38
Monitoring Compound (SMC), %
89
90
84
89
91
84
88 88 88
90 91 91
85 83 84
definitions are presented in Appendix
27C 41C
39
61
40
35
34
34
42
63
41
38
37
38
37C
45
66
45
42
41
40
Recovery
87
90
84
D.
88
90
84
89
91
85
BSD
31C
50
58
46
37
39
36
94
89
82
Results for control vials from manufacturer B*.
B20 B21 B22
B23
B24
B25 B26 B27
B28
B29
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
34C 38C 39C
41 40 36
59 61 59
42 42 37
39 38 32
38 38 31
38 38 30
38C
42
63
43
40
38
38
39C
41
62
41
38
37
37
43C 37C 32C
40D 44
59 64
41 44
37 40
36 39
36 39
38
60
40
35
34
33
31C
43
63
44
41
39
40
36C
45
64
45
42
41
41
System Monitoring Compound (SMC), % Recovery
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
87 88 87
89 89 89
88 87 85
84
90
83
88
90
85
88 89
90 91
84 86
87
90
84
87
90
89
86
91
85
B25D
38C
51
59
47
37
40
36
97
90
84
- data superscripts indicate flagged data. Flag definitions are presented in Appendix D.
B-2
-------
Table B2-5. Analytical Results for potentially leaky vials from manufacturer C ' .
Cl C2
C3
C4
C5 C6 C7
C8
C9
CIO
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
29C 26C
40 33
53 52
43 38
37 32
39 34
38 33
System
90 91
89 89
80 81
- data superscripts indicate flagged data. Flag
** - nr = not reported (see text for discussion).
Table B2-6. Analytical
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
30C
38
56
41
34
36
34
25C
34
52
38
30
33
31
27C 27C 24C
28 40 6D
49 54 30D
34 43 17D
28 36 13D
31 38 17D
31 36 20D
nrT
nrT
nrT
nrT
nrT
nrT
nrT
32C
27
50
35
29
32
32
26C
42
58
44
36
38
436
Monitoring Compound (SMC), % Recovery
89
88
81
88
89
83
89 87 87
89 89 90
81 80 79
nrT
nrT
nrT
93
90
82
90
89
79
C5D C7D C8D
27 32C 32C
46 50 48
56 58 57
40 46 44
32 37 33
34 39 37
30 35 33
88 90 89
94 95 98
91 86 88
definitions are presented in Appendix D.
Results for control vials from manufacturer C*.
C20 C21 C22
27C 29C 29C
39 33 39
55 53 56
43 37 40
36 30 33
38 33 35
36 32 33
C23
VOC
25C
38
55
41
34
36
34
C24
C25 C26
C27
C28
C29
Concentration, (ng/mL)
26C
42
58
44
38
38
37
29C 27C
43 21
56 43
44 30
38 24
39 27
37 28
25C
23
47
32
26
28
29
30C
22
47
32
25
29
29
26C
42
58
44
38
39
36
System Monitoring Compound (SMC), % Recovery
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
91 89 90
89 89 89
89 82 82
89
90
83
87
88
79
90 86
91 89
82 80
87
89
82
91
90
80
89
89
80
C25D
32C
49
58
43
34
37
33
90
93
88
- data superscripts indicate flagged data. Flag definitions are presented in Appendix D.
B-3
-------
Table B2-7. Analytical Results for potentially leaky vials from manufacturer D .
Dl D2
D3
D4
D5 D6 D7
D8
D9
DIG
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
nr 37
nr 32
nr 51
nr 36
nr 26
nr 29
nr 26
System
nd 90
nd 93
nd 84
46
29
52
36
29
33
32
38
37
56
41
33
36
35
35D 40 37
38 38 39
57 57 58
43 41 43
33 33 34
36 36 38
34 35 36
40
39
59
43
35
38
36
41
40
59
43
35
37
35
33
41
59
44
35
38
36
Monitoring Compound (SMC), % Recovery
92
91
84
* - data superscripts indicate flagged data. Flag definitions
Table B2-8. Analytical
Results for control vials from
D20 D21 D22
D23
88
91
85
89 88 87
91 90 90
84 84 85
are presented in Appendix
86
90
83
D.
88
90
85
90
90
83
DSD
21C
38
50
41
29
34
31
87
85
84
manufacturer D .
