Eyaluatimi o-f Sampling Methods fqfc
Gaseous Atfnospheric Samples
Research Triangle Inst.
Resear-ch Triangle Bark, NC
ftit
Sc?i-encss: Research Lab,
'Earl?-,: NO
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
-------
-------
-------
NOTICE
•This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
-------
ABSTRACT
A research program was conducted to test and evaluate several alternati-
ves for collecting and transferring.samples from the collection site to the
laboratory for the analysis of a variety of toxic organic pollutants by gas
chromatography (GC). Sample storage media included three types of polymeric
bags (FEP Teflon , Tedlar, five-layered aluminized bags), glass bulbs,
electropolished and Summa polished cannisters, Tenax GC and charcoal
cartridges, and nickel cryogenic traps. Twenty-seven test compounds including
hydrocarbons, airomatics, halogenated hydrocarbons, halogenated aromatics,
and oxygen, nitrogen and sulfur-containing compounds were used to test the
storage media. Dynamically flowing mixtures of these gases were synthesized
using a specially designed permeation/dilution system. Quantitative labora-
®
tory stability tests were conducted with Tenax GC, charcoal, and cryogenic
traps at 2 concentration levels of 50 parts per billion (ppb) and 200 parts
per trillion (ppt), for 15 of the 27 chemicals. Quantitative stability
tests were conducted with the remaining storage media at one concentration
level, nominally 50 ppb, for the same 15 chemicals. The stability tests
were conducted over a 7 day storage period. Also, quantitative stability
® ®
tests were conducted with Summa polished cans, glass bulbs, Tedlar bags
®
and Tenax GC cartridges at one concentration level, nominally 50 ppb, for
the remaining group of chemicals.
The potential effect of inorganic gases as interferences during the
®
collection of test compounds was quantitatively studied with Tenax GC,
§
charcoal, cryogenic traps, Summa polished cans, glass bulbs and Tedlar bags
at two concentration levels of inorganic substances for approximately one-
half of the test compounds. The remaining test compounds were also collec-
ted in the presence of inorganic gases using Tenax GC cartridges, Summa
polished cans, glass bulbs and Tedlar bags.
Sampling systems were designed and fabricated as necessary to collect
valid gas samples for the quantitative tests conducted with the various
111
-------
collection devices using a permeation/dilution system for the synthesis of a
dynamic flow, synthetic air vapor mixture.
A quality control and quality assurance (QC/QA) program was established
and maintained for all measured and analyzed data. This QC/QA program
included the following elements: (a) sampling procedures; (b) calibration
procedures; (c) analytical procedures; (d) data collection and reporting
procedures; (e) auditing procedures; (f) storage procedures; and (g) computa-
tional and data validation procedures.
The sample collection and transfer methods were tested, evaluated and
compared for the following elements: (a) limits of applicability; (b) col-
lection and recovery efficiency for gas chromatographic analysis; (c) analy-
tical accuracy and detection limits; (d) interferences by inorganic gases;
(e) sample stability in storage; (f) quality of chromatograms; and (g) sim-
plicity and convenience. Because of resource limitations, however, statisti-
cal analyses of data on comparing methods by chemical or group types and on
interference effects were not conducted.
The support coated open tubular (SCOT) capillaries coated with SE-30
were the best available when this research was initiated and performed ade-
quately throughout the study. They were rugged with two SCOTs per instru-
ment required over a three year usage. Fused silica capillaries were not
commercially available when the program started and, thus, there was no
opportunity to evaluate them.
An automatic two channel ambient air sampler utilizing sorbent cartridges
as the collection medium was designed and fabricated. The mechanical and
electronic systems of the automatic sampler were designed and built from
commercially available components to include the following features: (1)
two sampling channels with two sampling heads with provisions for six samples
per head and a blank, (2) capability for collecting 12 single or 6 duplicate
samples, (3) variable orifices, which are manually set for low flow rate
settings and range through each channel, (4) sampling rates settable from 7
mL/min to 1.5 L/min, (5) a mass digital flow meter (switchable to each
channel individually, or measurement of total flow for both channels) with
the ability to integrate total flow, (6) solid-state timer system with clock
integrator, printer and manual mode, (7) sampling periods selectable at 15,
IV
-------
30, 45 and 60 min, and 0.5, 2, 3, 6, 8 and 12 hr, (8) reset capability for
flow integrator after collection of each sample (but not after a power
failure), and (9) operative on 120V alternating current.
The automatic sampler's systems were checked to insure proper electrical
and mechanical functioning under laboratory simulated conditions. These
tests included stepping sequences, clocking, printing, resetting, and sampling
head sealing (pressure). Calibration of functions such as flow meter,
integrator, timers, etc. was conducted. Background tests on sorbent cart-
ridges sealed in sampling heads were also instituted.
-------
CONTENTS
Abstract iii
Figures viii
Tables xiv
Acknowledgements xix
1. Introduction 1
2. Conclusions 4
3. Recommendation . 10
4. Program Objectives . 13
5. Design and Fabrication of Permeation/Dilution System 20
6. Evaluation of Sample Collection Devices 34
7. Support and Quality Control Data 162
8. Design and Fabrication of an Automatic Sampler 216
9. Performance of Automatic Sampler 248
10. Preliminary Development of Diffusion Tubes for Low
Vapor-Pressure Compounds i • 263
References 272
Preceding page blank
Vll
-------
FIGURES
Number Pa
1 Overall schematic of permeation system 21
2 Permeation chamber, storage vessel, and enclosure 22
3 Dilution bulbs and heated enclosure 24
4 Dilution bulb inlet 25
5 Tenax cartridge sampling arrangement employed with
permeation/dilution system 43
6 Charcoal cartridge sampling arrangement employed
with permeation/dilution system 43
7 Purging (A) and storage arrangements (B) for
recovering vapors from Ni traps 45
8 Nickel cryogenic trap sampling arrangements employed
with permeation/dilution system 46
ii
9 Vacuum injection manifold 48
10 Inlet-manifold 52
11 Schematic of vaporization unit for loading organics
dissolved in methanol onto Tenax GC cartridges 55
12 2 mil Teflon Bag 61; low concentration study; Group I
compounds T ; June 13, 1980 61
13 2 mil Teflon Bag 61; low concentration study; Group I
compounds; T ; June 20, 1980 62
14 Tedlar Bag F; low concentration study; Group I
compounds, T ; June 12, 1980 63
15 Tedlar Bag F; low concentration study; Group I
compounds, T ; June 12, 1980 64
16 2 mil Teflon Bag #61, zero air bag; stored in clean air;
TIO, July 30, 1980 66
17 2 mil Teflon Bag #61, zero air bag; stored in zero air; T .,
August 9, 1980 67
Vlll
-------
FIGURES CONT'D.
Number Page
18 2 mil Teflon Bag #62, zero air; stored in lab air; TQ,
July 30, 1980 68
19 2 mil Teflon Bag #62, zero air; stored in lab air; T .,
August 9, 1980 . 69
©
20 Chromatogram of air in Teflon bag #13; T-, December
16, 1980 74
®
21 Chromatogram of air in Teflon bag #13; T , December
19, 1980 '. 75
©
22 Chromatogram of air in Teflon bag #13; T,, December
22, 1980 76
®
23 Chromatogram of air in steel box for Teflon bag
experiment; T , December 17, 1980 77
®
24 Chromatogram of laboratory air used for Teflon bag
experiment; December 16, 1980 78
©
25 Chromatogram of clean air for Teflon bag experiment;
T , December 16, 1980 79
®
26 Chromatogram of air in steel box used for Teflon bag
experiment; T , December 19, 1980 80
J
27 Chromatogram of air in Tedlar bag #IX; T_, December
30, 1980 82
28 Chromatogram of air in Tedlar bag #IX; T,,, January
2, 1981 : . . . 83
®
29 Chromatogram of laboratory air used for Tedlar bag
experiment; December 30, 1980 84
®
30 Chromatogram of air in steel box used for Tedlar bag
experiment; T , January 2, 1981 85
®
31 Chromatogram of air in steel box used for Tedlar bag
experiment; T , December 31, 1980 86
®
32 Chromatogram of laboratory air used for Tedlar bag
experiment; December 30, 1980 87
33 FID Chromatogram of T Tenax sample (high level study). . . 110
®
34 Background profile of Tenax GC cartridge used in
low level study 112
ix
-------
FIGURES CONT'D.
Number Page
35 Background profile for 30 L of air from permeation/
dilutor passed through a Tenax GC cartridge 113
36 ECD chromatogram of T carbon tube extract (high
level study) 115
37 Chromatogram for background observed with Tenax used
to sample 30 L of air containing low levels of inorganic
gases (no GFF) 131
38 Chromatogram of background observed with Tenax used to
sample 30 L air containing low levels of inorganic
gases (with GFF) 132
39 Chromatogram for sample taken with Tenax with test
compounds and low levels of inorganic gases present
(no GFF) 133
\
40 Chromatogram for sample taken with Tenax with test
compounds and low levels of inorganic .gases present
(with GFF) 134
41 Chromatogram of background for 30 L air sample taken
with charcoal trap in the presence of high levels of
inorganics and no test model compounds 138
42 Chromatogram for 30 L air sample in the absence of
inorganics (high level study) and with test model
compounds 139
43 Chromatogram of 30 L air sample with charcoal trap in
the presence of high levels of inorganics and test
model compounds 140
44 Chromatogram of background for 30 L air sample taken
with charcoal trap in the presence of low levels of
inorganics and no test compounds 141
45 Chromatogram for 30 L air sample with charcoal trap
in the absence of inorganics (low level study) and
test compounds present 142
-------
FIGURES CONT'D.
Number Page
46 Chromatogram of 30 L air sampling using charcoal trap
and low levels of inorganics and test compounds
present 143
47 - Chromatograms of background for cryogenic trap for
30 L of air sample containing high levels of inorganics
and no test compounds 148
48 Chromatogram of sample collected with cryogenic trap
without inorganics (high level study) and test compounds
present 149
49 Chromatogram of sample collected with cryogenic trap
with high levels of inorganic and test compounds-
present • 150
50 Chromatogram of background for cryogenic trap with low
levels of inorganic gases and no test compounds ... 151
51 Chromatogr.am of sample collected with cryogenic trap
with test compounds only (low level interference
study) 152
52 Chromatogram of sample collected with cryogenic trap
with low level inorganic gases and test compounds
present 153
53 GC/MS/COMP profile of chloroprene in vehicle carrier
(xylenes) from permeation tube (old source) 184
54 GC/MS/COMP profile of chloroprene in vehicle carrier
(xylenes) from permeation tube (new source) 185
55 Schematic configuration used in calibrating the
ozone monitor - 189
56 GC/MS/COMP profile of background for 30 L air sample
from permeation/dilution systems with 340 ppb 0~, 320
ppb NO , 200 ppb SO , and 90% humidity present. Major
sources of hydrocarbons traced to NO supply 205
57 Chromatogram depicting background for solvent desorbed
unexposed NIOSH charcoal tubes . 206
XI
-------
FIGURES CONT'D.
Number Page
58 Mass spectrum of background contaminant from charcoal
tube 207
59 Mass spectrum of background contaminant from charcoal
tube. 208
60 Automatic sampler 218
61 Main control unit 220
62 Functional diagram of flow measurement processes 222
63 Flow schematic of the main control unit 223
64 Exploded view of sampling head block for sampling mode
(upside down) 224
65 Exploded view of sampling head block for transportation/
storage mode (upside down) 225
'66 View of top of sample head block - approximately to
scale 226
67 Tenax cartridges for automatic sampler 227
68 Sample holder .228
69 Heated sample cover for sample cartridge collector .... 229
70 4-Port Manifold 230
71 Muffler 231
72 Auto sampler electronic schematic 232
73 Time pulse generator 233
74 Flow rate integrator 235
75 Valve sequencer : 236
76 74164 8-Bit parallel-out serial shift register 237
77 Printer interface buffer 238
78 Flow sensor power supply 239
79 7404 Hex inverter. . . . '. 240
80 7476 Dual JK flip flop with preset and clear 241
81 7490 Decade counter as divided by 10 242
82 7490 Decade counter in cascade 243
83 74148 8-Line to 3-line encoder 244
84 75417 Dual peripheral driver 245
xii
-------
FIGURES CONT'D.
Number Page
85 Arrangement of P.C. boards on edgecard frame (looking
into connectors) 246
86 Schematic of diffusion tube and GC column 264
87 Modified diffusion tube system for generator vapor/gas
mixture 267
Kill
-------
TABLES
Number Page
1 Test Parameter Relationships for Evaluation of Collection
Devices 15
2 Test Compounds Selected for Use in Quantitative Evaluation
Studies of Sample Collection Methods 16
3 Recovery of Radiolabeled n-Hexadecane from Permeation System 28
14
4 Recovery of C-Bromobenzene from Permeation System .... 30
5 Experimental Design for Obtaining Measurements on Three
Levels of Concentrations (Zero, Low, High) at Three
Points in Time (t , t , t ) Using Three Container Types 35
6 Experimental Design for Determining Effects of Humidity,
SO., NO , and 0« on High Concentration Samples .... 37
£. A J
7 Chromatographic Parameters for Analysis of Containers ... 50
8 Operating Parameters for Thermal Desorption and GC/FID of
Tenax Cartridges 51
9 Operating Parameters for GC/FID and GC/ECD Analyses of
Solvent. Desorbed Charcoal Tubes 56
10 Concentrations of Inorganic Gases Employed in Interference
Studies 58
11 Comparison of Storage Environment on Bag Background .... 72
®
12 Effect of Storage Environment on Teflon Bag Background . . 81
®
13 Effect of Storage Environment on Tedlar Bag Background . . 81
14 GC Parameters for Analysis of Containers 89
®
15 Average Percent Recovery of Group I Compounds from Teflon
Bags 90
16 Average Percent Recovery of Group I Compounds from Tedlar
Bags 91
©
17 Percent Recovery of Group II Compounds from Tedlar Bags. . 93
18 Percent Recovery from Glass Bulbs 95
19 Relative Percent Recovery from Electropolished Steel Con-
tainers 98
xiv
-------
TABLES CONT'D.
Number Page
20 Relative Percent Recovery from Summa Polished Steel
Containers 99
21 Test Parameter Relationships for Evaluation of Trap Type
Collection Devices - Storage/Stability Study 102
22 Concentrations (High Level Study) and Total Quantity of
Group I Compounds Delivered to Sampling Devices:
Tenax GC, Charcoal, and Cryogenic Traps 103
23 Concentrations (Low Level Study) and Total Quantity of
Group I Compounds Delivered to Sampling Devices:
Tenax GC, Charcoal, and Cryogenic Traps 104
24 Percent Recovery for High Levels of Test Compounds from
Tenax GC Cartridges - With Correction for Breakthrough
Volume 105
25 Percent Recovery of Group I Compounds from Tenax GC
Traps - Low Level Study 107
26 Percent Recovery of Test Compounds from Tenax GC Traps -
Disregarding Breakthrough Volume (High Level Study) . 109
27 Percent Recovery of Group I Compounds from Charcoal
Cartridges 114
28 Percent Recovery of Group I Compounds from Cryogenic Traps
Storage/Stability Study 117
29 Percent Recovery of Test Compounds in the Presence of
®
Potential Interferences from Tedlar Bags 119
30 Percent Recovery of Test Compounds in the Presence of
Potential Interferences from Glass Bulbs 121
31 Percent Recovery of Test Compounds in the Presence of
Potential Interferences from "Summa" Polished SS Cans 124
®
32 Absolute Recovery of Group I Compounds from Tenax GC
Cartridges - Interference Study with Correction for
Breakthrough Volume 126
®
33 Relative Percent Recovery of Test Compounds from Tenax
GC Cartridges - Interference Study with Correction for
Breakthrough Volume 128
xv
-------
TABLES CONT'D.
Number Page
®
34 Relative Percent Recovery of Group II Compounds from Tenax
GC Cartridges - Interference Study, Uncorrected for
Breakthough Volume 130
35 Percent Recoveries of Group I Compounds from Charcoal
Cartridges in the Presence of High Levels of Inorganic
Substances 135
36 Percent Recoveries of Group I Compounds from Charcoal
Cartridges in the Presence of Low Levels of Inorganic
Substances 136
37 Absolute Percent Recovery of Group I Compounds from
Cryogenic Traps in the Presence of High Levels of
Inorganic Substances 144
38 Absolute Percent Recovery of Group I Compounds Using
Cryogenic Traps in the Interference Study 145
39 Relative Percent Recovery of Group I Compounds Using
Cryogenic Traps in the Interference Study 147
40 Peak Height Calculations 157
41 Area Count Calculation 158
42 Calculated Concentration of Test Compounds in Solvent Mix-
ture Used to Desorb NIOSH Charcoal Tubes - High
Level Study . 159
43 Retention Times for Group I Compounds Analyzed by TD/HRGC. 163
44 Retention Times of Group II Compounds Analyzed by TD/HRGC. 164
45 Retention Characteristics of Model Compounds Recovered
from Charcoal Cartridges Analyzed by GC/FID 165
46 Retention Times of Model Compounds Recovered from Charcoal
Cartridges and Analyzed by GC/ECD 166
47 Group I Compound Calibration Data for Bags, Bulbs and
Cannisters, February, 1981 168
48 Group II Compound Calibration Data for Bags, Bulbs and
Cannister, June, 1981 169
49 Group I Compound Calibration Data with Bags, Bulbs and
Cannisters for Interference Studies, July, 1981 . . . 170
xvi
-------
TABLES CONT'D.
Number Page
50 Group I Compound Calibration Data for Tenax Cartridge
Analysis - High Level 171
51 Group I Compound Calibration Data for Tenax Cartridge
Analysis - Low Level 173
52 Group II Compound Calibration Data for Tenax Cartridge
Analysis 175
53 Permeation Tubes Prepared and Available During the Past
Year 177
54 Historical Record of Permeation Rates for Group II Com-
pounds During Storage-Stability and Interference
Studies 179
55 Historical Record of Permeation Rates for Group II
Compounds During Storage-Stability and Interference
Studies 180
fi)
56 Breakthrough Volumes for Tenax GC Cartridge 186
57 Ozone Analyzer Calibration Results 188
58 Experimental Parameters for Testing Dilution System .... 190
59 Observed Levels for Benzene and Tetrachloroethylene from
Dilution System 191
60 Group I Compound Calibration Data Using Flash Method -
High Levels 193
61 Comparison of Calibrations Between Permeation System and
Flash Unit Methods - High Levels of Group I Compounds. 194
62 Comparison of Slopes for Standard Curves Between Permeation
Tube and Flash Unit Calibration Methods - Low Levels
of Group I Compounds 195
63 Instrument Calibration Using Different Sources of Benzene
and Tetrachloroethylene 197
64 Comparison of Percent Recoveries for Test Compounds from
®
Tenax GC Traps Using Two Different Calibration
Techniques 198
65 Calibration Data from Liquid Injection of Group II
Compounds 199
xvii
-------
TABLES CONT'D.
Number Page
66 Instrument Calibration Using Flash Unit Method 201
67 Comparison of Slopes for Standard Curves Between Permea-
tion Tube and Flash Unit Calibration Methods 202
68 Chronological Record of Quality Control and Quality
Assurance Procedures Performed During Storage-Stabi-
lity and Interference Studies 210
69 Specifications of the Automatic Air Sampler ........ 219
70 Switch Positions and Control Settings at Power Up of
Automatic Sampler 221
71 Flow Meter Calibration of Automatic Sampler - Serial/
Manual Mode i 249
72 Flow Meter Calibration of Automatic Sampler - Parallel/
Manual Mode 250
73 Flow Integrator Calibration of Automatic Sampler -
Parallel/Manual Mode . 251
74 Calibration of Dynamic Range for Flow Meter on Automatic
Sampler 252
75 Automatic Sampler Output-Parallel Mode 253
76 Automatic Sampler Output-Serial Mode 254
77 Initial Leak Test Results 256
78 Chromatographic Parameters Used in Storage Study 257
®
79 Background Level Study for Tenax GC Cartridges Stored in
Phototype Sampling Heads Used for the Automatic
Sampler 258
80 Background Levels of Tenax Cartridges Stored in Sampling
Heads Compared to Cartridges Stored in Culture Tubes . 260
®
81 Background of Tenax GC Cartridges Stored in Sampling Heads 261
®
82 Background of Tenax GC Cartridges Stored in Bolt-Type
Sampling Head 262
83 Compounds Examined in Modified Diffusion Tubes 265
84 Diffusion Tube Stability for Naphthalene - 4 Hour Study . . 279
85 Diffusion Tube Stability for Naphthalene, Quinoline, o-
Chloronitrobenzene - 7 Hour Study 271
xviii
-------
ACKNOWLEDGEMENTS
The authors wish to thank the following individuals for their contribu-
tions to this research program: Dr. D. Wagner, J. Harden, J. Bunch, R.
Green, M. Honeycutt, and D. Utterbach. A special thanks to Dr. L. Ballard
of Nutech Corporation for the design and fabrication of the automatic sampler.
The authors wish to thank also Mr. S. Kopczynski and K. Krost, and Drs. A.
Ellison and B. Dimitriades of ESRL, EPA, RTF, for their helpful suggestions
and criticisms throughout this research program.
xix
-------
SECTION 1
INTRODUCTION
The chemical characterization of atmospheric pollutants is of major
importance in determining primary sources, elucidating chemical transformation
pathways in the atmosphere and determining their potential risk to the
environment and the populace. Since chemical compounds present in the
atmosphere include both vapor phase organics (1) and particulate organic
matter (2), it is important that collection and analytical techniques encom-
pass the entire spectrum of substances.
Air pollutants are classified as primary or secondary. Primary atmos-
pheric pollutants are natural (e.g. dust, vegetation, etc.) or anthropogenic
in origin (e.g. smoke stack and vehicular emissions 1-3). Secondary products
are generated from primary pollutants, i.e. via atmospheric (photochemical)
reactions.
The amount of vapor phase organics emitted anthropogenically in the
13
United States has been estimated at 1.9 x 10 g/year. World wide emission
13
has been estimated to be about 7.5 x 10 g/year. The levels of vapor-
phase organics are generally 10-50 times greater than particulate organics
(4).
Most atmospheric pollutant samples are extremely complex in nature
containing perhaps hundreds of different molecular species with a very large
dynamic range of concentration. This complexity may necessitate the use of
high resolution techniques such as capillary column gas chromatography
and/or high performance liquid chromatography. Pre-separation techniques
may be required and also methods which provide for definitive chemical
analysis of samples such as mass spectrometry (5-12).
The successful chemical analysis of the vapor phase organics depends
upon a number of important steps beginning with the collection and/or precon-
centration techniques. During the past decade significant advances have
-------
been made in the development of collection methods for vapor-phase organic
materials in the atmosphere. Collection methods now available include those
techniques which trap organic vapors on sorbent surfaces (6-13), condense/
freeze vapors in cryogenic traps (14,15) or confine the pollutants in evac-
uated stainless steel cannisters or bags (16,17).
One of the major problems encountered in sampling vapor-phase organics
is the presence of water. The relative abundance of water in the atmosphere
4
is high (often greater than 10 fold) compared to that of organic species of
interest and many of the collection and analysis methods (such as gas chroma-
tography columns) do not tolerate large quantities of water. Simultaneous
concentration of water with organic material causes partial dilution of the
samples thus impeding sensitive analysis of the vapor-phase organics.
Sample volatility also limits the ability to concentrate analytes in subse-
quent steps.
Several primary criteria for evaluation of collection devices for air
sampling of vapor-phase organics have emerged from previous studies: (a)
ability to discriminate against water and preferentially concentrate the
vapor-phase organics of interest; (b) low background contribution from the
sampling media during subsequent analysis; (c) minimal decomposition or
polymerization of the sampled constituents during collection and recovery;
(d) quantitative collection efficiency and recovery of trapped or confined
vapors; (e) high breakthrough volumes for sorbent-based collection devices,
and (f) collection systems that do not contribute to in situ formation of
artifacts. (5-13).
Although a number of collection devices have been reported and applied
by researchers, there are no reports in which a comparison has been made of
the various collection devices. The strengths and weaknesses of each of the
collection devices (containers and traps) have not been evaluated in a
concerted and thorough fashion in terms of the primary criteria outlined
above. None of the methods which are currently in use for collecting gaseous
atmospheric samples have been shown to be completely satisfactory for all
chemical species at all concentrations. Limits of applicability are not
well-defined and is the subject of this research report. Plastic bags out-
gas residues from the plastic film and are subject to diffusion through
-------
the bag walls. Certain chemicals species may be adsorbed on the walls of
glass or stainless steel containers or undergo reactions at active sites on
the walls. However, these factors have not been thoroughly compared among
the various collection devices. Concentrating gaseous samples in cold traps
is a cumbersome procedure and is not well suited for field sampling by
unskilled personnel at the present time. The relative high concentrations
of water in the atmosphere may plug the trap, interact with certain pollutants
causing artifacts or interfere with subsequent analysis. Traps packed with
adsorbent materials and operated at ambient temperature do not efficiently
collect the more volatile pollutants and the strongly adsorbed pollutants
may be difficult to remove from the adsorbent trap. Long term storage of
gases in any case may result in losses of pollutants through diffusion or
slow reactions from other chemical species or with materials of the container.
When this research program was initiated, a sampling system for sorbent
cartridges capable of collecting several sequential replicate or single
samples (unattended) was not commercially available. The need for an automa-
tic sampler which could collect organic vapors from ambient air over a
prescribed flow range (e.g. 5 mL/min to 1.5L/min) and sampling period inter-
vals (e.g. 15 min to 24 hr) precipitated a design and fabrication effort.
The U.S. Environmental Protection Agency has been concerned with these
analytical problems and a program to test and evaluate various methods of
collecting and analyzing gaseous atmospheric samples for a variety of toxic
organic pollutants by gas chromatography was performed and is reported here.
By defining the limitations and applicability of the techniques, a more
comprehensive approach to the analysis of organic pollutants may be formulated
for future studies.
-------
SECTION 2
CONCLUSIONS
Three type of polymeric bags (FEP Teflon , Tedlar, and five-layer poly-
ethylene-aluminized), glass bulbs, stainless steel cannisters (electro and
Summa -polished), and Tenax GC, charcoal and nickel cryogenic traps were
evaluated for: (1) simplicity and convenience; (2) collection and recovery
efficiency for GC analysis; (3) accuracy, reproducibility and limits of
detection; (4) analyte storage stability; (5) potential interferences from
inorganic gases [ozone, NO , SO-] and water, and (6) limits of applicability
[chemical/physical properties of chemicals, background, field conditions].
Because of a limitation in resources a rigorous statistical evaluation of
data addressing these six issues could not be conducted, and thus, in many
cases only qualitative trends can be described.
In order to test the various collection methods, a permeation/dilution
system was designed and fabricated for this research program. The general
difficulty in the use of this system was attributed to potential adsorptive
losses of the chemicals of interest at highly dilute levels. Adsorption
studies were conducted with radioactive dimethylamine, hexadecane, and
bromobenzene. Adsorptive losses were essentially 100% for dimethylamine at
the low ppb level while recoveries were 88-90% for bromobenzene. Attempts
to deactivate the glass surfaces in the system with silanizing agents and a
Carbowax 20M treatment were not successful as tested by the transmission of
dimethylamine. Approximately 40-50% of the radioactive hexadecane passed
through the system when the system was maintained at 150°C. Thus, chemicals
with boiling points higher than bromobenzene would require higher temperatures
to minimize adsorptive/condensation losses to the surface of the system.
The system was operated at 200°C.
Commercial sources of reference standard or NBS certified synthetic
air/vapor mixtures were not available for most of the chemicals when this
-------
program was initiated. None of the sources of chemicals used here for the
preparation of a dynamic flowing synthetic air/vapor mixture were traceable
to a reference standard. As such, all synthetic air/vapor mixtures were
synthesized utilizing a permeation tube concept. The permeation tubes for
many of the 27 test compounds exhibited permeation rates within +10% (RSD)
over a several month period. Permeation tubes of chloroprene, 1,2-dichloro-
propane, 1,1,2,2-tetrachloroethane, a-epichlorohydrin, and nitrobenzene
exhibited permeation rates with variability >+10%. Thus in the evaluation
of the collection methods, the uncertainties associated with the absolute
accuracy and reproducibility should be noted in drawing conclusions from the
reported data.
Although bags have the advantage of allowing 10-100 L of samples to be
collected for replicate measurements, they are easily punctured and clear
bags must be protected from light after sample collection. Thorough cleaning
to remove volatile organic background can be complicated since the bags
cannot be heated excessively without the seams developing leaks. Cleaning
with ozone and ultraviolet light was necessary to reduce levels of high
®
boiling contaminants in Teflon and Tedlar bags. .The background level in
the five-layered polyethylene-aluminized bags was so severe and unacceptable
after attempted cleaning, that they were deemed unsatisfactory for environmen-
tal sampling and were not further evaluated in this program.
Recoveries for 15 test compounds collected from a dynamic flowing
synthetic air/vapor mixture for Teflon and Tedlar bags were generally in
the range of 70-100%. The highest recoveries were found for the most volatile
chemicals in the synthetic air/vapor mixture while recoveries decreased with
the less volatile substances. However, with time the decrease in recovery
®
was generally more rapid with Teflon than with Tedlar bags. Both types of
bags exhibited both loss of compound and influx of contaminants by permeation
®
through the bag walls. Teflon and Tedlar bags should be stored in clean
environments or should be analyzed within 4 h after sample collection.
The potential interferences from inorganic gases present during the
sampling of test compounds decreased the recovery of most test compounds
from Tedlar bags. One adverse effect caused by the presence of inorganic
-------
gases was the release of unknown contaminants from the wall of the Tedlar
bags which appeared as background during analysis.
The amount of sample collectable in a glass bulb is limited with usually
1-2 L available for analysis. Glass bulbs were easily broken especially
during the filling step of the bulb. Bulbs should also be protected from
light after sample collection. Bulbs must be cleaned and this can be facili-
tated by heating while evacuating which improves efficiency of the cleaning
process though care should be taken not to heat the stopcock valves employed.
Cleaning with a solution was difficult and time consuming since the valves
must be disassembled and this operation leads to a high incidence of breakage.
The recovery of test compounds from glass bulbs decreased rapidly with
increased boiling point and was above 90% for only a few test compounds.
The sampling of test compounds in the presence of potential interferences
from inorganic gases generally decreased their recoveries over those obtained
with low level interferences.
As with glass bulbs, steel containers allow recovery of only limited
sample volume, typically 4-6 liters. They are however extremely rugged and
®
could be cleaned thoroughly by heating while evacuating. The Summa polished
containers generally show higher recoveries for high boiling compounds than
electropolished containers. Also the Summa polished containers exhibited a
better maintenance of recovery with time than the electropolished containers.
Sampling in the presence of high level of inorganic gases as potential
interferences decreased recoveries for some compounds and increased them for
others. These increases may be due to further displacement by water of the
test compounds and/or release of contaminants not released during cleaning.
The NIOSH charcoal cartridges evaluated in this program were found to
be generally inadequate as applied to the sampling of environmental levels
(low ppb) of test compounds. The limits of applicability of charcoal cart-
ridges are revealed in the overall poor recovery of organics. For the
analysis of test compounds, GC/FID and GC/ECD were employed and only when
using GC/ECD were the limits of detection adequate for some of the chemicals
of interest. None of the test compounds collected in the low ppb range and
a 30 L sample volume were detected by GC/FID. The recovery of chemicals
-------
measurable by ECD was poor and the precision was erratic which prohibited
the establishment of storage characteristics.
Initial storage-stability studies with nickel cryogenic traps (as
prescribed by EPA) cooled with dry ice yielded poor recoveries for all
chemicals at the low ppb level (1 L/min sampling rate). The cryogenic traps
were then modified by filling with clean glass beads and liquid oxygen was
used as the cryogen. Most of the compounds were detected but the absolute
recoveries were still low and the precision was poor. The applicability of
cryogenic traps of the design evaluated in this study was limited. This
method was labor intensive during sample collection, sample transfer and
storage, and sample recovery and analysis. Cryogenic traps were the least
convenient of the collection methods described.
®
The Tenax GC sampling cartridge was limited principally in the break-
through volume which directly determined the detection limits obtainable for
given measurement techniques. In this study a 30 L sampling volume was
employed and thus the breakthrough volumes for chemicals .that are less than
the sampling volume will severely limit their detection and quantification.
Similarly, the collection efficiency was directly related to the breakthrough
volume. Recoveries of chemicals were not significantly decreased by short-
term storage (7 da). The precision of recoveries was slightly less than
those observed for containers; however, with Tenax GC cartridges, the recovery
was based upon triplicate sample analysis and not measurement of the same
sample. A major attribute of a cartridge sampling concept is its simplicity
and convenience in its preparation, sample transport, and recovery and
analysis. It is one of the few techniques which is amenable to personnel
sampling. Large numbers of samples can be taken simultaneously, stored
until analysis and analyzed relatively rapidly.
Experiments with potential inorganic gas interferences demonstrated a
major problem which can occur with any collection device. Reactive inorganic
gases in the atmosphere can perturb the quantitative and qualitative composi-
tion of the air sample (parent compounds disappearing and new artifacts
appearing) during collection of organics. Substantial improvement without
absorptive loses was obtained by using a very small amount of mild reducing
®
agent to remove ozone prior to trapping, :L.e. , Tenax GC sampling.
-------
The following table exhibits the relative performance of the collection
devices for the sampling of synthetic air vapor mixtures with and without
the presence of inorganic gases in the parts-per-billion levels of test
organic compounds:
No. of Compounds in Recovery Range
Vjuj.xci.i-a.uu iicuuuu
(No. Compounds Tested)
Teflon® (15)*
Tedlar (27)*
Tedlar (27)**
Glass bulbs (27)*
Glass bulbs (27)**
Summa polished SS cannister (27)*
(R)
Summa polished SS cannister (27)**
Tenax GC cartridge (27)*
Tenax® GC cartridge .(27)**
>95%
4
5
2
4
2
10
4
12
11
90-95%
0
1
1
2
3
5
1
3
3
80-90%
2
6
10
2
5
2
4
4
0
70-80%
5
6
3
5
5
3
4
0
2
60-70%
1
3
2
6
2
1
2
2
0
*ppb levels of test compounds sampled and then analysis immediately
conducted.
**ppb levels of test compounds in the presence of low levels of inorganic
gases sampled and then analysis immediately conducted.
The two sampling techniques which show the greatest promise under all
®
of the laboratory tests conducted are the Summa polished stainless steel
(R) ' ®
cannisters and the Tenax GC sampling cartridge. The Summa polished can-
nister gave the highest recoveries for the more volatile chemicals in the
®
test group, while the Tenax GC sampling cartridge performed better for
those chemicals with breakthrough volumes larger than the sampling volumes.
Thus, these two collection techniques can compliment each other when samp-
ling a broad spectrum of vapor-phase organics in the atmosphere is desired.
Support coated open tubular (SCOT) capillaries were employed for re-
solving the test compounds. The stationary phase was SE-30. The SCOT
capillaries performed adequately in these studies. After much of the re-
search on this program had been performed with SCOTs, fused silica capil-
laries became commercially available; however, these was no opportunity to
evaluate their performance.
-------
The performance of a newly designed and built automatic sampler was
laboratory tested. Both in the parallel and serial modes (1 to 6 channels
in duplicate and 1 to 12 channels, singly) the stepping sequencer was found
to step through each channel in the proper order. Also, the sampler was
tested for sampling periods at 15, 30, 45 and 60 min, and 2, 12 and 24 hr
and was found to step through the channels correctly at the proper time
intervals.
The sampler's digital flow meter was calibrated to read the actual flow
rate being sampled by adjusting zero and gain controls. The sampler set and
actual flows (0.19% relative standard deviation, N = 6) agreed within 2% in
the parallel and serial modes. Initial evaluation of the printout from the
integrator (total volume) and its agreement with the actual volume sampled
revealed a 20% and 3-5% deviations in the serial and parallel modes, respec-
tively. After calibrating the integrator both modes were within 5%. No
significant drift was detected between the digital flow meter/integrator
registered volume and actual volume over 24 hr of operation.
Various sealing designs (phenolic screw cap and bolt-type arrangement
with Teflon or Viton o-rings) in the sampling heads were tested and the
bolt-type was superior. No leaks were found at 10 psi of helium when Viton
®
o-rings were used. Tenax GC sampling cartridges stored in bolted end plate
sampling heads exhibited background levels typical of cartridges stored in
culture tubes.
-------
SECTION 3
RECOMMENDATION
Upon conducting a number of laboratory tests using test compounds and
synthetic air vapor mixtures, a few collection devices have been found to be
sufficiently accurate with synthetic mixtures to be further tested with
atmospheric samples. The recommendations offered here specifically address
®
the further testing of Tenax GC sampling cartridges, "Summa" polished cans
and glass bulbs. Field studies should include at a minimum three industrial
sites and one rural area for the additional testing of these three collec-
tion devices. It is recommended that field testing protocols be developed
prior to the collection of field data and that these protocols incorporate
the following elements:
(1) the use of 20 or more deuterated surrogate test compounds selected
from those previously studied in laboratory experiments for the
evaluation of the collection devices;
(2) quantitative and qualitative analysis using capillary gas chroma-
tography/mass spectrometry;
(3) collection of triplicate samples at each field site;
(4) collection of a set of both day and night samples;
(5) the incorporation of quality control procedures, e.g_. , blanks,
etc.
(6) determination of potential interferences of each test compounds
where possible, and
(7) calculation of accuracy and precision of collection and analysis
methods.
The use of deuterated surrogate compounds allows for the distinction
between endogenous and exogenous pollutants when mass spectrometry is used
as the measurement technique, as well as the differentiation between endo-
genous' and artifactually formed compounds. Thus, by spiking the atmospheric
10
-------
samples in a continuous fashion with deuterated compounds, the true collec-
tion accuracy can be assessed as well as the reproducibility of analysis.
The absolute and relative recovery should be determined by the use of
internal standards.
It is recommended that the development and employment of quality control
and assurance practices be established in developing quality data and to
ensure the validity of the results obtained from the field evaluation for
these three collection devices. These quality control and assurance practices
should include at a minimum the following:
(1) gravimetric calibration of permeation tubes and statistical analysis
of permeation rates;
(2) GC/MS instrument calibration checks [e.g. instrument performance
and chromatography column performance], and
(3) flow calibration of pump systems used in the field studies.
In addition to the field testing of these three collection devices
further research is recommended regarding the automatic sampler. These
recommendations are in (1) design refinements and (2) the field evaluation.
Recommended design refinements include:
(1) removing the pump from the control console with a quick connect
coupling to the console making the console smaller, lighter and
cooler;
(2) placing the calibration port in front of the variable orifice
valves (i.e. parallel to the inlet-ports) so that the flow rate
through each channel can be set easier prior to the beginning of a
run, and
(3) the provision of external battery jacks on the front panel. A
feature for providing power to the timing circuitry for cases
where line power outages occur and print-out of the information in
the recording registry can be made.
Further laboratory and field testing of the automatic sampler should
include:
(1) determining the accuracy and reproducibility of sampling synthetic
air/vapor mixtures from a permeation dilution system as described
in this report (with the variables of sampling time and rate
evaluated);
11
-------
(2) long-term storage (months) of clean Tenax GC cartridges in
sampling heads to determine the background which may occur;
(3) the effect of transportation on background of the blank Tenax
cartridge in the sampling head.
(4) determine accuracy and reproducibility under atmospheric (field)
sampling conditions to include (a) sampling in light and heavy
®
particulate loads, (b) a comparison of Teflon vs. glass fiber
filter, (c) sampling times and rates, and (d) sampling under heavy
and light vapor-phase pollutant loads.
(5) determine its reliability under (a) extreme weather elements such
as temperature and humidity, and (b) power transcient effects
occurring during stormy weather and recovery of the sampling
system, and
(6) evaluate the automatic sampler using deuterated surrogate standards
@
in a field sampling protocol design as developed for the Tenax GC
sampling cartridge described above.
12
-------
SECTION 4
PROGRAM OBJECTIVES
The primary objective of the research project has been the comprehensive
evaluation arid testing of six collection devices and the particular analytical
procedures associated with each of these devices. The component objectives
of this study have included: (1) the preparation of a comprehensive program
design.for the evaluation and testing of the six devices (polymeric bags,
glass bulbs, metal containers, Tenax and charcoal absorbent traps, and Ni
cryogenic traps); (2) the selection of model compounds for systems evaluation;
(3) the preparation and calibration of permeation tubes; (4) the design and
performance testing of a portable permeation system; (5) the delineation of
sampling and analysis procedures; and (6) the design and fabrication of an
automatic sampler.
Although not a specific initial objective of this program, some experi-
ments on optimization of collection devices (e.g. polymeric bags, glass
bulbs, and nickel cryogenic traps) were needed and conducted as required by a
subsequent technical directive. Also, design changes of the automatic
sampler were instituted as specified in a technical directive.
COLLECTION DEVICE EVALUATION
As stated, the primary objective of this research project has been the
evaluation of six gas collection devices. The elements of the evaluation
included determination of the following:
(1) limits of applicability;
(2) collection efficiency;
(3) recovery (transfer) efficiency for gas chromatographic analysis;
(4) analytical accuracy and detection limits;
(5) effect of potential interferences (including ozone, S09, NO , and
^_ X
water vapor);
13
-------
(6) sample stability and storage;
(7) quality of chromatograms;
(8) simplicity and convenience of the sample collection and transfer
methods.
The objective included not only evaluation of each of these elements for
each collection device but also a comparison of the evaluation results
obtained with each device so that the overall "best" devices might be selected
for field testing in a future project.
The evaluation was to be performed using a variety of test conditions.
Mixtures of test compounds were to be collected under various conditions.
Using these compounds, the test parameter relationship for the collection
and .storage experiments to be used are given in Table 1.
In experimental design (A) the sampling volume and the relative humidity
(30%) were to be held constant. No potentially interfering substances (0~,
SO , NO ) were to be added. The variable parameter was to be the concentra-
ff X
tion of each individual substance and storage time. In this case, the
concentrations were to be zero, not less than 10 ppt (low) and not more than
100 ppb (high). In the second experimental design (B), the ozone, SO , and
NO concentrations, and relative humidity were to be varied while the sampling
volume, time and rate, and concentrations were to be constant. All sampling
devices were to be evaluated simultaneously by sampling the test atmosphere
at the same time to reduce possible variability in the performance of the
permeation system.
SELECTION OF MODEL COMPOUNDS
The experimental design incorporated an approach which provided for a
quantitative comparison of the various sampling methods. A set of 27 test
compounds was selected for this study and it is given in Table 2. The
initial criterion for selection of test compounds was based upon the produc-
tion level in the U.S. and their toxicity data. These data are incorporated
as provided by the U.S. EPA.
For each chemical group, several compounds were included to provide a
range of chemical and physical properties which would reveal the strengths
and weaknesses of the collection devices. For chloroalkanes, methyl chloride
represented the most volatile organic (b.p. -24.2°C) and 1,1,2,2-tetrachloro-
ethane (b.p. 146.2) as the least volatile (Table 2). In addition to a wide
14
-------
Table 1. TEST PARAMETER RELATIONSHIPS FOR EVALUATION
OF COLLECTION DEVICES
Experimental
Design
Test
Parameters
Constant
Parameters
Concentration
(Two levels > 10 ppt < 1 ppb
and > 1 ppb < 100 ppb)
Storage Time (0, 3, 7 da)
Volume, sampling time
and rate
RH = 30%
[03] = 0
[S02] = 0
[NO ] = 0
x
03/NOx/S02
Test mixture concen-
tration
Volume, sampling time
and rate
Relative humidity
15
-------
Table 2. TEST COMPOUNDS SELECTED FOR USE IN QUANTITATIVE EVALUATION STUDIES OF
SAMPLE COLLECTION METHODS
Chemical Group
Chloroalkanes
Chloroalkenes
Aromatics
Alkanes
Nitro compounds
Phenols
Acrylo compounds
Ethers
Sulfur compound
Compound
Methyl chloride
1 ,2-Dichloropropane
Chloroform
1,1,1-Trichloroethane
Vinyl chloride
Tetrachloroethylene
2-Chloro-l,3-butadiene
1 , 1-Dichloroethylene
Allyl chloride
ra-Dichlorobenzene
Benzyl chloride
Benzene
Toluene
1,2,3-Trimethylbenzene
Ethylbenzene
o-Xyleue
n-Decane
Nitrobenzene
o-Cresol
Acrylonitrile
Furan
Bis-(2-chloroethyl)ether
Propylene oxide
or-Epichlorohydrin
Methyl mercaptan
B.P.
-24.2
96.4
61.7
146 2
74.1
-13
121
59.4
37
45
132
173
215
80.1
110.6
176.1
136.2
139.1
174.1
210.8
190.9
77
31.4
178
34.3
116.5
6.2
Estimated U.S."
Production (106 Ib/yr)
460
(30)
260
75
4180
. 680-1210
349
260
290
690
80
1400
6940
-
-
1000
-
550
(30)
1410
-
1 (?)
-
500
-
Toxicity"
Carcinogenicity Mutagenicity Acute
+ H
H
+ - H
+ +11
• « ' tl
M
+ weak + H
+ - M
+ + H
+ «• L
1 + H
* * H
+ + H
r - M
M .
+ t
H
Promoter (?) M
+ +11
H
+ t- H
+ M
+ + H
M
Quantities in brackets are estimated from reported production of mixtures containing the com-
pound. Symbols (+ and -) indicate test results, +_ indicates uncertainty. Data provided by
EPA.
-------
range in boiling points, other compounds were selected for their unique
chemical reactivity. For example, 1,1,1-trichloroethane under certain
catalytic conditions will decompose to vinylidene chloride.
Among the chloroalkenes (Table 2), the volatility range was -13°C
(vinyl chloride) to 121°C (tetrachloroethylene). 2-Chloro-l,3-butadiene
(chloroprene), an isomer of vinyl chloride ("1-vinyl vinyl chloride")) is
highly reactive toward self-polymerization and destruction by ozone. Allyl
chloride is similarly reactive, but also can readily decompose to allyl
alcohol. All of the chloroalkenes exhibit varying degrees of reactivity
with ozone.
Among the chlorinated aromatics (Table 2), chlorobenzene represented
the most volatile (132°C) while benzyl chloride is the least volatile (215°C)
of all chemicals tested. Benzyl chloride is also a reactive species in the
presence of moisture and thus would be difficult to collect and store accu-
rately.
Benzene and 1,2,3-trimethylbenzene define the vapor pressure range for
the aromatic hydrocarbons (Table 2). Some sensitivity to ozone has been
shown for these chemicals, the extent of collection problems was to be
tested.
The recovery of nitrobenzene and phenol are suspect with polymeric bags
and metal containers and thus were included in this study. Also, background
®
from Tenax GC via ozone exposure producing phenol was to be studied.
The remaining substances (acrylo and sulfur compounds and ethers) are
all reactive in the presence of ozone. The stability of Bis-(2-chloroethyl)-
ether and a-epichlorohydrin during their collection in the presence of
moisture (particularly with HNO ) was also of interest. The ease of oxidation
of methyl mercaptan to dimethyl disulfide was to determine the reactivity of
the compound after its collection with inorganic substances.
Thus, the rationale for selecting the compounds in Table 2 is generally
obvious. Compounds which were polar and with low vapor pressures were of
interest since these substances have a propensity to adsorb to surfaces
making quantitative recovery often difficult. Also, they will, when atmos-
pheric conditions are optimum, partition between the aerosol and vapor-phase
17
-------
states. This is particularly important when filtering media are. used to
remove aerosols/particulates while collecting vapor-phase compounds.
The test compounds listed in Table 2 were used throughout the entire
(§) ®
program for the evaluation of FEP Teflon , Tedlar (polyvinyl fluoride) and
five-layer aluminized plastic bags; glass and metal bulbs; low temperature
®
condensation traps, and charcoal and Tenax GC adsorbent traps. The use of a
single set of test compounds for evaluating all the collection methods was a
major objective since a quantitative comparison between devices was desired.
PERMEATION AND DIFFUSION TUBES
In order to facilitate the evaluation of collection devices, it was
necessary to synthesize a continually flowing multi-component vapor/air
mixture. The accurate and reproducible synthesis of air/multi-component
vapor mixtures had not been reported prior to the initiation of this program.
Certified sources of organic vapors in air or permeation tubes for all the
chemicals in Table 2 were not available. Thus, the objective of this study
was to devise a means of delivering a flow of air/vapor mixture of a known
concentration.
To accomplish this, permeation tubes were fabricated and gravimetrically
calibrated in this laboratory. Permeation tubes were designed to yield
permeation rates for the chemicals in Table 2 to not exceed a range of 50
fold. The details of fabrication and calibration are given in Section 7.
For chemicals with lower vapor pressures than those listed in Table 2,
the development of diffusion tubes was an additional objective of this program.
This is discussed in Section 10.
PERMEATION/DILUTION SYSTEM
In order to evaluate the collection devices, accurate and reproducible
atmospheres of the model compounds were needed. Subsequently one of the
objectives of the project was to construct a permeation/dilution system,
using permeation tubes and diffusion tubes as sources, to deliver model
compounds (1-20 simultaneously) test atmospheres in the range of 10 ppt to
100 ppb. The system was to be designed for laboratory and field experiments
and thus be portable. Also, the system was to include a source of "zero"
dilution air and was to be designed to minimize loss of the test compounds to
the system component walls.
18
-------
SAMPLING AND ANALYSIS PROCEDURES
As stated, the primary objective of the project was to evaluate the six
sample collection devices described previously. This evaluation of the
devices cannot be separated from the methods of collecting the samples with
the devices and then recovering and measuring the analytes. Another objective
then was to develop and/or use the collection, recovery and measurement
procedures which are appropriately suited to the collection devices themselves.
Also these methods and the collection devices were to be "user friendly" if
possible; that is, their use should be practical and within the capabilities
of most appropriately trained analysts.
AUTOMATIC SAMPLER
A final major objective of this program was to develop an automatic
sampler to be used in an unattended fashion to collect vapor-phase organics
on adsorbent-type sampling traps. Specifically, the sampler was to be
designed so that it could be preprogrammed for prescribed sampling times and
rates, sequentially sampling over long periods of times (a maximum of 24 hr
per sample), and the collection of single or duplicate samples. A print-
out status report for each sample (date, time, volume collected) was to be a
feature. The design and fabrication was to include laboratory testing of
the sampler and the writing of an operating and maintenance manual.
-------
SECTION 5
DESIGN AND FABRICATION OF PERMEATION/DILUTION SYSTEM
The generation of accurate and reproducible test atmospheres was essen-
tial for testing and evaluation of the six collection devices. Prior to
initiation of the project, it was proposed that such atmospheres could be
produced using permeation and/or diffusion tubes in a permeation/dilution
system. The permeation and diffusion tubes were chosen as compound sources
as other sources, e.g., compressed gases in cylinders, were not available
for the variety of compounds to be used in the study. The permeation and
diffusion tubes could be prepared in the RTI laboratories without great
difficulty if they were not available from commercial sources. A system was
designed to deliver test compound atmospheres in the concentration range at
10 ppt to 100 ppb. The system was intended for both laboratory and field
experiments and thus was to be portable. The principle components of the
system were to be clean ("zero") air source, the permeation chamber and
three dilution stages. The permeation chamber was to be large enough to
accomodate about 20 permeation tubes so that mixtures of test compounds
could be generated.
SYSTEM FABRICATION
Two systems were constructed. The systems were essentially identical
except one contained two dilution stages while the other contained three
dilution stages.
A schematic of the system with three dilution stages is shown in Figure 1.