D24
D25 D26
D27
D28
D29
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
35 39 36
34 36 38
50 54 56
39 41 42
31 33 34
34 36 38
32 34 35
44"
nr
4"
nr
nr
2"
3"
33
38
56
42
33
37
34
40D 41
38 38
58 57
43 42
34 34
38 37
36 35
38
40
59
44
35
38
36
35
41
60
44
35
38
37
38
42
62
45
35
39
37
System Monitoring Compound (SMC), % Recovery
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
90 89 90
92 91 92
83 84 83
92
85
84
88
91
83
85 89
90 91
84 85
91
91
84
87
90
85
87
89
84
D25D D25D
23C 27CR
38 39R
49 49R
40 41R
27 28R
33 34R
29 30R
89 92
84 86
82 83
- data superscripts indicate flagged data. Flag definitions are presented in Appendix D.
- septum in vial PTFE-side up. Analytical results for information only, nr = not reported.
B-4
-------
APPENDIX C
Quality Assurance/Quality Control Report for the Vacuum Studies
-------
The frequency and types of quality control samples for the vacuum studies followed or
exceeded the specifications in the QAPP, "Integrity of VOA Vial Seals," January 7, 1998, Draft
1.2. The same vacuum pump and Grade AA vacuum gauge were used throughout the study.
The maximum vacuum attained by the system was recorded each day before analyzing
samples. Data for this calibration step are shown in Table A3-1.
Table A3-1. Check for vacuum system integrity on days that vacuum readings were made.
Date
12/10/97
12/15/97
12/21/97
12/22/97
12/23/97
3/31/98
4/3/98
4/6/98
4/7/98
4/10/98
4/13/98
Vacuum (mm Hg)
18620
18620
18620
18620
17860
18088
18468
18620
18620
18620
19380
Duplicate vacuum readings were performed every 10 samples for the plate-vacuum study
and for all samples in the septa-seal vacuum study. Data for the duplicate samples are given in
Tables B2-1 through B2-8. A summary of the percent difference (%D) for the plate-vacuum
study duplicates is given in Table A3-2. The reproducibility of the test was less than anticipated
with RPD values ranging from zero to 13.5%. The mean RPD was 3.8%.
C-l
-------
Table A3-2. Relative percent differences for duplicate vacuum plate samples.
Data presented are for 22 duplicate samples out of 216 vials from each
manufacturer.
Manufacturer Manufacturer Manufacturer Manufacturer
A B C D
6.3
12.5
6.5
3.6
5.7
1.6
3.1
5.1
3.1
1.8
3.0
6.2
1.1
1.8
5.5
1.1
4.5
1.1
1.1
0.8
0.6
1.0
3.6
0.7
6.7
3.8
3.6
5.4
10.4
11.5
9.1
6.2
0.9
4.3
3.4
7.7
0
8.2
6.3
3.4
4.5
8.7
2.6
1.1
10.9
0
8.7
0
4.8
13.5
0.9
9.5
6.1
7.9
1.3
0
2.5
8.2
11.5
0.5
6.3
2.3
2.8
1.1
1.1
1.1
12.1
2.7
0
1.3
2.5
0
1.2
1.2
1.2
1.2
1.2
0
1.1
8.4
0.5
1.1
1.8
3.5
0
7.0
0
2.2
C-2
-------
APPENDIX D
Quality Assurance/Quality Control Report for the VOC Loss Test
-------
The types of quality control samples, acceptance criteria, and qualifiers given in Table
A4-1 reflect those listed in Table 1 of the QAPP "Integrity of VGA Vial Seals," January 7, 1998,
Draft 1.2. These qualifiers appear in Tables B2-1 through B2-8, where appropriate.
TABLE A4-1. VOC PC SAMPLES and ACCEPTANCE CRITERIA
QC Sample
Duplicates
Continuing calibration
Internal standards
SMC recovery
IDL (7 replicates)
Acceptance Criteria
% D < 25% for all target analytes
% D < 25% for all target analytes
Area counts 50-200% of 50 [ig/mL std.
% R = 100 ± 25% for all target analytes
% R = 100 ± 25% for SMCs
Qualifiers
flag "D"
flag "C"
flag "T"
flag "R"
--
Duplicate Samples
One duplicate sample for each set often "potentially leaky" or "control" vials was
reported. Additional duplicate data are given for samples C7, C8, and D23 because the initial
data were questionable (either the internal standard or the system monitoring compound recovery
was poor).