In this system, gases emitted from permeation tubes were diluted in a series
of steps which each involve removal of a portion of the gaseous sample and
addition of diluent (air). Each system was constructed using two Marinite-
®
XL boxes covered with 0.5 in foil-coated polyurethane foam. The smaller
box (Fig. 2) encloses the permeation chamber and a purged storage chamber
for the permeation tubes not in use in the chamber. The temperature of
20
-------
VACUUM
PUMP
BENDIX
AIR
DRYER
IPS
_o
p
BALLAST
TANK
NEEDLE
VALVE
MANIFOLD
FURNACE
Figure 1. Overall schematic of permeation system.
21
-------
c
I HEATER!
AUX. ^
INLET C p
OUT
AIR
IN
Figure 2. Permeation chamber, storage vessel, and enclosure.
22
-------
these permeation tubes was originally controlled through control of the
temperature of the air in the box using two ceramic heaters controlled by a
Valco precision controller. This was found to be insufficient and the
permeation tube chamber was rebuilt so as to be water jacketed. Circulating
water at 30.0 + 0.1°C was provided via a Haake water bath and pump system.
The temperature of the dilution air passed into the permeation chamber was
controlled by first passing it through several meters of copper tubing
setting in the aforementioned water bath. The storage chamber was shock-
mounted and was constantly purged with clean air. The purged air was drawn
off by the vacuum pump and passed through the catalytic cleaner. The large
diameter of the permeation chamber allows several tubes to be inserted at
the same time so that complex mixtures may be generated. An auxiliary
injection port was added on the outlet of the permeation chamber to allow
the injection of radioisotope tracers for the initial studies and also to
allow gas mixtures from tanks to be bled to the dilution bulb for further
dilution.
The larger component (Fig. 3) houses the three dilution bulbs. These
1 L bulbs were interconnected by 20/12 spherical ground glass joints. A
1000 watt heater and an Omega proportioning heat controller maintained the
temperature in the box up to 150°C. .A glass manifold distributed the gas
mixture to six 0.25" glass sampling ports which passed through the wall of
the box. A larger vent port with 25/12 joint passed through the side of the
box for disposal of excess mixture.
The permeation system was designed to be portable. As such, it contained
its own clean air supply. Air entered the system through the compressor of
the Bendix Model 8833 heatless air dryer. The dry air at 60 psi passed into
a 12 L ballast tank and then flowed through a platinum/palladium catalyst at
300°C for the oxidation of hydrocarbons. The output from this furnace is
allowed to cool by flowing through a 3.5 m length of copper tubing before it
reaches the flow controllers. The air is used to dilute the gases given off
by the permeation tubes in the permeation chamber. The clean air flows
through the chamber after passing through Tylan flow controller (FC) #1, and
this mixture then goes into the first dilution stage. An expanded view of
the inlet of the dilution bulb is shown in Figure 4.
23
-------
VENT
—TU
tJ
HEAT
CONTROLLER
Figure 3. Dilution bulbs and heated enclosure.
24
-------
Figure 4. Dilution bulb inlet.
25
-------
A known portion of the gas mixture could be drawn off at (A, Fig. 4)
and passed through the needle valve and flow meter. A known flow of makeup
air could be added at (B) for a further dilution. The gas mixture plus
dilution air then flows into the mixing bulb. The withdrawal/addition
process could be repeated at the next two bulb junctions. Although an
infinite dilution of the original gas mixture is theoretically possible, the
practical dilution limit of this system was originally considered to be 10 .
The entire process can be illustrated through an example. If a permeation
tube gives off a compound at 1 (Jg/min, and the flow across the tube from FC
#1 is 1 L/min, the concentration at the outlet of the permeation chamber if
1 M8/L- If 990 mL/min of this flow is taken at point A through FM #1 and
990 mL/min air is added at B, the mixture in the first dilution bulb should
be 0.01 |Jg/2 concentration. This dilution of 1:100 at each stage was origi-
nally considered the maximum that should be attempted with this system. If
this maximum dilution were achieved at each of the three stages, the final
output concentration will be one one-millioneth of the concentration at the
outlet of the permeation chamber. After some initial trials with the system,
this dilution factor was found to be overly optimistic. The principal
difficulty arises from reproducibly withdrawing a large quantity of test
atmosphere at each dilution stage. Needle valves rather than electronic
flow controllers were used to control test atmosphere removal as flow control-
lers do not work well with less than 10 psi difference across the input to
the output sides of the controller. Greater than 10 psi pressure drop is
very difficult to maintain when withdrawing large volumes of gas, unless an
especially high volume vacuum pump is used. Needle valves are reproducible
to about +1 percent. Following the example given above, 990 + 10 would be
withdrawn, leaving great uncertainty in the amount remaining. Subsequently,
it was concluded that factors of 10-20 (maximum) per dilution stage were
much more practical. Again following the example, 900 + 10 would be withdrawn
leaving 100 + 10 mL to be diluted to 1000 ml for a dilution factor of 10.
Also, the Tylan flow controllers and meters were compared to NBS traceable
bubble flow meters and found to be 10-20% off in all cases. They were then
calibrated using the bubble flow meters as standards rather than attempting
26
-------
to make a correction for the type of gas being used or some other mathematical
correction.
An initial problem was experienced with the air compressor/dryer enclo-
sure which caused overheating of the pump motor. The compressor/dryer was
subsequently operated without an enclosure.
VALIDATION OF PERMEATION/DILUTION SYSTEM
Radiolabeled compounds were employed for validation of the portable
permeation system. Initial testing was conducted to determined the extent
of condensation in the permeation system of a non-polar high boiling compound
at relatively high concentrations (>100 ppb). For this purpose, n-[l,2
(n)- H] hexadecane with a specific activity of 4.86 x 10 dpm/g was procured
from Amersham Corporation. n-Hexadecane (b.p. 287°C) has a density of 0.773
g/mL, so that 4.7 pL injected was equivalent to 3.6 mg which produced 16,600
dpm. A TriCarb scintillation counter (Packard Inst., Chicago, IL) was
used.
The permeation system was assembled and calibrated; the permeation
chamber temperature was brought to 30°C, and the secondary dilution system
brought to 65°C. The hydrocarbon free air flowed at 250 mL/min. Four
midget impingers each containing 10 mL of Triton X scintillation counting
solution were connected in series downstream of the dilution for collection
of the labelled compounds. Then 4.7 |JL of radiolabeled n-hexadecane was
introduced into the system using a 10 (JL syringe (Run A). Collection was
effected for 60 min, at which time a new set of impingers was put in line.
The temperatures of the permeation chamber and the dilution system were
increased to 60°C and 93°C, respectively, and another 4.7 (JL of the radio-
labeled hexadecane was injected (Run B). This collection continued for 60
min, at which time new impingers were placed in line. The flow was increased
to 1000 mL/min, the temperature of the permeation chamber and dilution
system was reduced to initial conditions, and the effluent collected for
another 60 min period (Run C). The permeation chamber was then removed from
the system and rinsed with a 2 mL aliquot of toluene. This was added to
Triton-X scintillation counting solution and counted along with the samples
from the impinger. Results are listed in Table 3. The permeation system
27
-------
Table 3. RECOVERY OF RADIOLABELED n-HEXADECANE
FROM PERMEATION SYSTEM
Quantity Injected Dpm Observed
Run No. (dpm) Collection Set (collected)
A 16,000 1 1,192
B 16,600 2 7,102
C 0 3 2,509
Toluene Rinses 14,832
Total 25,635
28
-------
was then disassembled and thoroughly washed and dried to remove residual
radiolabeled compounds.
14
Due to the low recovery of the tritiated hexadecane, radiolabeled C-
bromobenzene (b.p. 156°C) was selected for further system condensation
testing. The radioisotope was diluted to a specific activity of 1,335
dpm/(jL so that a 5 |JL injection equivalent to 7.5 mg of bromobenzene-contained
6,675 cpm. The permeation chamber was brought to 60°C the dilution system
was at 65°C, and the hydrocarbon free air flowed at 250 mL/min.
The radioisotope was injected (Run A, Table 4) into the system and the
effluent collected at 250 mL/min in the impingers containing 10 mL of Triton .
X scintillation counting solution. After 60 min, these were removed and
replaced by a second set of impingers which collected for another 60 min.
These aliquots were then counted, and a total of 105% recovery was observed.
Duplicate injections (Run B and C, Table 4) of bromobenzene were then
made with the temperature of the permeation chamber at 30°C and all other
parameters remaining unchanged. As before, the effluent was collected at
250 mL/min and counted. Yields of 92% and 87% were observed.
The low concentration condensation experiment was executed using radio-
14
labeled C-bromobenzene in methanol, 90 M8/ML with a specific activity of
5,838 dpm/|jL. Duplicate injections (Run D and E, Table 4) of 1 |jL-equivalant
to 6 ppt were delivered into the permeation chamber which was at 30°C. The
secondary dilution system was at 65°C, and again the flow rate was 250
mL/min. These two injections netted 88 and 83% recoveries.
Thus, from these experiments, it appears that the permeation/diluter
system is acceptable for yielding high recoveries of semi-polar to non-
polar compounds that have boiling points of up to ^160°C.
The adsorption of volatile polar compounds onto the glass surface of
the permeation/dilution system was the subject of further investigation.
14
This was done with C-diethylamine hydrochloride in water, which had a
specific activity of 60 dpm/(jL and a concentration of 4 ng/(jL.
®
The permeation system was disassembled, washed with Alkonox . followed
® '
by Isoclean , rinsed with deionized water, baked overnight at 400°C, and
then reassembled. Thirty microliters of the diethylamine hydrochloride in
water were injected onto a glass wool plug impregnated with potassium hydroxide
29
-------
Table 4. RECOVERY OF 14C-BROMOBENZENE
FROM PERMEATION SYSTEM
Quantity Injected Dpm Observed Percent
Run No. (dpm) Impinger Set (collected) Recovery
A 6,672 1 6,433 106
2 606
B 6,672 1 6,142 92
2 31
C 6,672 1 5,794 88
2 64
D 5,838 1 5,157 89
2 24
E 5,838 1 4,864 84
2 36
30
-------
in the gas stream of the permeation chamber. The resulting reaction yielded
the free base diethylamine (b.p. 56.3°C). The permeation chamber was heated
to 30°C, the dilution system was at 65°C, and the hydrocarbon free air flow
set at 250 mL/min. The effluent gas stream was bubbled through 4 impingers
in series, each containing 10 ml of the Triton-X scintillation counting
solution. Subsequent analysis of the cocktail indicated no significant
14
amount of the C-diethylamine to be collected from the permeation/diluter
system. The glass wool plug was removed from the system and placed in
cocktail arid counted. Radiation level for the glass wool was 33 counts/min
above background indicating the radiolabeled compound was released into the
gas stream and had apparently adsorbed onto the interior surfaces of the
glass system. Collection of radiolabelled diethylamine in the 4 midget
impinger arrangement was verified by injecting the compound directly in
front of the collection assembly fitted with a heated injection port and gas
flow at 250 mL/min.
Deactivation of the adsorption sites on the glass surfaces of the
permeation/diluter system was attempted utilizing N-methyl-N-trimethylsilyl-
trifluo'roacetamide (MSFTA). A 10% mixture of the low boiling silanizing
agent in methylene chloride was prepared and three 10 (jL injections at 20
min intervals were made into the permeation system through the 70/60 injection
stopper unit. The chamber temperature was 65°C, the dilution system was set
at 150°C and the hydrocarbon free air was set at 250 mL/min and was maintained
in this stage for 18 hrs following the final injection.
Following silanization, the radiolabel diethylamine experiment was
conducted as before. Again, no significant amount of the radioisotope
passed through the permeation/dilutier system.
Following another silanization step, the radiolabel diethylamine experi-
ment was conducted as before. Again, no significant amount of the radioiso-
tope passed through the permeation/diluter system.
®
The system was again disassembled and cleaned with Isoclean , rinsed
with deionized water, and baked at 400°C overnight. The interior of the
glassware was washed with a solution of 0.1% Carbowax 20M in methylene
chloride and baked at 280° for 2 hrs while being purged with nitrogen gas.
After cooling to room temperature, it was again treated with 0.05% Carbowax
31
-------
20M in methylene chloride and again baked for 2 hrs at 280°C with nitrogen
gas purged. The diethylamine experiment was again conducted with no apparent
recovery of the isotope.
Two subsequent treatments with the Carbowax 20M followed by the diethyl-
amine radioisotope experiment indicated adsorption continued to be a problem.
Solutions of 0.1% and 0.05% Carbowax 20M and toluene were made and used
in an attempt to reduce the adsorption onto glassware. The subsequent diethyl-
amine experiment indicated no improvement in the adsorption problem.
Since the first group of model compounds were semi- and non-polar, the
problems associated with the glassware in the permeation/dilution system were
expected to be minimal and primarily influenced by condensation and not
adsorption. Thus the project proceeded with the system described. It was
anticipated that the system would be possibly modified when the low-volatility,
polar compounds were to be studied. These compounds were never studied
however.
TEST ATMOSPHERE GENERATION
When the permeation/dilution system was first used to generate standards,
poor reproducibility was noted in the concentrations generated. After study
of the permeation/dilution system, several ideas for improvement of the
precision of the system were arrived at. These ideas were as follows:
A. Temperature Control
(1) The water-jacketed permation tube chamber and the dilution
bulbs should be at the same temperature.
(2) The temperature of the entire system should be controlled to
+0.1°C.
(3) Both the perm tube chamber and the tube bringing "zero" air
from the constant-temperature water bath should be well
insulated. Also, this tube should be Teflon and not copper.
B. Input and Output Flows of the Dilution Stages
(1) Flows should be measured with a bubble flow meter before and
after use of the dilution system. Flow controllers may be
used to monitor drifts but readings from these devices cannot
be accepted as absolute values.
32
-------
(2) Flow controllers should only be used to control flows in the
range of 10%-90% of the designated controller range.
C. Miscellaneous
(1) Addition of ground glass joints to the tubes on the bulbs
through which dilution gas is added and gas is removed was
®
considered. The initial system used Swagelok fittings to
make the connections. This change, which would.have made
bulb removal for cleaning easier, was not implemented.
(2) The dilution system should not be used to generate low-
concentration samples following generation of high-concentra-
tion samples without first cleaning the bulbs.
Not all these ideas could be implemented. The dilution bulbs need to
be maintained at about 150°C to minimize loss of sample to the glass. Also,
though the permeation tube chamber is maintained at +0.1°C, this could not
be done with the dilution bulbs, where temperature control was about +0.5°C.
33
-------
SECTION 6
EVALUATION•OF SAMPLE COLLECTION DEVICES
INTRODUCTION TO EXPERIMENTAL DESIGN
The experimental design called for collection of various test mixtures
of compounds using six primary collection devices and then recovery and
measurement of these compounds. Various parameters were to be varied while
others were to be held constant. The relationships of these test parameters
were shown in Table 1. In experimental design A (Table 1), the sampling
volume and relative humidity (30%) were to be held constant. The variable
parameters were to be the concentration of each individual compound, and the
storage time. The compound concentrations were to be zero, not less than 10
ppt (low) and not more than 100 ppb (high). In the second experimental
design (B), the sampling volume, ozone, SO- and NO concentrations and the
£ X
relative humidity were to be varied while the sampling time- and rate and
test compound concentration were to be constant. The interferent concentra-
tions were to be as follows: relative humidity -30% and 90%; SO -VLO and
~200 ppb; NO - ^100 and ~500 ppb and 0 - ~75 and ^500 ppb. These concen- .
X ' j
trations reflect low and high ambient levels of SO., N0_ and 0_.
As stated previously the six collection devices were to be polymeric
bags, steel cannisters, glass bulbs, cryogenic traps and Tenax and charcoal
adsorbent traps. The containers and traps were to be tested principally for
their ability to retain the input level of concentration. Three levels of
concentration (zero, low, high) were to be utilized and measurements of
concentration were to be made at three points in time (t , t , t-) . The
layout of the design for the planned study was shown in Table 5. It was
intended that analysis of variance be applied to the data resulting from
this set of measurements to determine the mean differences among the types
of sampling devices. Table 6 presents the design layout for determining
effects of humidity, SO , NO and 0 on high concentration samples.
34
-------
Table 5. EXPERIMENTAL DESIGN FOR OBTAINING MEASUREMENTS ON THREE
LEVELS OF CONCENTRATIONS (ZERO, LOW, HIGH) AT THREE POINTS IN
TIME (t , t , t ) USING THREE CONTAINER TYPES3
. .
Lol-Lcction DGVICG osniple No .
Tedlar Bag 1
2
3
4
5
6
7
8
9
Stainless Steel Can 1
2
3
4
5
6
7
8
9
Glass Bulb 1
2
3
• 4
5
6,
7
Zero High Low
fcO fcl fc2 fcO tl fc2 fcO fcl t2
XX X
X XX
XXX
XX X
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
X X X
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
X
X
X
X
X
X
(continued)
35
-------
Table 5 (cont'd.)
Zero
Low
High
Collection Device
Low Temperature
Condensation Trap
Charcoal Adsorbent
Trap
Tenax Adsorbent
Trap
oamjjj.
1
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
9
e
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
•XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
X • X X
XXX
XXX
XXX
XXX
Concentration levels and storage time were tested. Three repli-
cates are included for each concentration.
36
-------
Table 6. EXPERIMENTAL DESIGN FOR DETERMINING EFFECTS OF HUMIDITY,
S00, NO , AND 00 ON HIGH CONCENTRATION SAMPLES3
2 x 3
Collection Device
Tedlar Bag
Stainless Steel
Can
Glass Bulb
Low Temperature
Condensation Trap
•
Charcoal Adsorbent
Trap
Tenax GC Cartridge
Sample No.
10
11
12
13
10
11
12
13
10
11
12
13
10
11
12
13
10
11
12
13
10
11
12
13
Experiment IB
X
X
X
X
X
X
X
• X
X
X
X
X
Experiment 2B
X
X
X
X
X
X
X
X
X
X
X
X
Sample volume, ozone, SO , and NO concentrations, relative humidity
are the test parameters.
Experimental Design IB 0 V500 ppb
SO^ ^200 ppb
NO ^500 ppb
RHX
Experimental Design 2B 0- . ^75 ppb
SOj ^10 ppb
NO^ ~100 ppb
RH -v-30%
Samples will be collected in triplicate.
37
-------
The experimental design for testing the cryogenic trap and the charcoal
and Tenax sorbents is essentially identical to that shown for the
containers. The principal difference between the containers and these
devices is that replicate samples can be drawn from jeach container whereas
only one sample can be recovered from the trap and sorbents. Thus, 6 sam-
ples/experiment were collected for each concentrating sampling device.
EXPERIMENTAL PROCEDURE
Sample Collection
Bags-
Three types of polymeric bags were selected for evaluation. These were
FEP Teflon, Tedlar (polyvinyl fluoride) and five-layered aluminized bags.
Polymeric bags present the greatest challenge as contamination can arise
from the walls of the bag, from the polymeric material itself and from
diffusion into the bag from the outside environment.
Cleaning—The cleaning procedure originally proposed and utilized
involved several cycles of evacuating and filling the bags with clean air.
@ ®
This procedure worked generally well for Teflon and Tedlar bags. The
five-layer bags could not be cleaned to any reasonable degree and thus were
eliminated from .the study. The cleaning procedure described above sometimes
® ®
failed even with Teflon and Tedlar and thus some different cleaning proce-
dures were tested. The four bag-cleaning methods investigated were: (1)
evacuation of contaminated contents; (2) clean air flushing; (3) clean air
flushing plus direct sunlight irradiation; and (4) clean air flushing plus
ozonation and direct sunlight irradiation.
Four Tedlar and four Teflon 10 liter bags (2 and 5 mil thickness)
were used for evaluation of the clean-up methods. These bags were prepared
with sheet polymer provided by the EPA; the bag seals were made using a
Vertrod (Brooklyn, NY) thermal impulse heat sealer. Each prepared bag was
filled with clean air containing approximately one part per million of an
aromatic-aliphatic hydrocarbon mixture. After several hours, each bag was
evacuated and refilled with clean air. The following day, the contents of
® ®
one Tedlar and one Teflon bag which had not been flushed previously were
® ®
chromatographed. A second set of one Tedlar and one Teflon bag were
flushed two times with clean air and then their contents were chromatographed.
38
-------
Another set was flushed two times with clean air, filled with clean air, and
© ®
irradiated in direct sunlight. A fourth set of Tedlar and Teflon bags was
flushed twice with clean air, filled with approximately 25 parts per million
ozone, and irradiated in direct sunlight. After several hours, the irradiated
bags were evacuated and refilled with clean air. The following day, these
remaining four bags were flushed two times with clean air and their contents
chromatographed.
Data acquired showed that the clean air flush was sufficient in removing
the majority of the volatile compounds from the Tedlar bags. The total
chromatographic peak area measured with samples taken from the cleaned bags
was found to be about two to four times that measured with samples of clean
laboratory air. This peak area level corresponds to about 1-2 ppb C. A
reduction to near non-detectable levels is observed with clean air plus
irradiation and little additional clean-up is gained with ozonation plus
irradiation.
®
The low molecular weight compounds were easily removed from Teflon
with clean air flushing but the heavier aromatic hydrocarbons were not. It
appears that a minimum of clean air flush plus irradiation is necessary to
clean Teflon bags satisfactorily. Ozonation plus irradiation is required
to remove all compounds to near non-detectable level.
Sampling Procedure—For polymeric bags, a typical field application
involves pumping sample into the bag through a metal bellows pump or placing
the bag in an airtight vessel with the inlet of the bag exposed to a sample
port, then evacuating the enclosure vessel to fill the bag. Passing the
sample through a stainless steel bellows pump could introduce a new variable
into the evaluation if any of the species in the mix reacted with stainless
steel. In order to avoid this complication, the mixture should enter the
bag directly from the sample port with no metal contact. In this program
the bags were attached to the glass manifold of the permeation/dilution
§
system with TFE Teflon fittings and allowed to fill by the pressure of the
system.
Glass Bulbs--
The glass bulbs used in this project were prepared from 2 liter, round-
bottom Pyrex glass flasks to which had been attached Teflon, high vacuum
39
-------
stopcocks. Six nun glass tubing was attached to the second port of each
®
stopcock so that connection could be made with a Swagelok fitting to metal
and/or polymeric tubing.
Cleaning—The procedure originally proposed and utilized for cleaning
glass bulbs was complex. The vacuum stopcocks were removed and the glass
bulbs were rinsed with a mixture of potassium dichromate and sulfuric acid,
rinsed several times with distilled water and then placed in an oven at
400°C for 24 hours. The bulbs were then evacuated, flushed with clean air
and heated three times each. A major problem encountered with the glass
bulbs was their fragility. Bulb input/out tubes were broken numerous times
while removing the stopcocks. It was thus decided that the cleaning procedure
for bulbs should consist of several cycles of filling the bulbs with clean
air and then evacuating them while the bulbs were heated to a temperature of
about 150°C. Viton o-rings must be used for the stopcocks. Other materials
will decompose and give rise to contaminants at this temperature.
Sampling Procedure—Typically glass bulbs are sent to the field evacuated
and then simply opening a stopcock to obtain a sample. This method however,
would not provide enough sample for the repetitive analyses to be performed.
Thus the glass bulbs were filled by means of a metal bellows pump. The
upstream side of the pump was connected to the output manifold of the permea-
tion/dilution system. The downstream side of the pump was then connected to
the bulb by means of metal tubing and a Teflon, Swagelok fitting. A small
pressure gauge on a T connected to this tube indicated bulb pressure. The
bulbs were filled to about 15 psi (above ambient pressure). They were not
filled above this pressure as a matter of safety.
Cannisters--
Two types of steel cannisters were evaluated. One type was prepared in
the RTI laboratory. The bodies of these containers as well as the tops were
made from 304 stainless steel. The container bodies were constructed from a
2-L stainless steel beaker manufactured by Vollrath (Sheboygan, Wisconsin).
The beakers were electropolished by filling them with a 1:1 (w/w) mixture of
concentrated sulfuric and phosphoric acids and applying a 6-volt charge at 8
amps, using the beaker as a cathode and a tin bar immersed in the acid
mixture as the anode. The tops of the containers were 1/8 inch (3.175 mm)
40
-------
thick stainless steel. The lips on the stainless steel beakers were cut
with a wet cutting wheel after the containers had been electropolished.
After the container was cut, it was cleaned with strong oxidizing reagents
(Nochromix,-Godax Labs) to remove any grease deposits on the inside of the
container.
A I/A inch (6.35 mm) x 2 inch (50.8 mm) stainless steel tube was heliar-
ced to the top of each container. The tops were then electropolished, and
the tops and bodies of the containers were joined by heliarcing under an
inert atmosphere to prevent any oxidation of the interior surfaces during
this process. The containers were then mounted with H series Nupro valves
constructed of stainless steel with metal bellow seals. Each container was
then engraved with a letter and number. The container was then ready for
testing. It was first pressurized to 60 psi with zero air and leak checked.
The container was immersed in clear water and visually inspected for leaks.
If there were no leaks, the container was ready to be cleaned.
®
The second type of steel container evaluated was the "Summa " polished
stainless steel container manufactured and sold by D&S Instruments Ltd.,
®
Pullman, WA. Summa is a proprietary electropolishing method of Molectrics
Corp. These 6 L containers are spherical with a cylindrical base welded to
each to serve as a stand. Each container is mounted with a single input/out-
put tube to which is attached two H series Nupro valves in a T configuration.
Cleaning—The cleaning of the steel cannisters consisted of several
cycles of filling with clean air and then evacuating while under conditions
of high temperature. Four containers at a time were connected to a manifold
®
in a Marinite box fitted with a large heating element and a temperature
controller from RFL Industries, Inc. (Boonton, NJ). The manifold was valved
so that either vacuum could be applied or clean air could be introduced.
The temperature in the box was raised to 150°C. The containers were then
evacuated to less than 0.5 mm Hg pressure and maintained at that condition
for 1-2 hours. The containers were then filled with clean air to a pressure
of about 60 psi and maintained at that condition for 1-2 hours. The contain-
ers were then evacuated once again and the whole cycle repeated. Generally
4-5 cycles were sufficient for cleaning. At the conclusion of the last
cycle, the containers were evacuated in preparation for sample collection.
41
-------
Sampling Procedure—The steel cannisters were loaded with test gases
from the permeation/dilution system using a metal bellows pump (Metal Bellows
Corp., Model MB-151). All connections from manifold to pump and from pump
®
to container were made with stainless steel or Teflon connectors and tubing.
The containers were filled at a rate determined by a critical orifice placed
between the pump and the container, which was about 300 mL/min. The contain-
ers were filled to a pressure of about 15 psig.
Tenax GC Cartridges—
Preparation—Virgin Tenax GC (Applied Science, State Park, PA) was
extracted in a Soxhlet apparatus for a minimum of 18 hr with methanol prior
to its use. The Tenax GC sorbent was dried in a vacuum oven at 100°C for
3-5 hr and then sieved to provide a fraction corresponding to 35/60 mesh.
This fraction was used for preparing sampling cartridges.
The sampling tubes were prepared by packing a 10 cm long x 1.5 cm i.d.
glass tube containing 6.0 cm of 35/60 mesh Tenax GC with glass wool in the
ends to provide support. Cartridge samplers were then conditioned at 270°C
with helium flow at 30 mL/min for 30 rain. The conditioned cartridges were
transferred to Kimax (2.5 cm x 150 cm) culture tubes, immediately sealed
using Teflon-lined caps and cooled. The culture tubes were placed in scalable
cans to provide a second seal during storage. This procedure was performed
in order to avoid recontamination of the sorbent bed.
Sampling Procedure—The sampling cartridges were assembled as shown in
Figure 5. The Teflon Swagelok union was attached to the small diameter
(0.25 in o.d.) end of the cartridge in a 416 Beckman union fitted with a
®
Luerlok was fitted to the other end. The hypodermic needle was attached to
the Luerlok and the 0.25 in Teflon Swagelok was tightened to the manifold
of the permeation/dilution system. Sampling at 1 L/min was then initiated.
After sampling of 30 L of air from the permeation/dilution system (blank) or
§
the air-vapor mixture, the Tenax cartridges were returned to the Kimax
®
culture tubes. All sampling cartridges were handled with Kimwipes or clean
tweezers to avoid their contamination.
Charcoal Cartridges —
Preparation--NIOSH charcoal tubes (200 mg) were purchased from Supelco
(Belfonte, PA) and used in these experiments as received.
42
-------
Luer-Lok/Septum
Tenax Cartridge•
Manifold
Pump^f—
•- V
•^—i //in i
7
1/4" i.d.
Latex
200 mg
N.I.O.S.H Manifol
]"
7
1/4" x 1/4" 1/4" x 1/41(
Beckman Swagelok
Figure 6. Charcoal cartridge sampling arrangement employed with
permeation/dilution system.
-------
Sampling Procedure—Figure 6 depicts the devices associated with
connecting the charcoal sampling tube to the permeation/diluter system. A
total of 30 L of the synthetic air-vapor was sampled at 1 L/min.
Cryogenic Traps--
Preparation—Cryogenic traps were made of 0.25 in o.d. 0.21 in i.d.
nickel tubing in 24 in lengths. Traps were coiled in a transaxial configura-
tion (Fig. 7). Cryogenic traps were cleaned with methanol and pentane to
remove cutting oil and then thermally conditioned at 160°C under helium flow
(V30 mL/min) for 2 hr. Initially the cryogenic traps were used empty.
Subsequently, when recoveries were observed to be low, they were filled with
clean 2 mm glass beads and used with liquid oxygen as the cryogen. Upon
cooling (under flow), the nickel traps were immediately sealed with Swagelok
fittings. The helium used during conditioning of the traps as well as for
purging the contents onto Tenax cartridges was passed through a liquid
nitrogen cryogenic trap to remove impurities, a step found to be essential
to achieve a low background.
Sampling Procedure—The configuration for collecting vapor-phase organics
using nickel cryogenic traps is shown in Figure 8. A transaxially coiled
trap was employed and the cryogenic trap was immersed midway into finely
crushed dry ice or, in later experiments, liquid oxygen. The opened end of
the quick-connect, the 0.25 in x 0.5 o.d. glass adaptor, and the 416 Beckman
were assembled as shown in Figure 7. One end of the trap was attached to
the pump while the other end was attached to the manifold of the permeation/
®
dilution system with a Teflon union. The nickel traps were then set in the
cryogenic medium which cooled only half of these transaxial traps. This
configuration was used while collecting 30 L of the test air-vapor mixture
at. 1 L/min. After sampling, the quick-connect fitting was removed and the
end was capped immediately. The nickel trap was removed from the cryogenic
bath and placed in a container of dry ice for storage. No solvents were
employed throughout the collection and analysis of the cryogenic traps.
Measurement Procedure
Bags, Bulbs, and Cannisters--
Recovery--The recovery and measurement procedure was identical for
samples in bags, glass bulbs, and steel cannisters. The container is first
44
-------
-p-
Cn
PURCE
CAS
SWAGELOK
QUICK CONNECT
S.S. 1/4" Tube
SWAGELOK
CAP
T.F.E. or S.S. 1/4" Tube
BECKMAN
REDUCER
T.F.E. 1/4" x 1/2" Tube
Tenax Cartridge
Figure 7. Purging (A) and storage arrangements (B) for recovering vapors from Ni traps.
-------
MANIFOLD
ON
SWAGELOK
UNION
T.F.E. 1/4" Tube
SWAGELOK
QUICK CONNECT
S.S. I/A" Tube
II fi
DEWAR FLASK
PUMP
Figure 8. Nickel cryogenic trap sampling arrangements employed with permeation/dilution system.
-------
placed in a box and loosely connected via a Swagelok fitting to a heated
sample transfer line as shown in Figure 9. The hox could be heated and was
done so for the glass and steel containers but not the bags; the temperature
used for the glass and steel containers ranged from 50° to 90°C. Valve A is
opened and the inlet line is purged with clean air. At this point, the
entire system can also be purged. After the lines have been purged, valve A
is closed and the connection to the container immediately tightened. The
multiposition valve is turned to the sample loop, and the corresponding
outlet valve (C) is opened. The sample loop used was a 15 cm length of 1 mm
i.d. stainless steel tubing in the shape of a U filled with glass beads.
Valves D and E are opened to evacuate the entire system and the sample loop
is immersed in liquid oxygen. Then valve D is closed, the container valve is
opened and the heated metering valve (B) is opened to allow sample to enter
the system.
The amount of sample passed through the sample loop is determined by
the pressure change on the Heise gauge. For cryogenically trapped samples,
the following equation is used.
V AP
V = gas volume passed through trap (mL)
V = total dead volume in the system (this was 536 mL in our system)
AP = pressure change registered on gauge (mm Hg)
P = ambient barometric pressure (mm Hg)
3
A typical sample volume was 200 mL. After the sample was trapped in
the cryogenic loop, the liquid oxygen Dewar was removed, and a heated sili-
cone oil bath (150°C) was substituted for this Dewar. This sudden heating
caused flash volatilization of the trapped organic compounds. Rotation of
the valve resulted in carrier gas sweeping the sample into the gas chromato-
graphic column.
The measurements were performed with a Perkin-Elmer Model 3920 gas
chromatograph modified to accept a 30 meter SE-30 SCOT column. The GC
system included a column effluent splitter to allow simultaneous FID and EC
detection of the mixture components. The FID and EC responses were recorded
47
-------
PUMP
OIL
TRAP
Figure 9. Vacuum injection manifold.
Analysis:
1.
2.
Sample placed in heated enclosure.
100 mL portion passed through
cryogenic traps.
3. Flash desorption onto column.
-------
on a dual pen strip chart recorder. The outputs of the two detectors were
also connected to analog-to-digital converters which are part of a Hewlett-
Packard 3352B laboratory data system. Peak retention times and areas were
determined by the computer and printed out in tabular form. The gas chroma-
tographic parameters used are listed in Table 7.
Calibration--Calibrations were performed for each group of compounds.
The FID response was standardized at three different compound concentrations
by drawing a sample directly from the glass manifold through a heated 1/8"
O.D. stainless steel sample line into the GC sampling system. Each concen-
tration was calculated based upon the permeation rate of the compound, the
chamber air flow, and the dilution air flow. Typical concentrations used in
establishing standard curves and the correlation coefficient for each compound
response curve are given in Section 7 in Tables 47 and 48.
As the results indicate, furan and acrylonitrile were not resolved on
the GC system used in this study. A percent recovery for furan plus
acrylonitrile is given for the "Summa" polished cans as well as the other
containers studied.
© "
Tenax GC Cartridges and Cryogenic Traps--
Recovery--The instrumental conditions for the thermal desorption/gas
®
chromatographic analysis of volatile organics on Tenax GC sampling cartridges
are given in Table 8. The inlet-manifold system is depicted in Figure 10.
®
The thermal desorption chamber and the six-port Valco valve were maintained
at 270°C. The helium purge gas through the desorption chamber was adjusted
to 15 mL/min. The Ni capillary trap on the inlet manifold was cooled with
liquid nitrogen. In a typical thermal desorption cycle, a sampling cartridge
was placed in the preheated desorption chamber and helium gas was passed .
through the cartridge to purge the vapors into the liquid nitrogen capillary
trap. After the desorption was completed, the six-port valve was rotated
and the temperature on the capillary loop was rapidly raised. The carrier
gas then introduced the vapors onto the high resolution GC column. The
glass capillary column was temperature programmed under the conditions
listed in Table 8. After all of the components had eluted from the column,
the column was cooled to ambient temperature and the next sample was processed.
49
-------
Table 7. CHROMATOGRAPHIC PARAMETERS FOR ANALYSIS OF CONTAINERS
Parameters
Setting Conditions
Column
Carrier Gas
Make-up Gas
Column Temperature
FID - Air flow
- H flow
GC
Detector Temperature
30 m SE-30 SCOT;
0.5 mm i.d.
He - 5 mL/min
He - 23 mL/min for FID
Ar/CH, - 15 mL/min for
ECD
0°C/4 min, 4°C/min, 150°-
0 min
40 psi
17 psi
Perkin Elmer Model 3920
200°C
50
-------
Table 8. OPERATING PARAMETERS FOR THERMAL DESORPTION AND GC/FID
OF TENAX CARTRIDGES
Parameter
Setting
Inlet Manifold
Desorption chamber and valve
Capillary trap - min
- max
Desorption time
He purge flow
65 m glass SCOT SE-30;
0.50 mm i.d.
Carrier (He) flow
FID - Air flow
- H- flow
Detector temperature
GLC
270°C
-195°C
240°C
8 min
15 mL/min
40°C for 6 min, 40-210°C, 4°C/min
^-2.5 mL/min
•^275 mL/min
~30 mL/min
250°C
51
-------
en
ro
PURGE
GAS
ALUMINUM
HEATING BLOCK
PLATINUM
PROBE SENSOR
HEATING
CARTRIDGE
CARRIER
GAS
TOGLC
CAPILLARY
TEFLON SEAL
SPRING
METAL SEAL
HEATING
CARTRIDGE
/ VENT TO
CARBON TRAP
TWO POSITION
SIX PORT
VALVE
HEATING
CARTRIDGE
HEATING AND
COOLING BATH
LIQUID NITROGEN
VALVE POSITION A
(SAMPLE DESORPTION)
VALVE POSITION B
(SAMPLE INJECTION)
Purge
Gas
Figure 10. Inlet-manifold.
-------
A Varian Model 3700 GC equipped with a thermal desorption inlet manifold
(Fig. 10) and a Varian CDS 111 integrator was used to obtain chromatographic
peak areas for each of the components in the mixture.
A GLC oven was used to heat the nickel cryogenic trap and transfer the
®
vapors to Tenax GC cartridges. This was accomplished as follows: the
helium gas was set to pass at 200 mL/min through a transaxial trap submerged
in liquid N? for cleaning the helium gas. Then this clean trap under He
flow was placed in Dewar flask No. 1 and liquid nitrogen was slowly added
until full and the system was equilibrated. The helium flow was then reduced
to 20 mL/min. A second Dewar was filled halfway with finely crushed dry
ice. The sample trap (which had been stored on dry ice) was immediately
inserted and the Dewar No. 2 filled full of dry ice. After equilibration,
®
the end cap (downstream) was removed and the Tenax GC "transfer" cartridge
was attached. The end cap (upstream) was removed and the stainless steel
flex tube flowing with clean He (20 mL/min) was connected using a 0.25 in x
0.25 in stainless steel union. The trap was lifted from the second Dewar
and placed across the open face of the GLC oven with the Tenax cartridge on
the outside edge of the oven. The oven and therefore the trap was programmed
from 30 to 160°C at 15°/min. The temperature was held at 160° for 2 min and
®
then the Tenax cartridge was' returned to the Kimax culture tube. After the
oven was' cooled to room temperature, the downstream end cap was placed on
®
the trap and then upstream was also capped. The analysis of the Tenax GC
transfer cartridges containing the content of the cryogenic traps was
identical to the procedures described for Tenax GC sampling cartridges.
®
Calibration--For the analysis of Tenax GC cartridges including vapors
transferred from cryogenic traps, the thermal desorption/gas chromatography
(TD/HRGC) data system was calibrated using two independent techniques. The
first method utilized permeation tubes in a permeation system used specifi-
cally for calibrating instruments (8-10). This permeation system was not
the one used in these studies. Quantities of each test compound were loaded
onto Tenax GC cartridges as a group and then analyzed according to the
conditions described in Table 8. Levels of the test compounds were loaded
at several different levels and a linear regression analysis was made (cali-
bration curve).
53
-------
The second technique was used for verification of the first and employed
the preparation of test compounds in a methanol solution and then injecting
2.0 (JL into a flash (250°C) evaporation unit (Fig. 11). The components were
©
swept by clean He (30 mL/min, 500 mL total volume) onto Tenax GC cartridges.
Subsequently, the cartridges were analyzed and the FID responses were measured
by a Varian chromatography data system. Standard curves using linear regres-
sion analysis were prepared. The use of the flash evaporation unit was
applicable to compounds with breakthrough volumes significantly larger than
methanol (ca. 500 mL) and thus verification of standard curves for vinyl
chloride, methyl chloride, acrylonitrile, furan and chloroprene were not
possible by this procedure.
Charcoal Cartridges--
Recovery--The content of the charcoal sampling trap was emptied into a 1
mL volumetric flask and a carbon disulfide/methanol solution (30/70 v/v) was
added to the mark. After 1 hr of desorption, aliquots were taken for gas
chromatography analysis with flame ionization and electron capture detection
(GC/FID and GC/ECD, respectively). Table 9 presents the operating parameters
for the GC/FID and ECD. These conditions were found to be the most suitable
for the test model compounds.
Calibration--Solutions of compounds for analysis by GC/FID and GC/ECD
were prepared in carbon disulfide and carbon disulfide/methanol, respectively.
Standard curves were prepared from four points (including zero) calibration
data (triplicate analysis per point) using linear regression analysis.
Microliter quantities were injected into the GC and peak heights or areas
were determined for calculating quantities and recoveries.
Air/Vapor Generation Procedure
Storage-Stability Study--
•The permeation/dilution system described in Section 5 was used for
generating synthetic air/vapor.mixtures. The list of 27 test compounds
(Table 2) was divided into two major groups, so that each chemical could be
resolved during instrumental analysis. The Groups I and II contained 15 and
13 compounds, respectively, with 1,2,3-trimethylbenzene included in each
group.
54
-------
Ul
TO FLOW
METER
TENAX GC CARTRIDGE
Z SEPTUM 3-WAY STOPCOCK
/ 1 •*— — H« FLOW (3
< WM ^tt
TEFLON UNIONS
\ "
m . -. M
LOADING TUBE WRAPPED
WITH ALUMINUM FOIL- 1 1
AND HEATING TAPE Q
CARBON TRAPS
/ [
Figure 11. Schematic of vaporization unit for loading organics dissolved in metha,nol
onto Tenax GC cartridges.
-------
Table 9. OPERATING PARAMETERS FOR GC/FID AND GC/ECD ANALYSES OF
SOLVENT DESORBED CHARCOAL TUBES
Parameter
Setting
GC/FID
Injection port temperature
Detector temperature
Carrier flow (N )
Hydrogen flow
Air flow
Column-programmed
GC/ECD
Injector port temperature
3
Detector temperature (Sc- H)
Carrier flow (N )
Pulse width
Pulse interval (adjustable)
Column - Isothermal Step 1
Step 2
270°C
270°C
20 mL/min
~30 mL/min
•v270 mL/min
8°C/min
270°C
300°C
20 mL/min
1.0 (jSec
25-1000 (JSec
150°C
100°C
56
-------
Each chemical was placed in a permeation tube (see Section 7 for prepara-
tion and calibration) and constant permeation established at 30°C. The group
of permeation tubes used in each experiment was placed in the permeation/dilu-
tion system and the system equilibrated overnight before initiating any
sampling of synthetic air/vapor mixtures. The flow through the glass manifold
on the permeation/dilution system was at 5 L/min, thus allowing for more than
one collection device to be attached during the experiments. The temperature
on the glass dilution bulbs and manifold assembly was maintained at 200°C +
1°C. Dilutions of the synthetic air/vapor mixture was made to achieve the
ppt range when low level studies were conducted.
The supporting performance data and quality control and assurance practi-
ces invoked on the permeation/dilution system are given in Section 7.
Containers were filled by attaching the empty polymeric bags directly to
the manifold port or by using a metal bellows pump to fill and pressurize the
bulbs. Stainless steel cannisters and glass bulbs were filled at a rate of
1 L/min and 300 mL/min, respectively. Each was pressurized to 15 psi.
Traps were sampled at 1 L/min for 30 min with a Nutech Model 220-A
sampler attached downstream to the collection device, with the device attached
to the manifold..
Table 1 gave the experimental conditions employed for the storage-
stability study.
Interference Study--
In this study ozone, SO and NO levels and relative humidity were
^ X
varied while the sampling time, storage time, and concentration were held
constant. The levels employed are given in Table 10. These concentrations
reflect high and low ambient levels of 0 , NO and SO .
.3 X L.
Preliminary Preparations—The portable permeation system was employed
for the execution of the interference study. The entire system was operated
at ambient pressure. All connections and other components were glass or
Teflon®.
Clean air was humidified by passing it though a fritted bubbler contain-
ing water and then irradiated with ultraviolet light to generate ozone. The
ozone/water vapor in air mixture was mixed with nitric oxide and sulfur
dioxide from certified gas cylinders to generate the required concentrations
57
-------
Table 10. CONCENTRATIONS OF INORGANIC GASES EMPLOYED IN
INTERFERENCE STUDIES
Containers
TJ *. *. ' „ T T-. 4- * — .C — — — — *.
High
Ozone 360a
N02 360
S02 200
H20 90% RHb
Low
75
100
10
26% RH
Traps
High
380
380
190
90% RH
Low
60
60
12
30% RH
Values in ppb. Levels were determined with Bendix 0. and NO monitors
by measuring at the point of sampling from the permeation/dilution
system.
RH = relative humidity, measured with a YSI Dew Point Sensor.
58
-------
of each component. The entire interference gas mixture was mixed with the
model compounds in the clean air stream within a mixing bulb contained in
the permeation/dilution system oven which was heated to approximately 200°C.
Zero-air was sampled to determine "zero" settings on the Bendix NO and
ozone monitors. A certified tank of NO (NBS standard) was used to calibrate
the NO monitor. A known concentration of NO was delivered and the "span"
x x
settings of NO, NO , and NO were adjusted. In all cases, NO + NO = NO .
^ X £* A
The concentrations delivered (600 ppb) was greater than the concentration to
be used experimentally and the meter readings below 600 ppb were assumed to
be linear. The ozone generator was then calibrated according to slide
settings, the position of which determined exposure of the air to the UV
light. In all cases, if NO was greater than ozone, then N0_ equalled the
inital ozone concentration. An audit (Section 7) was performed on the ozone
monitor and the results agreed with the originally determined values.
A certified tank of SO. (NBS standard) was used to deliver a known
concentration of SO. into the system.
. • During the experiments the relative humidity was monitored with a YSI
Dew Point Sensor. 'Ozone and nitrogen dioxide concentrations were monitored
with a Bendix ozone monitor. Ozone concentrations were measured directly
whereas nitrogen dioxide concentrations were measured by the difference in
ozone concentration after gas phase titration. Sulfur dioxide concentrations
were not monitored.
Air/vapor mixtures were generated using the Group I and II model compounds,
The organic vapor concentrations ranged from 1-100 ppb.
Containers—The sample containers for the high level interference study
were filled after allowing the dilution system to equilibrate overnight. The
"Summa" polished stainless steel cans were filled to a pressure of 15 psig by
drawing sample at 1 L/min from the glass manifold through a metal bellows
pump fitted with a flow restrictor on the pump outlet. The glass bulbs were
filled to a pressure of 15 psig at a rate of 300 mL/min in the same manner.
The Tedlar bags were filled by attaching them directly to the glass manifold.
Three containers of each type were filled with sample.
One container of each type was filled from the permeation/dilution
system after having removed the permeation tubes the night before and allowing
59
-------
only interferent gases to flow. These containers served as controls. The
same procedure was used in filling the containers for the low level interfer-
ence study.
Traps—Tenax GC cartridge and cryogenic trap sampling was conducted as
described for the storage-stability study with a few modifications to the
cryogenic procedure. The Ni traps were filled with clean glass beads and
liquid oxygen was the cryogen in this study.
RESULTS
Special Bag Studies
Polymeric bags have been extensively used in the past as viable contai-
®
ners for captive air irradiations. Both Tedlar (polyvinyl fluoride) and
®
Teflon (fluorinated ethene propene) transmit almost the full solar spectrum
®
and are easily fabricated. While Tedlar tends to be slightly sturdier than
Teflon , it has also been found to offgas hydrocarbons due to the solvent
used in making the Tedlar material (1). Teflon bags do not suffer from
this problem since solvents are not used in its manufacture. However, for
both types of polymeric bags substances have been found to permeate through
the film (17,18). It has also been observed that substantial bag-to-bag
®
differences exist with Teflon acquired from different manufacturing batches
(18).
The initial studies using Group I compounds and bags yielded chromato-
grams such as those shown in Figures 12-15. While initial chromatograms of
® ®
Tedlar and Teflon bag content indicated principally Group I compounds, the
chromatograms became increasingly complex with time. This increasing conta-
® ®
mination with time of the Tedlar and Teflon bags was thought to be due to
a number of possible factors: (1) leaking bags; (2) offgassing from the
polymeric film, or (3) diffusion of contaminants from ambient air into the
bags.
Even though the objective of this program was to evaluate containers,
under a directive, effort was devoted to determining whether the contamination
with time of the polymeric bags was due to offgassing of the film, permeation
of substances through the walls, leakage or a combination of factors.
All bags used in this study were first tested for leaks by filling them
three-quarters of the way full with clean air and placing a light book on
60
-------
Figure 12. 2 mil Teflon Bag 61; low concentration study; Group I compounds; T~, June 13, 1980.
-------
Figure 13.
2 mil Teflon Bag 61; low concentration study; Group I compounds; T7, June 20, 1980.
-------
CO
Figure 14. Tecllar Bag Fj'low concentration study; Group I compounds; Tn, June 5, 1980.
-------
ON
Figure 15. Tedlar Bag F; low concentration study; Group I compounds; T?, June 12, 1980.
-------
top. It should be noted that plastic bags are extremely fragile, especially
®
Teflon , so that care 'should be taken in the above procedure or the test
itself may cause the bag to start leaking.
The bags were stored either in the lab inside an aluminum suitcase (to
protect it from light) as in the low concentration study or inside a 3 ft x
2 ft x 1 ft steel box which was continuously flushed with clean air. The
top of the box was equipped with a rubber gasket and the lid was held on
with clamps so that the box remained air tight. A vacuum was held inside
the box for several hours before use so it was considered acceptable for our
purpose. It was decided to continually flush the box with clean air so that
any leakage into the box or offgassing from the outside of the bags would be
continually diluted instead of building up as in a static environment.
Several types of polymeric bags were used for the study. Five-layered,
polyethylene/aluminized bags (Calibrated Instruments, Inc., Ardsley, NY)
®
were tried but could not be cleaned sufficiently for use. Four Tedlar
bags, two which had been cleaned by irradiating at a high concentration of
ozone and then flushed with clean air, and two new bags which had only been
I1
flushed once with clean air, were utilized for the actual studies. The type
of cleaning treatment used may contribute to the amount of offgassing observed
from the walls. Also, two types of Teflon bags were used, three 2 mil and
two 5 mil bags. If permeation plays a role in the contamination, then
differences may be seen with film of different thicknesses.
Initial measurements of the bags containing clean air were taken as
soon as the bags were filled. It took approximately one day for all bags to
be filled and analyzed. An analysis was then taken seven days later, corres-
ponding to the time span of the low concentration study, and then ten days
later. Selected segments of example chromatograms taken on these days are
shown in Figures 16-19. The segments shown are of the same time span so
they can be compared quantitatively. This section of the chromatogram
corresponds to the elution of lower molecular weight hydrocarbons (~3
minutes to 10 minutes). The day initial measurements were taken, a contamina-
tion existed in our system which is the offscale peak shown in each chromato-
gram. This contamination did not exist on subsequent analysis days.
65
-------
Non-bag Contaminant
ON
Figure 16. 2 mil Teflon Bag //6l, zero air; bag stored in clean air; T_, July 30, 1980.
-------
Figure 17. 2 mil Teflon Bag //6], zero air; bag stored in zero air; T n, August 9, 1980
-------
00
Figure 18. 2 mil Teflon Bag #62, zero air; stored in air lab; TQ, July 30, 1980.
-------
Figure 19. 2 mil Teflon Bag //62; zero air; stored in lab air; TIQ, August 9, 1980
-------
As exemplified by these chromatograms, bags stored in ambient air
contained considerably more contamination than those stored in a clean
environment. Those stored in the clean air, however, did not remain entirely
contamination free though the degree of contamination would not prevent
their use in most sampling efforts. This could be due to several factors.