Instrument Detection Limits
The instrument detection limits (IDLs) were determined as 3.143 times the standard
deviation obtained from the analysis of 7 replicates of the calibration standard at a nominal
concentration of 5 ng/mL. The system monitoring compound (SMC) recoveries for all
compounds were within the ± 25% as specified. By this method, the detection limits, in ng/mL,
for the compounds were: acetone = 3.5, 1,1-dichloroethene = 0.8, methylene chloride = 1.0,
benzene = 0.3, trichloroethane = 0.3, toluene = 0.7, and chlorobenzene = 0.5 ng/mL. These
detection limits are approximately 2 orders of magnitude lower than the 50 ng/mL working VOC
concentrations required for this work.
Instrument Blanks
Instrument blanks were run prior to the analysis of an initial 10 ng/mL standard used in a
5 point initial calibration curve. Blanks were not run following the CCVs (continuing calibration
verification) prior to sample analysis. During preliminary analysis of VOC samples, it became
evident that the first sample of a batch run immediately following a blank yielded low
concentrations for the target analytes relative to the remainder of the batch. This was attributed
to the blank which conditioned the trap differently than a sample. Because the CCV closely
matches the nominal 50 ng/mL concentration of the sample target analytes, analysis of the CCV
D- 1
-------
conditions the trap in the purge and as if it were a sample. Since the objective of this project was
to determine differences in VOC concentrations as a consequence of vial seal integrity, not trap
dependency of previously run samples, analysis of blanks following a CCV was eliminated.
Continuing Calibration
Continuing calibration data are shown in Tables D2-1 through D2-4. Analytes in the
CCVs that did not meet the 25% D window are flagged. Samples that were run immediately
prior to the CCV that did not meet the 25% D window are also flagged.
The polar nature of acetone yielded widely variable recoveries for this analyte. Acetone
failed the CCV most often while the remaining analytes easily met criteria. It was deemed
impractical to run new calibration curve solely for acetone. This forced some portions of an
analytical batch to be run with acetone failing the midpoint CCV. These samples are flagged.
SMC Recovery
The system monitoring compounds, l,2-dichloroethane-d4, toluene-dg,
bromofluorobenzene were added to every sample and CCV. SMC recoveries of samples are
reported with the sample data (Tables B2-1 through B2-8). Those samples not meeting the ±
25% R criteria are flagged.
Samples
Analytical data for target analytes are presented in Tables B2-1 through B2-8. Each batch
of 20 samples from one vendor were run in the sequence of CCV1 XI, X20, X2, X21, X3, X22,
X4, X23, X5, X24, CCV2, X6, X25, X7, X26, X8, X27, X9, X28, X10, X29, and CCV3 where
X represents the vendor letter identification.
Tables B2-1 and B2-2 contain VOC test results from Manufacturer A vials. For vial A5,
acetone and methylene chloride did not meet %D criteria in the sample duplicate and were
flagged. Variability in duplicates was not unexpected. The duplicate samples came from 2
different analytical runs using two different spiking solutions that were quantified from two
different initial calibration curves. The polar nature of acetone introduces further variations in
recoveries depending on the conditioning of the trap as has been discussed previously. In the
sample duplicate A5D, recovery of the SMC, toluene-d8, was 2% low and was flagged. The
remaining 2 SMCs were low, but passed.
For vial A25 in Table B2-2 methylene chloride and trichloroethene did not meet %D
criteria in the sample duplicate and were flagged. In the sample duplicate A25D, recovery of the
SMC, toluene-dg, was low and was flagged. The remaining 2 SMCs were low but passed.
D-2
-------
Tables B2-3 and B2-4 contain VOC test results from Manufacturer B vials. The target
analyte acetone failed high in CCV2. Because this was a mid batch CCV, the samples preceding
the CCV which include, Bl, B20, B2, B21, B3, B22, B4, B23, B5, and B24 and the samples
following the CCV which include B6, B25, B7, B26, B8, B27, B9, B28, BIO, and B29 were
flagged for acetone. The target analyte acetone failed high in CCV2D. Because this was a mid
batch CCV, the sample BSD preceding the CCV and the sample B25D following the CCV were
flagged for acetone.
For vial B25 in Table B2-4, 1,1-dichloroethene did not meet %D criteria in the sample
duplicate and was flagged.
Tables B2-5 and B2-6 contain VOC test results from Manufacturer C vials. The target
analyte acetone failed low in CCV2 and CCV3. Consequently samples Cl, C20, C2, C21, C3,
C22, C4, C23, C5, C24, C6, C25, C7, C26, C8, C27, C9, C28, CIO, and C29 were flagged for
acetone. The target analyte acetone also failed low in CCV3D. Because this was a closing CCV,
the sample duplicates C7D and C8D preceding this CCV were flagged.