The clean air bags were exposed to ambient air for short periods of time
when analyses were being performed. It could be that this slight contamina-
tion is due to permeation of ambient air during these times. It also could
be that contamination is due both to permeation and offgassing. The bags
exposed to ambient air were stored in an aluminum suitcase. Only one type
of bag was stored in each suitcase. If the bags offgassed, the immediate
environment would be higher in hydrocarbons than the surrounding lab air.
This may cause enhanced diffusion of hydrocarbons in to the bags and, there-
fore, explain some of the high levels of hydrocarbons measured in the bags.
®
Support for this possibility comes from the Teflon bags. Near the end of
the chromatograms, where the higher molecular weight hydrocarbons elute, a
© ®
number of peaks not observed with the Tedlar bags occur in Teflon bags
exposed to ambient air and clean air. It is possible these peaks may be due
to offgassing. While the quantity of these compounds is greater in the 2
mil exposed to ambient air as compared to the 2 mil in a clean environment,
®
the quantities are close to equal in the two 5 mil Teflon bags. This
implies permeation also plays a role in this contamination. From our experi-
®
ments offgassing appears to be indicated to some degree in both Tedlar and
® ®
Teflon bags. Tedlar seems to offgass in the lower molecular weight hydro-
®
carbon region, while Teflon appears to offgass higher molecular weight
®
compounds. The different cleaning treatments of the Tedlar bags appeared
to make little or no difference in the amount of contamination observed with
time.
®
The question arose as to' whether or not different batches of Teflon
®
and Tedlar would show different amounts of contamination with time. Thus a
® ®
batch experiment was devised and carried out. Four Tedlar and four Teflon
bags were used in the study with two bags of each film type coming from one
batch and two bags of each film type coming from another batch.
70
-------
After flushing each bag once with clean air, it was refilled with clean
air and the air was analyzed. This constituted the day zero measurement.
For the purpose of gaining additional insight into the modes of bag
contamination, one bag from each batch was stored in an aluminum case exposed
to laboratory air while the other was stored in a steel box which was sealed
and continually flushed with clean air. After storage for seven days in
their respective environments, the bag contents were again analyzed.
An inter-batch comparison of areas at day zero is not justified since
the bags would have been contaminated to different degrees during manufacture.
A better comparison is the ratio of day 7 area to day 0 area for each bag.
This accounts for any initial contamination while showing the offgassing and
permeability properties of the bag. From Table 11, it can be seen that the
®
Tedlar bags in batch #2 had a greater increase in contamination than the
Tedlar bags in batch #1 in both environments. The Teflon bags in batch #1
had a greater increase in contamination in both environments than those in
batch #2. . '
The greater increase in total peak area for bags stored in the cases
and exposed to laboratory air is not proof of permeation of gases through
the bag wall, although data has previously been presented to suggest that
permeation does play a significant role in contamination. Offgassing of
material from the bag wall could give similar results. In the steel box
flushed with clean air, gases coming off the outside bag wall would be re-
moved. Gases coming off the inside bag wall could create a concentration
gradient which would enhance permeation of gases out of the bag where they
would be removed. Some offgassed material could still be detected if the
rate of permeation and removal was not as great as the rate of offgassing.
To test whether flushing with clean air enhanced loss of offgassed substan-
ces, and thus made permeation appear relatively more significant in labora-
tory-exposed bags, additional experiments were performed.
®
In the first experiment, three Teflon bags were flushed three times
with clean air, refilled with clean air, and the contents chromatographed.
After analysis, all bags were placed in a steel box which had been purged
overnight with clean air. The clean air purge was continued overnight to
insure removal of contaminants. A 'chromatogram of the air in the steel box
71
-------
Table 11. COMPARISON OF STORAGE ENVIRONMENT ON BAG BACKGROUND
Film Type
Batch #
Environment
Ratio of Total
Area Counts
Day 7/Day 0
Tedlar
Tedlar
Tedlar
Tedlar
Teflon
Teflon
Teflon
Teflon
1
2
1
2
1
2
1
2.
Clean air
Clean air
Lab air
Lab air
Clean air
Clean air
Lab air
Lab air
0.72
2.30
13.68
20.33
16.60
3.64
40.40
7.26
72
-------
on the morning of Day 1 revealed very little contamination compared to
laboratory air.
On Day 3, the air in the steel box was again chromatographed to deter-
mine the extent of leakage and self-contamination. Subsequently, the
contents of all 3 Teflon bags were chromatographed. Representative chroma-
®
tograms from Days 0, 3, and 6 are shown in Figures 20-22 for the Teflon
bags; chromatograms of steel box, room air, and clean air are shown in
Figures 23-26. The resulting peak areas are shown in Table 12.
®
To further test the mode of bag contamination, each Teflon bag was
stored in the laboratory air for an additional three days. The peak areas
after laboratory storage are also shown in Table 12. The increase in total
peak area between Day 0 and Day 3 can be attributed to leakage of room air
into the steel box and/or offgassing from the walls of the box and permeation
of these contaminants into the bag rather than offgassing within the bag.
The much greater increase in total peak area after storage in laboratory air .
supports this idea since offgassing would not have been promoted simply by
transfer of the bags to a laboratory air environment.
The second experiment was identical to the first except bags made from
Tedlar film were used in place of the Teflon . Representative chromato-
grams from Day 0 and Day 3 are shown in Figures 27-32 along with chromato-
grams of room air, clean air, and steel box air. Total peak areas are shown
in Table 13.
The contamination of air in the bags occurs principally by permeation
of organic molecules through the bag walls. There is also a slight possibi-
lity that water molecules in large excess will displace organic molecules
trapped within and on the bag walls. The extent of this displacement would
be expected to be small however because of the hydrophobic nature of the
§ ®
Teflon and Tedlar . The possibility of this phenomenon is supported by
these two studies since the air used to flush the steel box had been dried
to some extent. Regardless, it appears that storage could be feasible if
the bag environment is kept clean and relatively dry to prevent contamination.
Further support for the significance of the contribution of offgassed
®
material to Teflon bag contamination is given by Lonneman, e_t al. (19).
®
Their study showed that heat pretreatment of Teflon bags to 190°C was
73
-------
Figure 20. Chromatogram of air in Teflon bag #13; T , December. 16, 1980.
0'
-------
Figure 21. Chromatogram of air in Teflon bag #13; T , December 19, 1980.
-------
Figure 22. Chromatogram of air in Teflon bag #13; T,, December 22, 1980.
-------
Figure 23. Chromatogram of air in steel box used for Teflon bag experiment;
T December 17, 1980.
-------
Figure 24. Chromatogram of laboratory air used for Teflon bag experiment; December 16, 1980.
-------
Figure 25. Chromatogram of clean air used for Teflon bag experiment; T , December 16, 1980.
-------
CO
o
Figure 26. Chroinatogram of air in steel box used for Teflon bag experiment;
T December 19, 1980.
-------
Table 12. EFFECT OF STORAGE ENVIRONMENT ON TEFLON BAG BACKGROUND
Sample Source
Teflon® Bag #12
Teflon® Bag #13
Teflon® Bag #14
Steel Box
Room Air
Clean Air
Day 0
1028b
367
371
462
26,629
95
Day 3
520
2102
2173
7300
-
-
Day 6a
15,901
13,271
23,698
-
-
-
o
Bags stored in lab air on days 4-6.
Area in arbitrary units.
Table 13. EFFECT OF STORAGE ENVIRONMENT ON TEDLAR BAG BACKGROUND
Sample Source
Tedlar Bag #7
Tedlar Bag #9
Tedlar Bag #10
Steel Box
Room Air
Clean Air
Day 0
342a
545
659
438
20,278
162
Day 3
2648
1150
1423
14462
--
--
Area in arbitrary units.
81
-------
00
to
Figure 27.. Chromatograin of air in Tedlar bag #IX; T , December 30, 1980.
-------
00
Figure 28. Chromatogram of air in Tedlar bag #IX; T , January 2, 1981.
-------
A. JW^
Figure 29. Chromatogram of clean air used for Tedlar bag experiment; December 30, 1980.
-------
00
Cn
Figure 30. Chromatogram of air in steel box used for Tedlar
January 2, 1981.
bag experiment; T
3'
-------
CO
Figure 31. Chromatogram of air in steel box used for Tedlar bag experiment; T ,
December 31, 1980.
-------
©
Figure 32. Chromatogram of laboratory air used for Tedlar bag experiment; December 30, 1980
-------
. necessary to prevent extensive offgassing but that the heavy molecular
weight fraction was not affected as significantly as the low molecular
fraction. Even after taking these measures, the contamination problem was
not completely eliminated.
Regardless of the mode of contamination, storage of polymeric bags
containing air samples in containers where the bag environment is contaminated
is not acceptable if low level hydrocarbon measurements are to be made. The
degree of contamination of polymeric bags also appears to be batch dependent.
It was decided that experiments with bags would be continued although
these would involve only short-term storage of high concentration levels of
test compounds in a "clean" environment.
Storage-Stability Studies
Containers--
® ®
Bags—Group I compounds were loaded into Teflon and Tedlar bags on
©
January 8, 1981. The Teflon bags were filled with about 10 liters of
®
sample by attaching them directly to the glass manifold with a 1/4" Teflon
(6)
tube (Section 5, Figs. 1 and 3). The Tedlar bags were filled with 20
liters of sample by the same method. Three bags of each type.were filled
with sample and one bag of each type was filled with clean air to serve as a
control. The polymeric bags were stored in the steel box described previously.
After placing the bags in the box, it was flushed with clean, dry air for
about 20 hours and then sealed. The steel box was also flushed overnight
after the Day 4 analyses to remove any laboratory air contamination which
could have occurred during opening.
Analyses of the polymeric bags was performed on 0, 3 and 7 days of
storage. (Day 0 analyses preceeded placing the bags in the steel box).
Aliquots (200 mL) of the gas in each bag were taken into the cryogenic trap
described previously without heating the bags. The cryogenically trapped
organic compounds were then volatilized and measured using gas chromatography.
The GC conditions are given in Table 14. The rasults of the analyses are
shown on Tables 15 and 16. The recovery of the lower molecular weight
®
compounds from Teflon bags was highest on Day 0. The higher molecular
weight compounds however, showed lower recoveries on Day 0, usually 70-80
percent. All compounds, excluding chloroprene and chloroform, showed large
-------
Table 14. GC PARAMETERS FOR ANALYSIS OF CONTAINERS
Parameter
Setting/Conditions
Column
Carrier gas
Make-up gas
Column temperature
FID - Air flow
- H. flow
GC
Detector temperature
60 meter SCOT SE-30;
0.5 mm i.d.
He 8 mL/min
He 23 mL/min for FID
5% CH4 in Ar - 40'mL/min
for ECD
4 min @ 30°C, 8°C/min,
2 min @ 185°C
40 psi
17 psi
Perkin Elmer Model 3920
200°C
89
-------
Table 15. AVERAGE PERCENT RECOVERY OF GROUP I COMPOUNDS FROM TEFLON w BAGS
Compound
Vinyl chloride
Methyl bromide
Furan/acrylonitrile
Chloroprene
Chloroform
Benzene
1 ,2-Dichloropropane
Toluene
Tet rachloroethylene
Chlorobenzene
1 ,1 ,2,2-Tetrachloroethane
m-I)i Chlorobenzene
1,2,3-Trimethylbenzene
Bis(2-chloroethyl)ether
I'pb Sampled
28.6
72.2
142.8
5.2
14.3
53.7
25.2
22.7
27.7
15.9-
33.6
18.5
94.8
19.3
T (0 days)
95.5 + 4.2 (4.4)
100.7 + 0.8 (0.8)
99.2 + 2.2 (2.3)
84.6 + 16.7 (19.8)
74.1 + 5.6 (7.5)
101.9 + 10.8 (10.6)
37.7 + 1.2
85.0 J 2.6 (3.0)
79.4 + 0.0 (0.0)
75.5 + 0.0 (0.0)
72.3 + 1.8 (2.5)
70.3 t 5.4 (7.7)
69.6 + 1.1 (1.5)
72.5 + 5.2 (7.1)
Percent Recovery at Time3
T (3 days)
88.5 + 5.2 (5.9)
84.1 + 5.5 (6.6)
86.1 + 5.3 (6.1)
89.0 + 25.6 (28.7)
500.6 + 190 (38)
81.9 + 3.7 (4.5)
30.0 + 0.0
71.8 + 2.6 (3.7)
56.7 + 2.2 (3.8)
55.2 + 2.2 (4.0)
72.3 + 19.6 (27.2)
56.8 + 16.8 (29.5)
48.2 + 1.6 (3.3)
49.7 + 8.5 (17.1)
T (7 days)
79.4 + 14.7 (18.5)
66.1 + 14.4 (21.8)
69.3 + 18.6 (26.8)
136.5 + 86.3 (63.3)
265.7 + 189 (71)
78.8 + 15.5 (19.6)
31.0 + 3.0
64.8 + 11.0 (16.9)
49.5 + 5.4 (10.9)
44.0 + 8.2 (18.7)
78.3 + 30.1 (38.4)
67.0 + 26.5 (39.5)
48.5 + 2.1 (4.3)
36.3 + 4.9 (13.6)
Number in parenthesis is coefficient of variation.
-------
Table 16. AVERAGE PERCENT RECOVERY OF GROUP I COMPOUNDS FROM TEDLAR BAGS
Compound
Pph Samp 1ed
Percent Recovery at Time
T (0 days)
T (3 days)
T (7 days)
Vinyl chloride
Methyl bromide
Furan/acryloniLrile
Chloroprene
Chloroform
Benzene
1,2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2,2-Tetrachloroethane
m-Dichlorobenzene
1,2,3-Trimethy 1benzene
Bis(2-chloroethyl)ether
cf
b
28.6
72.2
142.8
5.2
14.3
53.7
25.2
' 22.7
27.7
15.9
33.6
18.5
94.8
19.3
98.9 + 2.1 (2.1)
100.7 + 2.1 C2.1)
100.4 + 0.4,(0.4)
100.0 + 8.8 (8.8)
66.1 t 1.5 (2.2)
94.4 t 3.9 (4.1)
31.7 + 1.5b
88.1 + 0.0 (0.0)
81.9 + 2.2 (2.6)
81.8 + 0.0 (0.0)
72.3 + 1.8 (2.5)
70.3 + 0.0 (0.0)
71.4 + 1.3 (1.8)
76.2 + 3.1 (4.1)
97.9 + 5.9 (6.1)
97.9 + 4.4 (4.5)
98.0 + 4.9 (5.0)
93.7 + 4.0 (4.3)
473.4 + 418 (86)
'94.9 + 1.9 (1.9)
33.0 + 1.0b
82.4 + 2.6 (3.2)'
78.3 + 2.2 (2.8)
71.1 + 3.8 (5.3)
66.4 + 1.8 (2.7)
63.2 + 3.2 (3.2)
63.6 + 0.6 (0.9)
58.5 + 3.1 (5.3)
110.3 t 5.2 (5.2)
97 ..9 + 5.8 (5.9)
98.9 + 5.0 (5.1)
467 + 323 (69)
70.6 + 6.9 (9.9)
94.9 + 6.7 (7.1)
33.0 + 1.7b
85.0 + 5.3 (6.2)
80.5 * 4.3 (5.4)
73.6 + 3.8 (5.1)
67.6 » 1.8 (2.6)
64.9 + 5.4 (8.3)
66.B + 3.2 (4.7)
60.6 + 3.1 (5.1)
lis(2-chloroethyl)ether 19.3 76.2 + 3.1 (4.1) 58.5 + 3.1 (5.3)
Number in parenthesis is coefficient of variation.
'low recovery due possibly to variability/failure of permeation tube or system
-------
decreases in recovery from Day 0 to Day 7. Unknown interferences prevented
quantitation of the chloroprene and chloroform. The extent of interference
was variable and could not be accurately quantified by subtraction of the
blank determined using the control bag. The large increase in the variability
of the results on Day 7 indicates bag variability rather than measurement
variability. There was a significant increase in the background levels in
the control bag, even with storage in the clean atmosphere of the steel box.
A contributing source of control bag background apparently is compound
permeation out of the sample bags and into the control bag.
The recoveries from Tedlar bags (Table 16) were generally higher than
®
those with Teflon both on Day 0 and Day 7. Also the recovery variability
®
was much less with Tedlar on Day 7. Interferences again prevented quantita-
tion of chloroprene and chloroform.
A second group of compounds (Group II) was used for testing in June,
1981. The test samples were generated with the permeation/dilution system
described earlier but which was modified for these measurements. The permea-
tion rates were low for these compounds and thus extensive dilution was not
necessary. The mixing bulbs were replaced with a straight, one-inch diameter
glass pipe and dilution air was flowed directly across the permeation tubes
to achieve below 100 ppb for each compound. This modification was also used
in testing traps.
®
Of the polymeric bags only Tedlar bags were tested with Group II com-
®
pounds. Tedlar was used as it showed better overall recovery and recovery
®
precision than the Teflon . Again the bags were.filled directly from the
manifold. All previous experiments indicated that the bags could only be
used for storage for a short period of time unless they were maintained in a
®
"zero air" environment. As a practical matter then, the Tedlar bags were
analyzed on Day 0 only and without storage, or within a matter of hours
after they were filled; this was considered a realistic test of the usefulness
of the bags. By applying the appropriate calibration factors, the ppb level
of each compound in the bags on Day 0 was determined. The relative percent
recovery was calculated from these measured ppb levels and the expected
levels of each compound. These results are presented in Table 17. The best
recovery was obtained with methyl chloride, vinylidene chloride and allyl
92
-------
Table 17. PERCENT RECOVERY OF GROUP II COMPOUNDS FROM TEDLAR BAGS '
Compound
Methyl chloride
Methyl mercaptan
Propylene oxide
Vinylidene chloride
Allyl chloride
1,1, 1-Trichloroethane
a-Epichlorohydrin
Ethylbenzene
o-Xylene
n-Decane
1,2, 3-Trimethylbenzene
o-Cresol
Nitrobenzene
ppb Sampled
56.6
59.6
27.6
20.9
91.4
12.9
67.2
7.5
21.2
7.8
72.2
21.4
27.2
r%
Percent Recovery (Day 0)
83.3 + 1.1 (1.3)
64. 4b
Not detected
85.2 +5.7 (6.7)
100.3 + 2.1 (2.1)
74.4 + 6.2 (8.3)
Not detected
88.0 +. 10.7 (12.2)
34.4 + 3.3 (9.6)c
76.9 + 20.5 (26.6)
29.5 + 1.5 (5.1)c
Not detected
65.8 + 6.6. (10.0)
Number in parenthesis is coefficient of variation.
Single observation.
c
Low recovery due possibly to variability/failure of permeation
tube or system.
93
-------
chloride. Propylene oxide and a-epichlorohydrin were not detected in the
®
Tedlar bags and only weakly and erratically detected as standards. o-Cresol
® ~
was not detected in the Tedlar bags or as a standard, possibly because of
its boiling point which is 190.9°C.
Glass Bulbs—The two groups of compounds used to test Tedlar bags were
also used to test the glass bulbs. Each glass bulb was pressurized to about
10 psig by passing sample at a rate of 300 cc/minute through a metal bellows
pump and into the bulb. The bulbs were stored in boxes to protect them from
light. At the appropriate times, the samples were withdrawn from the bulbs
with the bulbs at room temperature. The results of these analyses are shown
in Table 18. Glass bulbs show a general decrease in recovery of compounds
with increasing boiling point. A small decrease in recovery occurred gene-
rally from Day 0 to Day 3 and an increase occurred from Day 3 to Day 7.
There is no ready explanation for this increase which occurred with the
groups of compounds studied in January and June, respectively. One possibi-
lity though, is a relatively rapid loss of the organic compounds to the
glass and then slow displacement from the glass by the small amount of water
in the sample gas. A large interference was observed on some of the sample
chromatograms of the first sample group on Days (3) and (7) between 5 and 15
minutes retention time. This interference, which prevented the measurement
of the methyl mercaptan, was thought to be due to contamination from the o-
®
rings to seal the Teflon stopcocks.
Stainless Steel Cannisters--The two types of steel containers, electro-
polished and "Summa" polished, were tested with the first group of compounds
whereas only the "Summa" polished containers were also tested with the
second group of compounds. The electropolished containers were not included
in the second test as they yielded recoveries similar to but not quite as
good as those obtained with the "Summa" polished containers. A metal bellows
pump which was used to fill the cans with sample was first equilibrated by
drawing sample through it for several minutes. The pump outlet was then
connected to the container valve and each container filled at a rate of 300
cc/min. The electropolished stainless steel cans were pressurized to 15
psig which corresponded to 4 liters of sample. The "Summa" polished cans
were pressurized to 10 psig which corresponded to 12 liters of sample.
94
-------
Table 18. PERCENT RECOVERY FROM GLASS BULBS
Compound
Methyl chloride
Vinyl chloride
Methyl bromi de
Methyl me reap tan
Fu ran/ aery loni tri le
Propy'lene oxide
Vinylidene chloride
Allyl chloride
Chloroprene
Cliloroform
1,1, 1-Trichlo roe thane
Benzene
1 ,2-Dichloropropane
Toluene
a-Epichlorohydrin
Tetrachloroethylene
Chlorobenzene
Ethylbenzene
o-Xylene
1 ,1 ,2,2-Tetrachloroethane
m-Di chlorobenzene
n-Decane
1, 2, 3-Trimethyl benzene
ppb Samp led
56.
-.28.
72.
59.
142.
27.
20.
91.
5.
14.
12.
53.
25.
22.
67.
27.
15.
7.
21
33
18
7
94.
6
6
2
6
8
6
9
4
2
3
9
7
2
7
2
7
9
5
.2
.6
.5
.8
.3
TQ (0 days)3
76.
96.
98.
97.
105.
92.
94.
76.
65.
8 +
9 +
8 +
.8 +
7 t
3 +
2 +
9 ±
9, ±
87.
3.0
2.1
1.7
3.6
18.7
6.4
5.4
0.0
10.8
.5
(3.9)
(2.2)
(1.7)
(3.7)
(17.1)
(6.9)
(5.7)
(0.0)
(16.4)
74.
95.
97.
Interference
95.
Not detected
100.
92.
91.
62.
34b
85.
79.
75
66
25
66
63
51
62
.0 +
.4 +
.5 +
.7 +
.5 +
.4 t
.2 +
.3 +
.2 +
2.6
0.0
0.0
1.3
1.9
1.8
3.2
5.1
2.1
(3.1)
(0.0)
(0.0)
(1.9)
(7.4)
(2.7)
(5.1)
(9.9)
(3.4)
83.
Not detected
78
75
65
24
65
63
43
60
T3 (3 days)3
5 +
5 +
4 +
8.1
2.1
0.8
(10.9)
(2.2)
(1.2)
T? (7 days)3
79.7 + 7.6 (9.5)
96.9 t 2.1 (2.2)
98.8 + 0.8 (0.8)
- not detected
0 +
9 +
0 +
3 +
69.
0 +
81.
1.1
16.3
6.4
4.0
9
7.8
,9
(1.1)
(16.2)
(7.0)
(4.4)
(12.6)
34b
.7 +
.3 +
.5 +
.3 +
.0 +
.5 +
.2 +
.6 +
.2 +
0.0
2.2
0.0
5.3
1.9
0.0
3.2
3.8
4.4
(0.0)
(2.8)
(0.0)
(B.I)
(7.9)
(0.0)
(5.1)
(8.7)
(7.4)
99.2 * 4.6 (4.6)
75.1 + 13.4 (17.8)
96.1 + 5.9 (6.1)
91.3 + 1.3 (1.4)
69.9
68.2 + 3.9 (5.7)
83.8
32b
85.9 + 3.1 (3.6)
81.2 + 2.5 (3.1)
77.4 + 3.8 (4.9)
86.7 + 8.0 (9.2)
34.9 + 4.7 (13.5)
69.3 + 1.8 (2.6)
68.6 + 3.2 (4.7)
64.1 + 19.2 (30.0)
66.8 + 3.2 (4.7)
(continued)
-------
Table 18 (cont'd.)
Compound
Bis(2-chloroethyl)ether
o-Cresol
Nitrobenzene
Benzyl chloride
aNumber in parenthesis
ppb
19
21
27
29
Sampled
.3
.4
.2
.2
is coefficient
Tfl (0 days)
79.3 + 3.1 (3.9)
66.9 + 9.2 (13.8)
42.5 + 3.4 (8.0)
of variation.
72
Not detected
61
37
T3
.5 +
.4 +
.7 +
(3
0,
0,
2
days)
.0 (0.0)
.4 (0.6)
.0 (5.3)
T7
77.7 +
62.5 +
47.6 +
(7
0.
0.
6.
d
0
4
2
ays)
(0.0)
(0.6)
(13.0)
-------
il
Three of each container type were filled with sample, and one of each type
was filled with clean air to serve as a control. Analyses of each steel
container involved placing the container in the small oven described previo-
usly and heating it to >90°C for at least 5 minutes prior to sample removal.
As with the other containers, 200 mL of sample was taken for each measurement.
The percent recovery for Group I compounds from the electropolished
containers is presented in Table 19. The percent recovery decreases with
increase in boiling point. The low boiling compounds show a modest loss in
recovery with time while the high boiling compounds generally show a drop in
recovery from Day 0 to Day 3 and then a leveling off in recovery. Two
compounds which produced inconsistent results with variation in the boiling
point were 1,2-dichloropropane and bis-(2-chloroethyl)ether. The low recovery
of the former and high recovery of the latter could have been due to a
change in the permeation tubes occuring between filling the containers and
calibration of the detector response.
The recovery values for the "Summa" polished containers are presented
in Table 20. As with the electropolished container, recovery generally
decreases with increase in boiling point. Also there is generally a decrease
in recovery from Day 0 to Day 3 but no analytically meaningful change in
recovery from Day 3 to Day 7 for a majority of the compounds.
As with the glass bulbs, recovery of some of the compounds was higher
on Day 7 than on Day 3 or Day 0. One possible explanation for this is that
the "Summa" cans were heated for a longer period of time on Day 7 than on
either Day 3 or Day 0 before beginning to trap-out a sample. The average
heating time on Day 0 was five minutes, on Day 3 was 20 minutes, and on Day
7 was 30 minutes. This could have lead to a greater fraction of the compounds
being desorbed from the can wall on Day 7. Even though this variance in
heating time was unintentional, some useful information may have been revealed.
The "instability" of many compounds may not be a decomposition with irrever-
sible loss but only the adsorption of the compounds on a cold surface.
Complete recovery of these compounds may be possible by heating the container
to a sufficient temperature to promote desorption but not thermal decomposi-
tion.
97
-------
Table 19. RELATIVE PERCENT RECOVERY FROM ELECTROPOLISHED STEEL CONTAINERS
00
Compound
Vinyl chloride
Methyl bromide
FII ran/aery Ion itrile
Chloroprene
Chloroform
Benzene
1 ,2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1 ,1,2,2-Tetrachloroethane
m-l)i Chlorobenzene
1,2,3-Trimethylbenzene
Bis(2-chloroethyl)ether
Ppb Sampled
30.9
78.3
154.9
5.65
15.5
58.3
27.2
24.2
30.1
17.3
36.4
20.0
102.9
20.9
T (0 days)
98.1 + 1.9 (1.9)
93.6 + 3.8 (4.1)
93.6 + 4.8 (5.2)
87.3 + 5.3 (6.1)
66.5 + 3.9 (5.8)
96.6 + 6.5 (6.7)
57.0 + 4.4b
86.6 + 2.4 (2.8)
93.0 + 6.6 (7.1)
86.7 + 0.0 (0.0)
49.2 + 25.3 (51.4)
46.9 + 21.9 (46.7)
44.9 + 22.7 (50.6)
81.3 + 19.1 (23.5)
Percent Recovery at Time
T (3 days)
96.1 + 1.9 (2.0)
91.1 + 8.2 (8.9)
89.3 + 6.9 (7.7)
77.9 +. 10.8 (13.9)
70.9 + 0.0 (0.0)
89.2 + 2.9 (3.3)
55.3 + 3.0b
82.5 + 4.9 (5.9)
92.0 + 4.9 (5.4)
80.9 + 0.0 (0.0)
41.8 + 26.4 (63.1)
38.2 + 14.9 (39.1)
38.2 + 15.8 (41.5)
79.9 * 13.9 (17.4)
T (7 days)
94.8 + 6.8 (7.2)
85.6 + 7.2 (8. A)
81.9 t 9.7 (11.8)
69.6 + 5.7 (8.1)
70.3 + 11.6 (16.5)
93.8 + 10.5 (11.2)
57.0 + 7. Ob
84.1 + 6.1 (7.3)
94.0 + 15.6 (16.6)
86.7 + 9.8 (11.3)
37.1 + 25.0 (67.4)
35.5 + 10.5 (29.6)
36.2 + 6.9 (19.1)
87.6 + 7.2 (8.2)
Number in parenthesis is coefficient of variation.
''Low recovery due possibly to variation/failure in permeation tube or system.
-------
Table 20. RELATIVE PERCENT RECOVERY FROM SUMMA POLISHED STEEL CONTAINERS
Compound
Methyl chloride
Vinyl chloride
Methyl bromide
Methyl mercaptan
Furan/acryloni tri le
Propylene oxide
Vinylidene chloride
Ally! chloride
Chloroprene
Chloroform
1,1, 1-Trichloroethane
Benzene
1 , 2-Dichloropropane
Toluene
a-Epichlorohydrin
Tet rachloroethylene
Chlorobenzene
Ethy Ihenzene
o-Xylene
1 ,1,2,2-Tetrachloroethane
m-Di Chlorobenzene
n-Decane
1 ,2,3-Trimethylbenzene
Ppu Sampled
56
30
78
59
154
27
20
91
5
15
12
58
27
24
67
30
17
6
21
36
20
7
72
.6
.9
.3
.6
.9
.6
.9 "
.4
.65 .
.5
.9
.3
.3
.6
.2
.1
.3
.5
.2
.4
.0
.8
.2
T (0 days)
81
101
99
' 97
125
97
95
72
69
.97
51
90
90
90
80
34
71
102
93
.4 + 6.4
.3 + 1.9
.6 + 0.0
50.0
.9 + 3.9
.4 + 21.5
.7 + 7.6
.0 + 3.7
.9 + 3.9
.0 + 2.3
.8 + 7.9
.0 + 4.4°
.7 + 2.4
.7 + 4.9
.8 + 6.9
.0 + 4.0
.4 + 1.4
75 * 1.6
.5 * 3.0
.6 + 15.4
.5 + 15.2
(7.9)
(1.9)
(0.0)
(4.0)
(17.1)
(7.8)
(3.9)
(5.3)
(3.3)
(8-1)
(2.7)
(5.5)
(7.6)
(5.0)
(4.1)
(2.2)
(4.2)
(15.0)
(16.2)
83.
97.
95.
_ _ 89 ._
Not detected
119.
97.
83.
70.
69.
90.
45.
85.
Not detected
87.
86.
80.
34.
69.
60.
79.
70.
T (3 d
2- + 3.9
1 + 0.0
4 t 0.8
40. 7a
1 + 3.6
6 + 6.7
7 + 7.8
7 + 5.1
9 + 0.0
8 + 7.0
4 + 2.6
7 + 5.7
4 + 4.1
4 + 4.9
7 + 5.8
0 + 12.
0 + 5.2
5 + 3.3
0 + 5.0
5 1 5.1
9 + 8.4
v at Tim<>a
ays)
(4-7)
(0.0)
(0.1)
(4.1)
(5.6)
(8.0)
(6.1)
(0.0)
(10.0)
(2.8)
c
(4.8)
(5.7)
(6.7)
0 (15.0)
(15.3)
(4.7)
(8.3)
(6.4)
(11.8)
T (7 days)
' 78.
98.
97.
58,
86.
113.
98.
81.
75.
74
95
52
88
89
88
94
41
72
56
96
89
.1 +
.1 +
.4 +
.8 +
.7 +
,9 +
.8 +
.9 +
.5 +
.4 +
•5 1
.7 +
.2 +
.7 +
.4 +
•7 ±
.5 t
.3 +
.0 +
.2 +
.7 +
27.2 (34.8)
1.9 (2.0)
1.5 (1.6)
24.7 (50.6)
5.6 (6.4)
-
45.9 (40.3)
7.2 (7.3)
6.2 (7.6)
3.9 (5.1)
2.3 (3.1)
6.9 (7.2)
3.2C
4.9 (5.5)
5.6 (6.3)
8.7 (9.8)
8.0 (7.8)
2.8 (6.7)
4.1 (5.7)
7.0 (12.5)
16.7 (17.4)
22.2 (24.7)
(continued)
-------
Table 20 (cont'd.)
Percent Recovery at Time
Compound
Bis(2-ch loroethyl )ether
o-Cresol
Nitrobenzene
Benzyl chloride
Ppb Sampled
20.9
21.4
27.2
29.2
T (0 days)
106.7 t 11.0 (10.
123.2 + 43.8 (35.
41.4 + 12.0 (29.
T (3 days)
.3)
.6)
0)
99.0
Not detected
103.3
Not
+ 10.0 (10.1)
+ 15.8 (15.3)
detected
. T (7 days
101.9 + 11.0
124.3 + 48.0
0
(10.8)
(38.6)
Not detected
aNumber in parenthesis is coefficient of variation.
Single observation.
GLow recovery is due possibly to variation/failure of permeation tube.
o
o
-------
Traps —
The test parameters employed for the storage-stability study are given
in Table 21. The sampling volume (30 L) and the relative humidity (30%) were
held constant. No ozone, NO or SO. were added. The variable parameters
X ^
were the concentration of each individual substance and storage time.
Sampling rate and time were held constant throughout all of the experiments.
All sampling devices were evaluated simultaneously by sampling the
air/vapor mixture concurrently to eliminate possible variability in the
performance of the permeation/dilution system.
Tables 22 and 23 present the concentrations and the total quantity of
the test compounds which were delivered to the sampling devices. The levels
in ppb, total weight (ng) , and the breakthrough volumes (Table 22) for these
compounds on Tenax GC cartridges are indicated. The range of concentrations
were 15-100 ppb and 2-858 ppt for the high and low level studies, respectively.
The total quantity delivered to the sampling devices was based upon sampling
30 L of the air-vapor mixture from the permeation/dilution system.
Tenax GC Cartridges—The storage-stability study using Tenax GC cartrid-
ges was conducted according to the experimental design described above.
Triplicate samples were collected for each experimental parameter (variable).
All samples and blanks were collected during T_ (T~ = 0 day of storage .
study). Analysis of T_ samples was performed within three hours of .collection.
The absolute areas, which were obtained for the samples analyzed at T_,
T- and T? were used to obtain the quantity of the material recovered by
interpolation from calibration curves [response (area) vs. quantity]. Using
the quantity (nanograms) of the material measured on the sampling cartridge
the percent recovery was calculated as a ratio of observed to expected times
100. .
The results for high and low level storage-stability are given in
Tables 24 and 25, respectively.
All compounds were detected in the high level study (Table 24). Quanti-
tative recoveries were observed for most compounds that had breakthrough.
volumes greater than the sample volume (30 L). The apparent lower recovery
of benzene is probably due to an uncertainty in the initial instrument
calibration, since subsequent studies have shown recoveries at least 25%
101
-------
Table 21. TEST PARAMETER RELATIONSHIPS FOR EVALUATION OF TRAP TYPE
COLLECTION DEVICES - STORAGE/STABILITY STUDY
Constant Parameters Variable Parameters
Volume - 30 SL Concentration - ppt and ppb
RH - 30% Storage time - t , t , t
(da) ° J 7
[03] = 0
[NO ] = 0'
[S02] = 0
102
-------
Table 22. CONCENTRATIONS (HIGH LEVEL STUDY) AND TOTAL QUANTITY OF GROUP I COMPOUNDS DELIVERED TO
SAMPLING DEVICES: TENAX GC, CHARCOAL, AND CRYOGENIC TRAPS
Compound
Vinyl chloride
Methyl bromide
Acrylonitrile
Furan
Chloroprene
Chloroform
Benzene
1 , 2-Dichloropropane
Toluene
Tetrachloroethylene
1,1,2, 2-Tetrachloroethane
Chlorobenzene
Bis- (2-chloroethyl) ether
m-Di chlorobenzene
Concentration
(ppb)
48
80
93
100
4.5
15
60
24
24
30
41
17
48
20
Total Wght: Delivered
(ng)
3,672
5,587
6,054
8,340
487
2,196
5,742
3,326
2,714
6,102
8,438
2,346
8,410
3,606
Breakthrough Volume (£)
(@80°F)
1.0
1.0
5
3
15
18
38
81
173
144
173
344
234
948
On standard Tenax GC cartridge.
-------
Table 23. CONCENTRATIONS (LOW LEVEL STUDY) AND TOTAL QUANTITY OF GROUP I COMPOUNDS DELIVERED TO
SAMPLING DEVICES: TENAX GC, CHARCOAL, AND CRYOGENIC TRAPS
Compound
Vinyl chloride
Methyl bromide
Furan
Acrylonitrile
Chloroprene
Chloroform
Benzene
1 , 2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2, 2-Tetrachloroethane
Bis-(2-chloroethyl)ether
m-Di chlorobenzene
Concentration
(ppt)
17
429
542
858
2
82
328
- 184
132
161
95
351
252
122
Total Wght. Delivered
(ng)
13
50
46
57
2
12
32
25
15
33
13
73
44
22
Concentration
(pg/L)
428
1,675
1,517
1,887
57
400
1,050
845
502
1,095
437
2,420
1,460
735
-------
Table 24. PERCENT RECOVERY FOR HIGH LEVELS OF TEST COMPOUNDS FROM TENAX GC CARTRIDGES -
WITH CORRECTION FOR BREAKTHROUGH VOLUME
o
Ul
Compound
Methyl chloride
Vinyl chloride
Methyl bromide
Methyl mercaptan
Furan
Propylene oxide
Acrylonitrile
Vinylidene chloride
Allyl chloride
Chloroprene
Chloroform
1,1, I-Trichloroethane
Benzene
1 ,2-Di chloropropane
Toluene
or-Epichlorohydrin
Tetrachloroethylene
Chlorobenzene
Ethylbenzene
o-Xylene
1 ,1,2,2-Tetrachloroethane
m-Di chlorobenzene
n-Decane
B.P. Breakthrough Volume ppb
(°C) (L) Sampled
-24.2
-13
3.4
6.2
31.4
34.3
37
45
59.4
61.7
74.1
77
96.4
110.6
116.5
121
132
136.2
139.1
146.2
173
174.1
3
1
1
-
3
4
5
1
6
15
18
9
38
81
173
54
380
344
344
-
173
948
-
56.3
48
80
60.2
100
27.6
93
20.7
93.4
4.5
15
13.4
60
24
24
38.3
30
17
7.4
8.4
41
20
7.7
Expected Quantity
(ng/cartridge)
3,480
122
186
3,548
1,009
1,965
834
2,452
8,752
243
1,318
2,182
3,326
2,714
4,335
6,102
2,346
968
1,095
8,438
3,606
1,335
TQ = 0 da
9.7 + 3.5 (36)
52 + 29 (58)"
51+7 (14)
9.0 + 1.7 (19)
90+8 (8)
46 + 15 (32)
112 + 12 (15)
112 + 18 (16)
17.4 + 2.6 (15)
80 + 12 (15)
111 + 21 (19)
23 + 3.4 (15)
65+9 (14)
101 + 14 (14)
91 + 10 (11)
67 + 5.5 (8.2)
97+7 (7)
105 + 7 (7)
88 + 13 (15)
88 + 11 (12)
93+4 (4)
103 + 6 (4)
105 + 32 (31)
' Storage Period
T3 = 3 da
7.3 + 2.4 (33)
41+7 (17)
63 + 7 (11)
11.8 + 8.5 (71)
89 + 23 (26)
44 + 5.9 (13)
142 + 19 (13)
174 + 4.5 (3)
19.6 + 6.4 (32)
83 + 1 (1)
139 + 36 (26)
26 + 10 (39)
63+9 (14)
99+9 (9)
89 + 10 (12)
55 + 6.1 (11)
95 + 13 (i4)
104 + 8 (8)
88 + 12 (14)
88 + 3.0 (4)
88+6 (7)
125 + 5 (4)
111 + 33 (30)
Ty = 7 da
6.6 + 3.7 (56)
50 + 7 (14)
79+2 (3)
6.7 + 2.6 (38)
89 + 36 (40)
41 + 7.8 (19)
165 + 18 (11)
105 + 26 (24)
19.5 + 4.0 (20)
88+3 (3)
142 + 29 (20)
23 + 1.4 (6)
71+3 (4)
113 + 6 (6)
95 + 2 (2)
59 + 9.8 (16)
96+4 (4)
106 + 2 (2)
89 + 4.3 (5)
84 + 4.7 (6)
97+7 (3)
131 + 5 (4)
106 + 20 (19)
(continued)
-------
Table 24 (cont'd.)
Compound
1 ,2,3-Trimelhylbenzene
Bis-(2-chloroethyl)ether
o-Cresol
Nitrobenzene
Benzyl chloride
B.P. Breakthrough Volume ppb
(°C) (L) Sampled
176.1
178 234
190.9
210.8
215 830
61.1
48
21.8
28.2
12.1
Expected Quantity
(ng/cartridge)
8,992
8,410
2,888
4,252
1,868
To
85
106
138
115
146
= 0 da
+ 14 (16) .
1 5 (5)
+ 20 (14)
+ 22 (19)
+ 26 (18)
Storage Period
T3 = 3 da
79 + 9.0 (11)
85+2 (2)
89 + 26 (29)
108 + 9.7 (90
88 + 29 (33)
T
86
97
147
117
165
? = 7da
+ 5.8 (7)
+ 3 (3)
+ 22 (15)
+ 10 (8)
+ 30 (18)
o
o\
-------
Table 25. PERCENT RECOVERY OF GROUP I COMPOUNDS FROM TENAX GC TRAPS - LOW LEVEL STUDY
Compound
Vinyl chloride
Methyl bromide
Furan
Acrylonitrile
Chloroprene
Chloroform
Benzene
1 ,2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2,2-Tetrachloroethane
Bis-(2-chloroethyl)ether
m-Di chlorobenzene
Breakthrough Volume ppt
(1.) Sampled
1
1
3
5
15
18
38
81
173
380
344
173
234
948
17
429
542
858
2
82
328
184
132
161
95
351
252
122
Kxpected Quantity
(ng/cartridge)
0.428
1.67
4.55
9.43
0.85
7.2
31.5
25.3
15.1
32.8
13.1
72.6
43.8
22.0
T0
NDa
ND
ND
ND
ND
ND
BIb
49.2 + 8.9 (21)
71.7 + 9.3 (13)
BI
43.2 + 9.3 (22)
BI
BI
51.4 + 6.4 (12)
Storage Period
T5
ND
ND
ND
ND
ND
ND
BI
37.7+7
53. 3C
89.1
37.6 + 24 (64)
BI
BI
22.0 + 3.4 (16)
T7
ND
ND
ND
ND
ND
ND
BI
63.2 + 39
89.4
110.8
77.6
BI
BI
32.4 + 4.
.8 (63)
7 (14)
-------
higher than indicated here. There was no statistically significant trend
indicating a decrease in recovery as a function of storage, nor was there a
decrease in precision (Table 24). The reactivity of benzyl chloride and
a-epichlorohydrin made it extremely difficult to calibrate, collect and
analyze samples.
For chemicals with breakthrough volumes less than the sampling volume,
the absolute recoveries were poor for many compounds as predicted (Table
26). After applying a correction for breakthrough, the recoveries were
significantly better. For some chemicals, relative recoveries were still
low. In all cases it is evident that the percent relative standard deviation
for precision was considerably higher for these chemicals with low break-
through volumes. Several reasons may apply. The first is that the absolute
®
quantity accumulated on the Tenax GC cartridges was small and coupled with
high background of volatile organics, precise measurements were not possible
with flame ionization detection. The use of a more specific detector, such
as mass spectrometry (mass chromatography) reduces potential interferences.
Secondly, these chemicals are easily influenced by displacement from more
tenacious chemicals in the mixture sampled, an important factor to consider
when using "chromatographic adsorbent traps". Finally, primary sources of
standards for instrument calibration as certified standards are not available
and thus uncertainty exists with accuracy of measurements.
It is interesting to note that even a compound (e.g. furan) with a
breakthrough volume 1/10 of the sampling volume gave good recoveries based
upon the breakthrough volume. Also, even at a total level of upto 600 ppb
of vapors (for Group I compounds), premature breakthrough of furan, acryloni-
trile, chloroprene, chloroform, benzene, etc. did not occur. Total levels
higher than 600 ppb were not tested in this study. Thus, these data imply
that only semi-quantitative determination for some vapor-phase organics can
be at best expected for constituents with breakthrough volumes less than the
sampling volume.
A chromatogram representing the analysis of a Tenax cartridge stored
for seven days is depicted in Figure 33 (negative deflections are integrator
event marks in all chromatograms).
108
-------
Table 26. PERCENT RECOVERY OF TEST COMPOUNDS FROM TENAX GC TRAPS -
DISREGARDING BREAKTHROUGH VOLUME (HIGH LEVEL STUDY)
Storage Period
Compound T = 0 da T, = 3 da T = 7 da
Vinyl chloride 1.72 + 0.98a 1.38+0.22 1.67+0.24
Methyl bromide 1.69 + 0.24 2.11 + 0.24 2.62 + 0.08
Furan 7.89 + 0.68 7.87 + 1.99 7.85 + 3.15
Acrylonitrile 20+4.2 25+3.37 29+3.15
Chloroprene 19+2.9 19+0.15 21+0.63
Chloroform 52+9.8 65+16 66+13
aPercent + S.D.
109
-------
o
8-
0>
CL
r
•B
0>
9
Figure 33. FID chromatogram of T Tenax sample (high level study)
-------
In contrast, only a few of the test compounds were detected in the low
level study (Table 25) using TD/HRGC with flame ionization detection. The
presence of background interferents was attributed to the air from the
permeation/dilution system. Figure 34 presents the background observed for
a typical Tenax GC blank prior to its use in the sampling analysis studies
at the low levels (electrometer sensitivity was set for low level nanogram
detection). Figure 35 presents the background observed from the portable
permeation system when sampling 30 L. Because of the low levels employed in
this study and the use of an integrating collection device, background was
experienced with the air from the permeation/dilution system. For these
reasons, it was difficult to obtain precise and accurate determinations for
several of the compounds. The recoveries observed after 7 days of storage
for 1,2-dichloropropane, toluene, tetrachloroethylene, chlorobenzene were
relatively better than for m-dichlorobenzene.
Charcoal Cartridges—The results of analysis of solvent desorbed charcoal
tubes (High and Low Level Studies) by GC with electron capture detection are
given in Table 27. The percent recovery for 1,2-dichloropropane and
bis(2-chloroethyl)ether was considerably higher than for tetrachloroethylene
and 1,1,2,2-tetrachloroethane. A decrease in recovery with storage time was
observed. The lower recovery for the latter two compounds was probably due
to the poor desorbing qualitities of the solvent mixture (carbon disulfide/-
methanol:.30/70) . However, the use of a higher concentration of carbon
disulfide which has been shown to be effective for desorbing substances from
charcoal cannot be used with GC/ECD since the ECD exhibits a large response
to this solvent. The percentage of carbon disulfide in methanol was selected
to circumvent this problem.
A chromatogram for a sample stored three days prior to analysis is
shown in Figure 36.
In the low level study no peaks were detected from charcoal traps
(Table 27). The high background produced by the carbon disulfide/methanol
obscured even the qualitative detection of tetrachloroethylene which was
present at 33 pg/mL. An attempt was made to decrease the carbon disulfide
concentration; however, the desorptive properties were then decreased to the
extent that the problem of detection was aggravated.
Ill
-------
3
I
S
cc
Figure 34. Background profile of Tenax GC cartridge used in low level study.
-------
o
Q.
tn
0>
OC
Figure 35. Background profile for 30 L of air from permeation/dilutor passed through a
Tenax GC cartridge.
-------
Table 27. PERCENT RECOVERY OF GROUP I COMPOUNDS FROM CHARCOAL CARTRIDGES3
High Level
Compound
Vinyl chloride
Methyl bromide
Furan
Acrylooitrile
Chloroprene
Chloroform
Benzene
1 ,2-Dichloropropane
Toluene0
Tetrachloroethylene
Chlorobenzene
1,1, 2 ,2-Tetrachloroethane
m-Di chlorobenzene
Bis-(2-chloroethyl)ether
ppb
Sampled
48
80
100
93
4.5
15
60
24
24
30
17
41
20
48
Expected
Quantity
(ng/mL)
3,672
5,587
8,340
6,054
487
2,196
5,742 •
3,326
2,714
6,102
2,346
8,438
3,606
8,410
0 da
NDb
ND
ND
ND
ND
ND
ND
90 + 0.3 (0.33)d
ND
34 + 0.8 (2.3)
ND
35 + 1.6 (4.6)
ND
86 + 0.5 (0.6)
Storage Time
3 da
ND
ND
ND
ND
ND
ND
ND
88 + 5.6 (6.4)
ND
35 + 2.0 (5.7)
ND
32 + 1.5 (4.7)
ND
77 + 4.3 (5.6)
7 da
ND
ND
ND
ND
ND
ND
ND
77 + 5.5 (7.1)
ND
35 + 1.5 (4.2)
ND
33 + 1.0 (3.0)
ND
66 + 2.0 (3.0)
ppt
Sampled
17
429
542
858
2
82
382
184
132
161
95
351
122
252
Low Level
Expected
Quantity
(ng/mL)
13
50
46
57
2
12
32
25
15
33
13
73
22
44
Storage Time
0 da
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3 da
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
7 da
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
aGC/ECD analysis.
bND = not detected.
Not sensitive in ECD.
dMean + S.D: (C.V.).
-------
(E
0
u
Figure 36. ECD chroniatogram of T_ carbon tube extract (high level study)
-------
Cryogenic Traps—The percent recovery of Group I compounds from cryogenic
traps is shown in Table 28 (High and Low Level Studies). In general the
recoveries were extremely poor. This may be attributed to a low collection
efficiency of unpacked nickel traps and/or utilizing dry ice as the cryogen
in the high level study.
For the low level study the nickel cryogenic traps were packed with
glass beads to increase the surface area prior to testing the collection
efficiency. Concentrations of test substances down to 10 ppt were examined
and again the results were poor for the very volatile organics and a large
number of interfering substances inhibited successful low level detection of
compounds (Table 28).
Interference Studies
Containers--
Interference studies were performed to determine the effects of CL,
NO., S0? and water vapor upon the recovery of the test compounds from the
various containers. In the study, Group I and II compounds described
earlier and "Summa" polished stainless steel containers, glass bulbs and
Tedlar bags were used.
By applying the appropriate calibration curve, the ppb level of each
compound in each container was determined. The relative percent recovery
was calculated from these measured ppb levels and the expected levels of
each compound.
Anomalies—It has been seen in stability studies that the FID response
to low concentrations of chloroprene and chloroform is small relative to
most of the other compounds. In the interference studies, both the
chloroprene and chloroform peaks were obscured by a large peak of unknown
origin, preventing their quantitation. This interfering peak was present
in the calibration analyses and analyses from all container types, but
was not present in the control samples. This points to a permeation tube
as a likely source for the contaminant. No data was reported for chloro-
prene or chloroform.