Vial C7 in Table B2-5 did not meet the %D criteria for all analytes and was flagged
although the SMCs met their QC criteria. Visual inspection of the vial did not reveal observable
defects. It is suggested the low recoveries of the analytes are a consequence of a loose cap due to
incomplete tightening rather than defects in the sealing lip of the vial.
Vial C8 in Table B2-5 did not meet the %D criteria for all analytes and the recoveries for
the SMCs were all significantly below 50% so this sample was flagged. Visual inspection of the
vial did not reveal observable defects. It is suggested as for vial C7, the loss of analytes and
SMCs are a consequence of a loose cap causing VOCs to be lost during the purge cycle rather
than defects in the sealing lip of the vial.
Tables B2-7 and B2-8 contain VOC test results from Manufacturer D vials. The target
analyte acetone failed low in CCV2D. Consequently samples DSD, D23D, and D25D were
flagged for acetone.
The CCV and the sample D25D following the CCV were flagged for acetone. The
recovery for the SMC 1,2-dichloroethane-d4 was low in CV3D and all analytes in sample D25D
were "R" flagged.
Analytical results for sample Dl in Table B2-7 was not reported. The glass lip of the vial
was so badly chipped, that a seal could not be obtained between it and the septum cap. Any
remaining VOCs as well as SMCs added immediately prior to the purge were lost. Sample D5
did not meet %D criteria for acetone and was flagged. The septum cap in vial D23 had the
Teflon-septum facing exterior to the vial which allowed permeation of VOCs through the
septum. Interestingly, the polar VOC acetone, showed no loss by permeation into/through the
silicone septum. All SMCs in the sample meet criteria indicative of a valid purge. Analytical
D-3
-------
data for sample D23 are for information only as the vial cap was improperly configured. Sample
D25 did not meet %D criteria for acetone and was flagged.
D-4
-------
Table D2-1. Continuing calibration results for samples and duplicate samples from manufacturer A.
CV1
CV2
CV3
CV1D CV2D
CV3D
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
46
52
58
52
52
51
52
50
48
57
48
50
49
50
57
49
55
47
49
48
50
System Monitoring
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
Table D2-2. Continuing
97
103
104
calibration
CV1
98
99
100
95
96
98
results for samples
CV2
CV3
57
43
41
45
42
43
50
Compound (SMC), %
86
82
98
48
41
43
44
42
44
51
40
40
42
45
42
44
50
Recovery
90
83
102
and duplicate samples from
CV1D CV2D
188
83
98
manufacturer A
CV3D
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
59
50
52
48
51
48
51
64C
51
54
49
52
50
51
40
46
50
46
49
47
49
System Monitoring
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
95
98
106
99
102
105
93
97
102
45
54
52
51
48
50
48
Compound (SMC), %
101
100
97
71C
57
60
53
52
53
51
43
53
57
53
48
50
48
Recovery
113
106
105
110
101
100
- data superscripts indicate flagged data. Flag definitions are presented in Table A4-1.
D-5
-------
Table D2-3. Continuing calibration results for samples and duplicate samples from manufacturer C
CV1
CV2 CV3
CV1D CV2D
CV3D
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
43
51
51
52
51
51
52
36C 31C
48 48
50 50
49 50
49 49
49 50
50 50
System Monitoring
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
- data superscripts indicate
104
101
103
flagged data.
98 99
99 101
101 102
Flag definitions
Table D2-4. Continuing calibration results for samples
CV1
CV2 CV3
50
56
57
53
53
55
53
Compound (SMC), %
105
109
110
are presented in Table
40
54
56
51
51
47
50
37C
57
60
54
51
54
49
Recovery
99
102
105
A4-1.
and duplicate samples from
CV1D CV2D
111
112
102
manufacturer D
CV3D
VOC Concentration, (ng/mL)
Acetone
1,1-Dichloroethene
Methylene Chloride
Benzene
Trichloroethene
Toluene
Chlorobenzene
50
48
52
49
51
50
51
47 52
49 48
53 53
57 48
51 51
49 49
51 52
System Monitoring
l,2-Dichlorethane-d4
Toluene-d8
Bromofluorobenzene
99
100
104
96 98
100 101
104 114
50
53
52
52
52
52
51
35C
44
51
47
46
47
47
44
49
45
42
41
41
39
Compound (SMC), % Recovery
107
104
102
97
95
99
56R
81
79
- data superscripts indicate flagged data. Flag definitions are presented in Table A4-1.
D-6
------- |