In our analytical scheme, furan and acrylonitrile peaks are generally
not resolved and the data is reported as a sum of the two compounds. At
the beginning of this calibration it was noted that the acrylonitrile
116
-------
Table 28. PERCENT RECOVERY OF GROUP I COMPOUNDS FROM CRYOGENIC TRAPS -
STORAGE/STABILITY STUDY
Compound
Vinyl chloride
Methyl bromide
Furan
Acrylonitrile
Chloroprene
Chloroform
Benzene
1 , 2-Dichloropropane
Toluene i
Tetrachloroethylene
Chlorobenzene
1,1,2 , 2-Tetrachloroethane
Bis-(2-chloroethyl)ether
m-Di Chlorobenzene
0 da
NDa
ND
ND ,
68 +
4 +
39 +
7 +
6 +
1.6 +
4 +
5 +
28 +
lll' +
27 +
3 (4)b
3 (75)
8 (21)
5 (79)
5 (76)
1.9 (113)
3.5 (85)
6 (122)
37 (130)
13 (12)
3 (10)
24
14
4
1
8
61
77
81
High
3 da
53
ND
94
19
ND
+ 6
ND
+ 4
+ 1
+ 0.
+ 1.
+ 15
+ 6
Level
7 da
(26)
(29)
(17)
5 (50)
2 (15)
(25)
(21)
+ 13 (16)
-
20
14
3
3
13
59
70
61
ND
ND
ND
ND
ND
± 3
ND
+ 3
+ 0.
+ 0.
+ 11
+ 7
+ 12
+ 3
-
(16)
(25)
7 (27)
02 (0.6)
(83)
(12)
(17)
(5)
0 da
ND
ND
ND
ND,
ND
ND
102 + 5.1 (5)
ND
97.9 + 1.7 (2)
72.5C
Id
I
I
I
Low Level
3 da
ND
ND
ND
ND
ND
ND
BI
ND
I
96.2 + 39 (41)
101.8
I
I
I
7 da
ND
ND
ND
ND
ND
ND
BI
ND
I
I
I
I
I
88.7
ND = not detected.
bMean + S.D. (C.V.)
GSingle value.
I = Interference.
-------
permeation tube had polymerized. For this reason the assumption was made
that the contribution of acrylonitrile was very little in the calibration
and container samples, and calculations for this peak were based on
characteristics of furan only.
Tedlar Bags—The recovery from the Tedlar bags is presented in Table
29. Recovery is observed to be generally higher with low level interfer-
ences than with high level interferences. The interferences generally
decreased the recoveries over those found in the storage study though some
increases were observed. The fluctuations observed may be due to loss
through chemical reaction, solubilization and/or displacement by water vapor
on the container walls, and generation and/or release of other (organic)
interferences.
Thus absolute recovery may have been higher for some compounds in the
test mixtures than in the standard mixtures. The results of 1,2,3-trimethyl-
benzene are in question because the permeation rate doubled between weighings
before and after the study. The recoveries of nitrobenzene are also question-
able due to problems in calibration. Another trend seen was the large
number of unknown peaks present in. bags containing high level interferences
as compared to bags with low level interferences. These interfering peaks
prevented the quantification of bis-(2-chloroethyl)ether and m-dichlorobenzene.
These interferences may have been released by the water vapor in the test
sample.
No particular trends are noted in the precision of the recovery values.
In fact, the precision values are similar for both high and low level inter-
ferences for most compounds.
Glass Bulbs—The recovery from the glass bulbs is presented in Table
30. The majority of the compounds showed higher recovery with the low-
level interferences than with the high-level interferences. Recoveries were
generally less than 100% for both high and low level interferences, though
it is interesting to-note that several compounds show substantially better
recovery than in the storage study. The results of 1,2,3-trimethylbenzene
and nitrobenzene are again questionable.
There appears again to be no particular trends with regard to preci-
sion.
118
-------
Table 29. PERCENT RECOVERY OF TEST COMPOUNDS IN THE PRESENCE OF
POTENTIAL INTERFERENCES FROM TEDLAR BAGS
Compound
ppb Sampled
High Level
Low Level
Methyl chloride
Vinyl chloride
Methyl bromide
Methyl mercaptan
Furan/acrylonitrile
Propylene oxide
Vinylidene chloride
Allyl chloride
Chloroprene
Chloroform
1,1,1-Trichloroethane
Benzene
1,2-Dichloropropane
Toluene
a-Epichlorohydrin
Tetrachloroethylene
Chlorobenzene
Ethylbenzene
o-Xylene
1,1,2,2-Tetrachloroethane
m-Dichlorobenzene
55.8
27.7
60.1
59.0
39.6
27.3
20.9
87.5
2.1
8.1
12.4
32.4
17.5
13.6
54.9
12.4
9.2
7.4
23.5
18.5
11.0
65.7 + 5.9 (8.9)
72.5 + 4.3 (5.9)
78.9 + 2.7 (3.5)
Not detected
38.6 + 3.3 (8.6)
Not detected
37.5 + 6.5 (17.4)
57.7 + 1.5 (2.6)
85,7 + 3.7 (4.3)
67.8 + 4.0 (5.9)
109.3 + 4.3 (4.0)
73.6 + 4.6 (6.2)
Not detected
109.7 + 7.4 (6.7)
90.2 + 2.9 (3.2)
72.0 + 4.5 (6.3)
58.2 + 1.2 (2.1)
71.0 + 3.6 (5.1)
b
81.3 + 2.3 (2.9)
92.6 + 4.2 (4.6)
85.4 + 3.8 (4.4)
Not detected
76.0 + 12.7 (16.7)
Not detected
71.2 + 4.9 (6.9)
85.8 + 2.2 (2.6)
81.5 + 13.6 (16.7)
114.4 + 15.3 (13.3)
113,7 + 9.6 (8.4)
87.5 + 4.5 (5.2)
Not detected
85.7 + 3.5 (4.1)
88.3 + 4.1 (4.6)
86.5 + 1.4 (1.6)
68.5 + 2.2 (3.2)
72.5 + 2.9 (4.0)
62.5C
(continued)
-------
Table 29 (cont'd.)
Compound
n-Decane
1,2, 3-Trimethylbenzene
Bis-(2-chloroethyl)ether
o-Cresol
Nitrobenzene
Benzyl chloride
ppb Sampled
6.5
66.0
2.4
20.9
22.2
28.6
High Level
63. 5C
42.3 + 2.4 (5.8)
___b
Not detected
76.4 + 0.3 (0.4)
69. 6d
Low Level
87. 5C
428.6 + 22.7 (5
62. 5C
Not detected
84.4 + 1.8 (2.
85. 7d
.3)
1)
r-o
o
Based on furan only, acrylonitrile tube polymerized.
Not quantified due to interfering peaks.
"Single observation.
Only one measurement achieved due to interference.
-------
Table 30. PERCENT RECOVERY OF TEST COMPOUNDS IN THE PRESENCE OF~
POTENTIAL INTERFERENCES FROM GLASS BULBS
Compound
ppb Sampled
High Level
Low Level
Methyl chloride
Vinyl chloride
Methyl bromide
Methyl mercaptan
Furan/acrylonitrile
Propylene oxide
Vinylidene chloride
Allyl chloride
Chloroprene
Chloroform
1,1,1-Trichloroethane
Benzene
1,2-Dichloropropane
Toluene
«-Epichlorohydrin
Tetrachloroethylene
Chlorobenzene
Ethylbenzene
o-Xylene
1,1,2,2-Tetrachloroethane
m-Dichlorobenzene
55.8
27.7
60". 1
59.0
39.6
27.3
20.9
87.5
2.1
8.1
12.4
32.4
17.5
13.6
54.9
12.4
9.2
7.4
23.5
18.5
11.0
79.0 + 1.2 (1.5)
94.2 + 4.1 (4.3)
82.3 + 1,2 (1.4)
Not detected
49.3 + 2.6 (5.2)
Not detected
70.5 + 24.3 (34.5)
93.2 + 3.1 (3.3)
101.0 + 13.7 (13.5)
60.0 + 3.4 (5.6)
107.2 + 2.0 (1.9)
80.6 + 0.9 (1.1)
Not detected
84.7 + 14.6 (17.2)
89.9 + 3.5 (3.9)
78.1 + 4.2 (5.4)
62.6 + 6.9 (11.0)
83.8C
92.4 + 5.5 (5.9) .
75.4 + 8.9 (11.8)
99.5 + 0.5 (0.5)
96.9 + 2.2 (2.3)
Not detected
78.6 + 2.9 (3.7)
Not detected
66.1 + 8.0 (12.1)
85.3 + 3.0 (3.5)
85.4 + 4.9 (5.8)
90.3 + 0.5 (0.5)
96.0 + 1.6 (1.6)
89.5 + 2.3 (2.5)
Not detected
93.8 + 6.7 (7.1)
90.2 + 5.4 (6.0)
88.2 + 2.9 (3.3)
72.3 + 1.4 (2.0)
73.5 + 1.6 (2.1)
81.6 + 9.1 (11.1)
(continued)
-------
Table 30 (cont'd.)
Compound
n-Decane
1,2, 3-Trimethylbenzene
Bis-(2-chloroethyl)ether
o-Cresol
Nitrobenzene
Benzyl chloride
ppb Sampled
7.5
66.0
2.4
20.9
22.2
28.6
High Level
71.5 + 6.6 (9.2)
44.7 + 8.7 (19.4)
83.3 + 8.4 (10.0)
Not detected
75.9 + 0.4 (0.6)
79.7 + 7.8 (9.8)
Low Level
82.9 + 2.1 (2.6)
539.3 + 18.9 (3.4)
60.4 + 4.2 (7.0)
Not detected
80.7 + 4.2 (5.2)
87.5 + 8.1 (9.3)
Based on furan only, acrylonitrile tube polymerized.
Not quantified due to interfering peaks.
^Single observation.
-------
Summa Polished Steel Containers—The recovery from "Summa" polished
steel containers is presented in Table 31. The level of interference
appeared to have a significant effect upon the recovery of the majority of
the compounds from "Summa" cans. The interferences generally decreased the
recoveries over those found in the storage study, though, in fact, several
recoveries were elevated by the presence of the interferences. Again fluctua-
tions may be due to loss through chemical reaction, displacement by water
vapor, and generation and/or release of other (organic) interferences.
As with the other containers, no general trends in precision are to be
seen. It is noted however that very low recoveries and recoveries greater
than 100% show exceptionally poor precision.
Traps--
The effects of inorganic pollutants on the recovery of test compounds
®
using Tenax GC, charcoal and cryogenic traps were investigated. Recovery of
all compounds were evaluated only at the ppb .level (Table 32) and were
examined under several different experimental conditions (see Experimental
Methods/Interference Study) which included sampling of test compounds (1)
in the absence of inorganic pollutants, (2) in the presence of inorganic
pollutants (with and without a glass fiber filter impregnated with ca. 5 mg
®
of sodium thiosulfate prior to the Tenax GC cartridge), and (3) at hi'gh and
low levels of inorganic pollutants.
Tenax GC Cartridges—Table 32 presents the absolute recovery of test
compounds for the control sample prior to the addition of inorganic pollutants,
The levels of test substances ranged from 1 to 48 ppb. Because the break-
through volumes were exceeded for"the first six compounds listed and lower
levels were employed than in the high level storage-stability studies several
ii
compounds were at trace levels or near background levels. The hydrocarbon
background was traced to the NO supply.
Table 32 also rlist the absolute recovery of the test compounds in the
presence of "high levels" of inorganic pollutants with and without a glass
®
fiber filter preceding the Tenax GC cartridge. These data indicate that the
use of a glass fiber filter impregnated with sodium thiosulfate, a mild
reducing agent, decreases the effect of inorganic gases on recovery of test
®
compounds from Tenax GC cartridges. Conversely, the percent recovery
123
-------
Table 31. PERCENT RECOVERY OF TEST COMPOUNDS IN THE PRESENCE OF
POTENTIAL INTERFERENCES FROM "SUMMA" POLISHED SS CANS
Compound
Methyl chloride
Vinyl chloride
Methyl bromide
Methyl mercaptan
Furan/acrylonitrile
Propylene oxide
Vinylidene chloride
Allyl chloride
Chloroprene
Chloroform
1 ,1,1-Trichlo roe thane
Benzene
1 , 2-Dichloropropane
Toluene
-------
Table 31 (cont'd.)
Compound
n-Decane
1 ,2 ,3-Trimethylbenzene
Bis-(2-chloroethyl)ether
o-Cresol
Nitrobenzene
Benzyl chloride
ppb Sampled
7.5
66.0
2.4
20.9
22.2
28.6
High Level
62.0 + 3.6 (5.8)
23.7 + 3.4 (14.3)
83.4 + 5.9 (7.0)
Not detected
81.6 + 0.9 (1.1)
76.9 + 4.2 (5.4)
Low Level
94.2 + 13.1 (13.9)
487.7 + 18.9 (3.9)
60.4 ± 2.4 (4.0)
Not detected
88.9 + 3.3 (3.7)
89.6 + 4.8 (5.4)
Based on furan only, acrylonitrile tube polymerized.
Not quantified due to interfering peaks.
-------
ro
ON
Table 32. ABSOLUTE RECOVERY OF GROUP I COMPOUNDS FROM TEN-
HIGH LEVEL POTENTIAL INTERFERENCES
GC CARTRIDGES -
Compound
Vinyl chloride
Methyl bromide
Furao
Acrylonitrile
Chloroprene
Chlorofom
Benzene
1 ,2-Dicbloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2,2-Tetracbloroethane
Bia-(2-chloroethyl)ether
m-Di ch 1 o robenzeae
3Synthetic air/vapor
Synthetic air/vapor
ppb Sampled
17
43
41
48
1
8
32
19
14
16
10
24
25
13
mixture sampled,
mixture sampled,
Expected Quant
(ng/cartridge
165
42
339
520
60
720
1,545
2,580
1,530
3,440
1,350
4,864
4,350
2,310
no 03, N02
potential
(-) Inorganics
i t u
ity
) No GFF
T"
T
'
-
'
71 + 25 (35)
76+5 (7)
63+4 (6)
78 + 5 (6)
78+3 (4)
77+3 (4)
118 + 33 (28)
69 + 13 (19)
64+6 (9)
, S0? or humidity was
interferences present
(+) Inorganics
No GFF
T
T
ND
-
ND
54+9 (17)
40+4 (10)
76+5 (7)
78+3 (4)
79+4 (5)
90+5 (6)
131 + 30 (23)
71 + 16 (22)
77 + 11 (14)
present.
CFFC
T
T
ND
-
ND
BI
77+8 (10)
92 + 17 (18)
82 + 1 (1)
75 + 15 (20)
84 + 5e
118 + 11 (9)
69+7 (10)
67 + 0.2 (0.3)
CGFF = glass fiber filter impregnated with sodium thiosulfate (CJK 5 mg) was used prior to the
cartridge.
dT = trace, - = weak signal not resolved from background, ND = not detected, BI = background
interference.
Duplicate analysis only.
-------
increased for benzene and 1,2-dichloropropane. A possible explanation is
that ozone is quenched by the reducing agent prior to reaching the adsorbent
and thus the adsorbed test compounds are not destroyed. In the case of
furan and chloroprene, these compounds may have reacted (depleted) with the
high levels of ozone while in transit through the permeation/dilution system
prior to reaching the sampling device.
Table 33 gives the relative percent recoveries for all compounds
tested. Relative percent recoveries were calculated as a ratio of absolute
recovery observed in the presence of inorganic pollutants to the absolute
recovery observed in the absence of inorganic pollutants times 100 percent.
These data suggest a trend since inorganic pollutants lower the recovery of
furan, chloroprene, chloroform, and benzene (see Tables 32 and 33/no GFF).
Table 33 also gives the relative percent recoveries, for sampling and
analysis of test compounds in the presence of "low levels" of inorganic
pollutants. These data indicate that lower levels of inorganic pollutants
did not completely .react with furan, chloroprene, and acrylonitrile during
their transit time in the permeation/dilution system. Since sufficient test
®
substance reached the Tenax GC cartridge differences between sampling with
and without a GFF were demonstratable for these compounds. Except for ct-
epichlorohydrin the recoveries for other test substances were similarly
unaffected by "lower levels" of inorganic pollutants ^.e. their recoveries
were already as high as the "control" experiment.
Table 34 gives the relative percent recovery for Group II compounds
observed without correction for breakthrough volume. Thus, comparison of
Tables 33 and 34 reveals the differences when taking the breakthrough volume
into account relative to the sampling volume.
Figures 37 through 40 depict a few example chromatograms for various
experimental conditions.
Charcoal Cartridges—Tables 35 and 36 present the percent recoveries of
test compounds from charcoal cartridges in the presence of high and low
levels of inorganic gases. Because of a contaminant from charcoal (see
Section 7 for identification), it was not possible to quantify bis(2-chloro-
ethyl)ether. Thus, only results for 1,1,2,2-tetrachloroethane and tetrachloro-
ethylene were obtained. The data (Tables 35 and 36) includes absolute
127
-------
Table 33. RELATIVE PERCENT RECOVERY OF TEST COMPOUNDS FROM TENAX GC CARTRIDGES -
INTERFERENCE STUDY WITH CORRECTION FOR BREAKTHROUGH VOLUME
K>
00
Compound
Methyl chloride
Vinyl chloride
Methyl bromide
Methyl oercaptan
Furan
Propylene oxide
Acrylonitrile
Vinylidene chloride
Ally! chloride
Chloroprene
Chlorofora
1 , 1 , 1-Trichloroetbane
Benzene
1,2-Dicbloropropane
Toluene
o-Epichlorohydrin
Tetrachloroethylene
Chlorobenzene
Ethylbenzene
o-Xylene
1,1,2, 2-Tetrachloroethane
n-Dichlorobenzene
B.P. Breakthrough Volume ppb
(°C) (C) Sampled
-24.2
-13
3.4
6.2
31.4
34.3
37
45
59.4
61.7
74.1
77
96.4
110.6
116.5
121
132
136.2
139.1
146.2
173
3
1
1
-
3
4
5
1
6
15
18
9
38
81
173
54
380
344
344
-
173
948
55.8
17
43
59
41
27.3
48'
20.9
87.5
1
8
12.4
32
19
14
54.9
16
10
7.4
23.5
24
13
High
No GFF
3.5 + 2.1 (60)
TC
T
ND
NC
T
NC
50 + 21 (42)
4.9 + 1.2 (24)
ND
76 + 36 (47)
6.7 * 1.2 (18)
53+9 (17)
120 + 11 (9)
100 + 7 (7)
39 * 6.1 (15)
101 + 7 (7)
116 + 8 (7)
40 + 1 (2.5)
51 + 2.2 (4.4)
111 + 40 (36)
120 + 17 (14)
Level3
OFF"
10 + 1.7 (17)
T
T
ND
ND
19 + 8.9 (5)
NC
239 + 131 (55)
18 + 3.4 (18)
ND
NC
101 + 10 (10)
101 + 8 (8)
146 + 19 (12)
105 + 6 (6)
104 + 14 (13)
96 * 11 (11)
109 * 9 (8)
63 + 11 (17)
85 + 15 (18)
100 + 30 (30)
104 + 6 (5.7)
Low
No OFF
2.9 + 2.6 (91)
T
T
ND
42+7 (17)
18 + 11 (65)
I
39 + 27 (70)
7.4 + 4.0 (54)
48 + 13 (27)
100 + 44 (44)
8.0 + 3.6 (45)
47 + 10 (11)
82+7 (9)
97+3 (3)
34 + 15 (44)
94+6 (6)
96+2 (2)
43 + 18 (42)
55 + 23 (42)
80 + 22 (28)
83+6 (7)
Level
GFF
4.6 + 1.3 (28)
T
T
ND
111 + 21 (19)
37 + 4.7 (13)
I
127 + 34 (27)
30 + 2.1 (7.2)
112 + 39 (35)
I
77 + 10 (13)
100 + 10 (10)
129 + 12 (9)
103 + 5 (5)
92 + 12 (13)
111 + 7 (6)
106 + 2 (2)
75 + 7.5 (10)
96 + 11 (11)
126 + 50 (40)
91 + 11 (12)
(continued)
-------
Table 33 (cont'd.)
Compound
n-Decane
1 , 2 , 3-Tr imethylbenzene
Bis-(2-chloroethyl)ether
o-Cresol
Nitrobenzene
Benzyl chloride
B.P. Breakthrough Volume pp'b
(°C) (£) Sampled
174.1
176.1
178 234
190.9
210.8
215 830
7.5
66
25
20.9
22.2
28.6
No
24 +
13 +
103 +
I
79 +
24 +
GFF
4.0
0.3
31
0.7
2.3
High
Level
Low Level
GFF
(8.4)
.(2.2)
(30)
(0.9)
(9.7)
38
24
100
139
58
+
+
•f
I
+
+
6.4 (17)
4.0 (16)
20 (20)
19 (14)
7.9 (14) '
No GFF
I
I
98 + 21 (21)
I
112 + 47 (42)
I
GFF
I
I
111 + 38
I
93 + 35
I
(34)
(38)
10
VO
See Table 10 for levels of inorganic gases employed. .
DGFF = glass fiber filter impregnated with sodium thiosulfate.
'T = trace, NO = not detected, NC = not calculated, I = interference, values are mean +_
S.D. (C.V.) for triplicate samples.
Permeation tube was highly suspect.
-------
OJ
o
Table 34. RELATIVE PERCENT RECOVERY OF GROUP II COMPOUNDS FROM TENAX GC CARTRIDGES -
INTERFERENCE STUDY, UNCORRECTED FOR BREAKTHROUGH VOLUME
Compound
Hethyl mercaptan
Vinylidene chloride
Hethyl chloride
Propyleae oxide
Allyl chloride
Ijljl-Trichloroe thane
ot-Epichlorohydrio
n-Decane
o-Cresol
o-Xylene
Ethylbeazeoe
1,2,3-TriBethylbenzene
Benzylchloride
Nitrobenzene
Breakthrough Vol.
(L) g 90°F
-
1.0
3
4
6
9
54
>30
>30
>30
344
>30
830
>30
ppb
Sampled
59.0
20.9
55.8
27.3
87.5
12.4
54.9
7.5
20.9
23.5
7.4
66.0
28.6
22.2
High Level
No. GFF
2.5 + 1.0 (42) .
1.7 + 0.7 (42)
0.4 + 0.2 (50)
Ta
1.0 + 0.2 (23)
5.4 + 1.0 J18J
39 + 6.1 (15)
24 + 2.0 (8)
Ib
51 t 2.2 (4)
40 + 1 (2)
13 + 0.3 (2)
24 + 2.3 (10)
79 + 0.7 (0.9)
GFF
8.3 + 8.0 (98)
9.6 + 5.3 (55)
1.4 + 0.2 (17)
3.2 + 1.5 (47)
5.5 + 1.0 (19)
40+4.1 {10]
104 + 14 (13)
38 + 6.4 (17)
I
85 + 15 (18)
63 + 11 (18)
24 ± 4.0 (16)
58 + 8.0 (14)
139 + 19 (14)
Low Level
No GFF
4.6 + 2.3 (51)
1.3 + 0.9 (70)
0.3 + 0.3 (100)
2.3 + 1.5 (65)
1.5 + 0.8 (54)
6.5 » 2.9 {45}
34 + 15 (44)
I
I
55 + 23 (42)
43 + 18 (42)
I
I
112 + 47 (43)
OFF
9.6 + 2.5 (25)
5.1 + 1.4 (27)
0.6 + 0.2 (30)
6.2 + 0.8 (13)
8.9 + 0.6 (7)
31 + 4.2 {14}
92 + 12 (13)
I
I
96 + 11 (11)
75 + 7.5 (10)
I
I
93 + 35 (38)
T = trace.
I = interference.
-------
o
a
2
cc
\Y
Figure 37. Chromatogram for background observed with Tenax used to sample 30 L of air
containing low levels of inorganic gases (no GFF).
-------
I
o
Q.
IA
0)
DC
Figure 38. Chroinatogram of background observed with Tenax used to sample 30 L air containing
low levels of inorganic gases (with GFF).
-------
o
a
cc
a
II
II
Figure 39. Chromatogram for sample taken with Tenax with test compounds and low levels of
inorganic gases present (no GFF).
-------
oc
q
u,'
0>
•a
Figure 40. Chromatogram for sample taken with Tenax with test compounds and low levels of
inorganic gases present (with GFF).
-------
Table 35. PERCENT RECOVERIES OF GROUP I COMPOUNDS FROM CHARCOAL CARTRIDGES IN THE
PRESENCE OF HIGH LEVELS OF INORGANIC SUBSTANCES3
w
01
Compound
m-Dichlorobenzene
Bis- (2-Chloroethyl) ether
1,1,2 , 2-Tetrachloroethane
Absolute/-!. S.b
_d
BIe
ND
Percent Recovery + S.D. (C.V.)
Absolute/+ I.S.
_
BI
43 + 11 (26)
RelativeC/+ I.S.
_
BI
ND
Chlorobenzene
Tetrachloroethylene
Toluene
1,2-Dichloropropane
Benzene
Chloroform
Chloroprene
Acrylonitrile
Furan
Methyl bromide
Vinyl chloride
54 + 12 (22)
80 + 42 (52)
148 + 84 (57)
See Table 10 for concentrations of inorganic substances.
IS = inorganic substances absent (-) or present (+).
"Recoveries are relative to "control" which was collection in the absence of inorganic substances.
Compounds not detected by this analysis procedure.
:a
'ND = not detected, BI = background interference.
-------
Table 36. PERCENT RECOVERIES OF GROUP I COMPOUNDS FROM CHARCOAL CARTRIDGES IN THE
PRESENCE OF LOW LEVELS OF INORGANIC SUBSTANCES3
Compound
m-Dichlorobenzene
Bis- (2-Chloroethyl)ether
1 , 1 , 2 , 2-Tetrachloroethane
Absolute/-!. S.b
_d
BIe
ND
Percent Recovery + S.D. (C.V.)
Absolute/+ I.S.
_
BI
Te
RelativeC/+ I.S.
_
BI
ND
Chlorobenzene
Tetrachloroethylene
Toluene
1,2-Dichloropropane
Benzene
Chloroform
Chloroprene
Acrylonitrile
Furan
Methyl bromide
Vinyl chloride
117 + 7 (6)
60 + 31 (52)
51 + 27 (52)
See Table 10 for concentrations of inorganic substances.
IS = inorganic substances absent (-) or present (+).
Recoveries are relative to "control" which was collection in the absence of inorganic substances.
Compounds not detected by this analysis procedure.
£J
ND = not detected, BI = background interference.
-------
recoveries in the absence and presence of inorganic gases and the relative
recoveries between the two. This phenomenon was also observed but to a
smaller degree with low levels of inorganic gases. A possible explanation
may be that charcoal tubes prior to sampling are highly activated and while
sampling is conducted in the presence of humidity, the charcoal sorbent
becomes deactivated and thus the adsorbed analytes are more easily desorbed
with organic solvent than when the analytes are adsorbed to an activated
charcoal surface.
The other analytes in Tables 35 and 36 are listed only for reference
purpose. Electron capture detection does not detect these compounds because
of their poor electron affinity. Figures 41 through 46 present typical
chromatograms obtained during the interference studies.
The results utilizing charcoal in both the storage-stability study and
in the pollutant interference study were somewhat disappointing since the
number of compounds which could actually be detected (FID and ECD) using
this sorbent were rather few.
Cryogenic Traps—Since sampling and analysis with Ni cryogenic traps
packed with glass beads using dry ice was not fruitful in the storage-
stability studies, liquid oxygen was used as the coolant in these experiments.
Tables 37 and 38 present the absolute percent recovery of test compounds
i.
from cryogenic traps in the presence of high and low levels of inorganic
gases. Once again inconsistent background of the cryogenic trap yielded
large coefficients of variation. Given in each Table are the results for
the presence and absence of the standard inorganic pollutants. Differences
in recovery were observed between the two experimental conditions. The
recovery of vinyl chloride, methyl bromide, furan, chloroprene, chloroform,
and benzene all decreased when inorganic gases were present during sampling.
The coefficient of variations for triplicate analysis were large for other
compounds. The absolute recoveries appeared to be less than those observed
®
with Tenax GC traps.
Another problem arises with the cryogenic traps when using liquid
oxygen. Excessive amounts of water were trapped and subsequently transferred
to the Tenax cartridge during the purging step. In order to analyze by
TD/HRGC, the large quantities of water were removed by placing calcium
137
-------
o
I
cc
a
o
teAteJY'
Figure 41. Chromatogram of background for 30 L air sample taken with charcoal trap in the
presence of high levels of inorganics and no test model compounds.
-------
o
a
IA
a>
CC
a
u
UJ
Figure 42. Chromatogram for 30 L air sample in the absence of inorganics (high level study)
and with test model compounds.
-------
c
o
Q.
S3
cc
o
u
01
C4H10S2
tetrachloro-
ethylene
2-ethylhexanol
Figure 43. Chromatogram of 30 L air sample with charcoal trap in the presence of high levels
of inorganics and test model compounds.
-------
a
o
Q.
M
«
cc
o
(J
UJ
Figure 44. Chromatogram of background for 30 L air sample taken with charcoal trap in the
presence of low levels of inorganics and no test compounds.
-------
.e-
10
9j
o
a
IXI
0)
cc
a
u
w
Figure 45. Chromatogram for 30 L air sample with charcoal trap in the absence of inorganics
(low level study) and test compounds present.
-------
c
o
Q.
18
DC
Q
O
01
Figure 46. Chromatogram of 30 L air sample using charcoal trap and low levels of inorganics
and test compounds present.
-------
Table 37. ABSOLUTE PERCENT RECOVERY OF GROUP I COMPOUNDS FROM CRYOGENIC TRAPS IN THE
PRESENCE OF HIGH LEVELS OF INORGANIC SUBSTANCES3
Compound
m-Dichlorobenzene
Bis -(2-chloroethyl) ether
1,1,2,2-Tetrachloroethane
Chlorobenzene
Tetrachloroethylehe
Toluene
1 , 2-Dichloropropane
Benzene
Chloroform
Chloroprene
Acrylonitrile
Fur an
Methyl bromide
Vinyl chloride
(-) Standard Pollutants
15+3 (20)
20+5 (25)
55 + 11 (20)
35+8 (23)
30 + 11 (37)
42 + 13 (31)
22+6 (27)
41 + 14 (34)
93 + 28 (30)
18+7 (39)
BI
T
T
T
(+) Standard Pollutants
28+1 (4)
28+5 (18)
65+9 (14)
32+7 (22)
28 + 10 (36)
26+9 (35)
25 + 21 (84)
21 + 19 (90)
20+7 (35)
ND
BI
ND '
T
T
See Table 10 for concentrations of inorganic substances employed. Recoveries are
calculated amounts in the synthetic air/vapor stream sampled. Mean + S.D. (C.V.)
plicate samples.
BI = background interference, ND = not detected, T = trace.
based upon the
are for tri-
-------
Table 38. ABSOLUTE PERCENT RECOVERY OF GROUP I COMPOUNDS USING CRYOGENIC TRAPS -
LOW LEVEL POTENTIAL INTERFERENCES
•F-
Cn
Compound
Vinyl chloride
Methyl bromide
Furan
Acrylonitrile
Chloroprene
Chloroform
Benzene
1,2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2 , 2-Tetrachloroe thane
Bis-(2-chloroethyl)ether
ni-Di chlorobenzene
ppb Sampled
17
43
41
48
1
8
32
19
14
16
10
24
25
13
(-) Inorganics
23+8 (35)
53 + 25 (47)
50+1 (2)
BI
32+7 (22)
126 + 26 (21)
50 + 22 (44)
72 + 5 (7)
54+3 (6)
61+5 (8)
66+6 (9)
107 + 47 (44)
33+9 (27)
58 + 10 (17)
(+) Inorganics
22 + 28 (127)
27 + 28 (104)
NDb
BI
22 + 16 (73)
66 + 33 (50)
26 + 18 (69)
30 + 16 (53)
31 + 16 (52)
31 + 18 (58)
34 + 11 (32)
42 + 18 (43)
18 + 6 (33)
30+7 (23)
Synthetic air/vapor mixture sampled without and with inorganic gases present.
ND = not detected, BI = background interference, values are mean + S.D. (C.V.) of triplicate
samples.
-------
sulfate in the bottom of the culture tube to adsorb moisture from the glass
®
wool used to secure the Tenax GC in the cartridge. This procedure solved
the problem of excess water. Other studies have shown that this drying step
is quantitative (20).
Table 39 presents the relative percent recovery of test compounds from
cryogenic traps in the presence of high levels of inorganic substances.
These data were calculated as recoveries relative to the control which was
the collection of test compounds in the absence of inorganic gases. Even
though a large coefficient of variation was observed, considerable differences
between the control (no standard pollutant) and experimental sample are
evident. Inconsistent trapping of organic test vapors and background from
the cryogenic traps contributed to the observed large coefficient of varia-
tions. Figures 47 through 49 present typical gas chromatograms obtained for
the analysis of cryogenic traps (attenuation is constant).
Comparison of the absolute recoveries (Table 37) obtained in the absence
of inorganic gases with the "high level" storage-stability study reveals
that use of liquid CL as the coolant was more efficient than with powdered
dry ice. However, the recoveries are still unacceptably low.
Figures 50 through 52 present examples of chromatograms for the "low
level" interference study using cryogenic traps.
Because of the poor recoveries of Group I compounds using Ni traps
packed with glass beads and liquid oxygen as the coolant, no recovery experi-
ments with Group II compounds were conducted.
DISCUSSION
Containers
Bags--
Bags have the advantage of allowing collection of 10-100 liters of
sample but they are easily punctured and clear bags must be protected from
light after sample collection. Thorough cleaning can be complicated as the
bags cannot be heated. Cleaning with ozone and/or ultraviolet light appears
to reduce the levels of high boiling contaminants; actually these cleaning
procedures may be producing compounds which are much less volatile or adsorb
more strongly than the parent compounds.
146
-------
Table 39. RELATIVE PERCENT RECOVERY OF GROUP I COMPOUNDS USING
CRYOGENIC TRAPS IN THE INTERFERENCE STUDY3
Compound
Vinyl chloride
Methyl bromide
Furan
Acrylonitrile
Chloroprene
Chloroform
Benzene
1 , 2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2,2-Tetrachloroethane
Bis- (2-chloroethyl) ether
m-Dichlorobenzene
Low
Mean +
96 +
51 ±
NDb
I
69 +
52 +
52 +
42 +
57 +
51 +
51 +
39 +
54 +
52 +
Level
S.D. (C.V.)
125 (130)
56 (102)
48 (66)
25 (48)
41 (80)
22 (52)
29 (51)
28 (55)
15 (29)
24 (61)
30 (56)
13 (25)
High Level
Mean + S.D. (C.V.)
T
T
ND
I
ND
21 + 19 (93)
51 + 53 (105)
114 + 67
62 + 37 (60)
93 + 41 (45)
92 + 32 (35)
119 + 28 (24)
140 + 25 (18)
201 + 42 (21)
Relative recoveries were calculated between absolute recovery with no
inorganic gases and inorganic gases. See Table 10 for levels of
inorganic substances employed.
ND = not detected, BI = background interference, T = trace.
147
-------
00
c
o
Q.
Figure 47. Chromatograms of background for cryogenic trap for 30 L of air sample containing
high levels of inorganics and no test compounds.
-------
C
o
Q.
1/1
01
DC
Figure 48. Chromatograni of sample collected with cryogenic trap without inorganics (high level
study) and test compounds present.
-------
c
o
a
3
cc
en
O
Figure 49. Chromatogram of sample collected with cryogenic trap with high levels of inorganic
and test compounds present.
-------
unknown
Jl
Figure 50. Chromatogram of background for cryogenic trap with low levels of inorganic gases and
no test compounds.
-------
I
o
•5 =
Figure 51. Chromatogram of sample collected with cryogenic trap with test compounds only
(low level interference study).
-------
Figure 52. Chromatogram of sample collected with cryogenic trap with low level inorganic gases
and test compounds present.
-------
Recoveries from Teflon and Tedlar bags are generally similar shortly
after sampling. However, with time, the decrease in recovery is generally
more rapid with Teflon than with Tedlar . Both types of bags show both
loss of compounds and influx of contaminants by permeation through the bag
walls. It would seem that the aluminum coated bags should prevent this
permeation; we could not test this possibility as we were unable to satisfac-
torily clean the aluminum coated bags. It appears than that bags not having
sealed surfaces or not stored in clean environments should be trusted no
more than 4-24 hours (depending on the levels of contaminants in the bag
storage area) after sample collection.
The interferences decreased the recovery of most test compounds from
®
the Tedlar bags though a few compounds such as tetrachloroethylene and
chlorobenzene showed an increase in recovery over that obtained in the
absence of interferences. This increase may have been due to the water
vapor blocking adsorption of and/or displacing already adsorbed organic
molecules. One adverse effect caused by the interferences was the release
®
of unknown contaminants from the wall of the Tedlar bags.
Glass Bulbs--
The amount of sample which can be collected in a glass bulb is limited;
usually 1-2 liters are available for analysis after collection. Glass bulbs
are easily broken, especially the filling tube/valve part of the bulb. The
bulbs should also be protected from light after sample collection. Bulbs
can be cleaned by evacuation; heating while evacuating improves the efficiency
of the cleaning process though care should be taken not to heat the valve.
Cleaning with a solution is very difficult and time consuming since the
valve must be disassembled, this latter operation leads to a high incidence
of breakage.
The recovery of the test compounds from the glass bulbs decreased
rapidly with increased boiling point and was above 90% for only a few com-
pounds. The low level interferences increased the recoveries of many of the
compounds above those obtained in the absence of interferences. The high
levels of interferences generally decreased the recoveries over those obtained
with low level interferences. The relatively high recoveries obtained in
154
-------
the presence of the interferences may well be due to the rapid occupation of
sorption sites by water molecules, which prevents loss to the glass surface.
Steel Containers—
As with glass containers, steel containers allow recovery of only
limited sample volumes, typically 4-6 liters. They are, however, extremely
rugged and can be cleaned thoroughly by heating while evacuating.
The "Summa" polished containers generally show higher recovery for high
boiling point compounds than the RTI-electropolished containers. Also the
"Summa" electropolished containers show a better maintenance of recovery
with time than the RTIpelectropolished containers. As with bags, some test
compounds show an actual increase in recovery from Day 3 to Day 7. This may
well be due to the relatively slow displacement of organic molecules by
water molecules in the test mixtures from the steel surface; this process
would be expected to be slow in the test mixture as the water level is low
r
in this case.
The low level interferences do not decrease recoveries but usually
increase them. High level interferences increase recoveries for some com-
pounds and decrease recoveries for others. These increases may be due to
further displacement by water of the test compounds and/or release of contami-
nants not released during cleaning.
Detection Limits-- i
A limitation of cannisters (and bulbs) is their small volume, generally
2-9 liters. Even if a cannister is pressurized to two atmospheres, only two
1 liter samples can be acquired practically using the trapping and measurement
system described previously. This is, of course, much less than the 10
liters than can be taken from a bag or the 50 liters that can be passed
through a trap, Tenax, for example. The small volume that can be taken from
a rigid cannister will lead to higher detection limits when compared to
other bags or traps. This concern was tested using the cryogenic trapping
and measurement system described and FID detection. Calibration mixtures at
the 10 ppt and 1 ppb level were loaded into 2 L cannisters and then measured.
Estimates of detection limits were calculated based on peak heights (2 chart
divisions) and computer-calculated area counts (50 area counts) - two chart
divisions and 50 area counts were assumed to be distinguishable from
155
-------
background signals. The detection limits are shown in Tables 40 and 41.
The only compounds detected at the lower level were furan and toluene. This
is to be expected; calculations from the higher level indicate that furan
and toluene are the only compounds present in the low level calibration at
greater than or near their minimum detection level.
Precision--
As stated previously, very few if any trends in precision are observed.
The only significant trend seem to be a decrease in precision with storage
time for some compounds. The precision decrease is more pronounced with the
bags and steel containers relative to the glass bulbs. This may be related
to the apparent rapid interaction between water and glass and the relatively
slow interaction between water and plastic or steel.
Traps
Charcoal Cartridges—
The charcoal cartridges evaluated in this program for applicability to
the collection of environmental levels of vapor-phase organics are widely
used for industrial workplace monitoring. They have been endorsed by NIOSH
for this purpose; however, their sampling requirements are significantly
different since ppm levels are sought in the absence of ozone. As such 200
mg charcoal cartridges have not been adequately studied for environmental
applications.
The limits of applicability of charcoal cartridges are revealed in the
overall poor recovery and detection of organics. For the analysis of test
compounds used in this study, GC/FID and GC/ECD were employed. An examination
of the limits of detection for GC/FID indicated that none of the test com-
pounds collected in the ppb range and a 30 L sample volume would be detected
(Table 42). In fact, the chemicals were barely detectable in the calibration
standard prepared at a concentration expected in the "high" level storage-
stability study. When a selected number of T_ cartridges were solvent
desorbed and analyzed no chromatographic peaks were detected.
Sampling volumes larger than 30 L were not investigated. Presumably,
volumes of 300 L or greater would be necessary to detect chemicals in the
low ppb level. However, factors such as the breakthrough volume for the
more volatile chemicals may prohibit adequate collection.
156
-------
Table 40. PEAK HEIGHT CALCULATIONS
Minimum Detectable
Concentration
Calculated Concentration Peak Height Based on 2 Chart
Compound in S.S. can (ppt) (Chart Division) Divisions (ppt)
High Level (^1 Par.t-Per-Billion)
Toluene
m-Dichlorobenzene
1 , 2-Dichloropropane
Furan
1,1, 2 ,2-Tetrachlo roe thane
Low Level (-\-100 Part-Per-Trillion)
Toluene
m-Dichlorobenzene
1 , 2-Dichloropropane
Furan
1,1,2, 2-Tetrachloroe thane
887 20,21,21 85.8
814 6,7,8 233
1220 2,5,3,3 914
3630 51,47,46 151
2350 8,7,8 613
88 1.5,1.3,2 110
81 N.D. >81
120 N.D. >120
359 5,5,3.5 160
232 N.D. >232
N.D. = Not detected.
-------
Table 41. AREA COUNT CALCULATION
Cn
Co
Compound
High Level (^1 Part-Per-Billion)
Toluene
m-Dichlorobenzene
1 ,2-Dichloropropane
Furan
1,1,2,2-Tetrachloroethane
Low Level OlOO P.art-Per-Trillion)
Toluene
m-Dichlorobenzene
1 ,2-Dichloropropane
Furan
1,1,2, 2-Tetrachloroe thane
Calculated Concentration
in S.S. can (ppt)
887
814
1220
3630
2350
88
81
120
359
232
Computer Calculated
Area Counts
636,640,639
34,N.D.,238
N.D.
1211,1206,1117
191,182,204
N.D.
N.D.
N.D.
121^164,168
N.D.
Minimum Detectable
Concentration
Based on 2 Chart
Divisions (ppt)
69
171
>1220
154
610
>88
>81
>120
119
>232
N.D. = Not detected.
-------
Table 42. CALCULATED CONCENTRATION OF TEST COMPOUNDS IN SOLVENT MIXTURE USED TO
DESORB NIOSH CHARCOAL TUBES3 - HIGH LEVEL STUDY
Compound
Vinyl chloride
Methyl bromide
Acrylonitrile
Furan
Chloroprene
Chloroform
Benzene
1,2-Dichloropropane
Toluene
Tetrachloroethylene
1 , 1 , 2 , 2-Tetrachloroethane
Chlorobenzene
Bis-(2-chloroethyl)ether
m-Di chlorobenzene
Maximum Wght. Collected
(ng)
3,672
5,587
6,054
8,340
487
2,196
5,742 .
3,326
2,714
6,102
8,438
2,346
8,410
. 3,606
Concentration in MeOH/CS,
(ng/p£)
3.67
5.59
6.05
8.34
0 . 49
2.20
5.74
3.33
2.71
6.10
8.44
2.35
8.41
3.61
Based upon 30 L collection of ppb vapors from permeation/dilutor system.
1.0 ml solvent mixture was used to desorb charcoal.
-------
On the other hand, GC/ECD did possess the inherent sensitivity to
detect seven of the chemicals in Table 42. In fact, five of the chemicals
possess a sufficient high electron affinity (21) to be detected in a 30 L
sample of ppt levels.
The recovery of chemicals measurable by ECD was in general poor. The
precision was erratic which prohibited the establishment of storage character-
istics. Their evaluation was terminated with the "high" level .interference
experiments with Group I compounds.
Cryogenic Traps--
Initial storage-stability studies with.empty cryogenic traps constructed
of nickel and cooled with dry ice yielded poor recoveries for all chemicals
at the ppb level. The cryogenic traps filled with clean glass beads (cryogen-
dry ice) gave better recoveries for a few chemicals at the ppt level; however,
the background was too elevated to measure accurately most of the chemicals.
At neither level were the more volatile chemicals detected.
In subsequent experiments involving potential interferences from inorga-
nic gases, liquid oxygen was used to-cool nickel traps containing glass
beads. Most compounds were detected, but the recoveries were still low and
the precision was poor.
The applicability of cryogenic traps of the design evaluated in this
study is limited. Because this method is labor intensive during sample
collection, sample transport and storage (requiring continual cryogen), and
sample recovery and analysis it is not a highly regarded technique. Substan-
tial quantities of water accumulate during sampling and must be removed
®
prior to sample analysis. Nafion tubing has been used to remove water;
however, polar substances are also lost (23).
The cryogenic trap theoretically has merits; particularly for very
volatile substances. Further work is needed in an optimal design and address-
ing the inherent problems of atmospheric sampling.
®
Tenax GC Cartridges--
§
The Tenax GC cartridge is limited principally by the breakthrough
volume which directly determines the detection limits attainable for a given
measurement technique. In this study it was limited by the 30 L sampling
160
-------
volume for chemicals with breakthrough volumes less than this. Similarly
the collection efficiency is directly related to the breakthrough volume.
Recoveries of chemicals were not significantly decreased by the short
term storage (7 days); in.fact storage of 2 to 3 months has been demonstrated
in other studies (22). Precision of recoveries was slightly less than those
®
observed for containers; however, with Tenax GC cartridges the recovery was
based upon triplicate sample analysis and not measurement of the same sample.
A major attribute of a cartridge sampling concept is its simplicity and
convenience in preparation, sampling, transport, and recovery and analysis.
Large numbers of samples can be taken simultaneously, stored until analysis
and analyzed relatively rapidly. Cartridges must, however, be protected
from sunlight.
Experiments with potential inorganic gas interferences readily demonstra-
ted a major problem which can occur with any collection device. Reactive
inorganic gases in the atmosphere can perturb the quantitative and qualitative
composition of the air sample during collection of organics. Substantial
improvement without adsorption losses can be attained by using a very small
amount of mild reducing agent to remove ozone prior to trapping.
161
-------
SECTION 7
SUPPORT AND QUALITY CONTROL DATA
SUPPORTING INFORMATION
Chromatographic Data
Analysis of Compounds from Cannisters—
The permeation tubes used to generate the test mixtures were placed in
the permeation/dilution system in small groups; samples were then taken and
analyzed. Using elution orders established by earlier Tenax studies (see
below) the GC peaks for the various.compounds were verified.
Analysis of Compounds from Tenax GC Cartridges--
The retention times for authentic compounds employed in this research
program were established and divided into two groups for conducting the
storage-stability and interference studies. These data for thermal desorp-
tion glass capillary gas chromatographic analysis are given in Tables 43 and
44.
Analysis of Compounds from Charcoal Cartridges--
. The compounds adsorbed to charcoal cartridges were recovered by solvent
desorption and microliter aliquots were analyzed by GC with flame ionization
and electron capture detection. The retention data which were established
by injection of authentic compounds are given in Tables 45 and 46. Charcoal
cartridges were not evaluated with the Group II compounds.
Calibration Data
Analysis of Compounds from Cannisters—
The FID response was standardized at three different compound concentra-
tions by drawing a sample directly from the glass manifold through a heated
1/8" o.d. stainless steel tube into the GC sampling system. Each concentra-
tion generated was calculated based upon the permeation rate of the compound,
the dilution air flow, and the volume of sample trapped-out. The
162
-------
Table 43. RETENTION TIMES FOR GROUP I COMPOUNDS ANALYZED BY TD/HRGC£
Retention Time + S.D. (C.V.)
Compound (min)
Vinyl chloride 4.99 + 0.035 (0.7)
Methyl bromide 5.65 + 0.023 (0.4)
Furan 6.88 + 0.058 (1.2)
Acrylonitrile 7.95 + 0.072 (0.9)
Chloroprene 10.71 + 0.055 (0.5)
Chloroform 11.81 + 0.030 (0.2)
Benzene 14.38 + 0.084 (0.5)
1,2-Dichloropropane 16.34 + 0.015 (0.09)
Toluene 20.81 + 0.091 (0.4)
Tetrachloroethylene 24.43 + 0.010 (0.04)
Chlorobenzene 25.20 + 0.050 (0.02)
Tetrachloroethane 27.92 + 0 (0)
Bis-(2-Chloroethyl)ether 31.82 + 0.023 (0.07)
m-Dichlorobenzene 33.39 + 0 (0)
1,2,3-Trimethylbenzene. 34.39 + 0.010 (0.02)
See Table 8 for operating parameters.
163
-------
Table 44. RETENTION TIMES OF GROUP II COMPOUNDS ANALYZED BY TD/HRGC3
Compound
Methyl chloride
Propylene oxide
Vinylidene chloride
Allyl chloride
1,1, 1-Trichloroethane
a-Epichlorohydrin
Methyl mercaptan
Ethylbenzene
o-Xylene
Benzyl chloride
n-Decane
1,2, 3-Trimethylbenzene
o-Cresol
Nitrobenzene
Retention
Time (min)
6.60
8.80
9.56
10.05
15.13
18.65
20.23
26.97
28.59
33.92
34.61
35.04
36.08
37.15
Temp. (°C)
10.4
19.2
22.2
24.2
44.5
58.6
64.9
91.9
98.4
120
122
124
128
133
See Table 8 for operating parameters, mean of three determinations.
164
-------
Table 45. RETENTION CHARACTERISTICS OF MODEL COMPOUNDS RECOVERED
FROM CHARCOAL CARTRIDGES AND ANALYZED BY GC/FID3
Compound Retention Time (Min)
Acrylonitrile
Furan
Chloroform
Chloroprene
1 , 2-Dichloropropane
Benzene
1 , 1 ,2 , 2-Tetrachloroethane
Tetrachloroethylene
Toluene
Chlorobenzene
Bis-(2-chloroethyl)ether
m-Dichlorobenzene
1,2, 3-Trimethylbenzene
4.0
4.8
9.6
16.0
16.5
19.2
24.6
25.8
27.6
28.8
29.2
36.8
46.5
Temperature (°C)
50
' 50
70
110
113
130
160
168
180
186
190
210
210
Chromatographic conditions were given in Table 9, mean of three
determinations.
165
-------
Table 46. RETENTION TIMES OF MODEL COMPOUNDS RECOVERED FROM
CHARCOAL CARTRIDGES AND ANALYZED BY GC/ECD3
Compound Retention Time (min)
Chloroform 2.0
Chloroprene 3.2
1,2-Dichloropropane 3.7
1,1,2,2-Tetrachloroethane 11.5
Tetrachloroethylene 13.5
Bis-(2-chloroethyl)ether 22.5
o
Operating parameters were given in Table 9, mean of three determina-
tions .
166
-------
concentrations used in establishing a standard curve and the correlation
coefficient for each curve are given in Table 47.
Tables 47 and 48 give the calibration data for Group I and Group II
compounds, respectively. Using a four point calibration (including zero) a
linear regression was established. Table 49 gives calibration data for
Group I compounds in the interference study. The interference study for
Group II compounds was performed immediately following the storage stability
study for Group II compounds. A calibration check showed no change in the
GC system response to standards and so the data presented in Table 48 was
used as calibration data for the interference study with Group II compounds.
The correlation coefficients for each compound are also provided. Each
measurement was performed in triplicate.
®
Analysis of Compounds from Tenax GC Cartridges--
The absolute retention times for model compounds were determined on an
SE-30 SCOT column (Table 8).
In order to determine the quantities of each of the compounds collected
on the sampling devices, the gas chromatographic system was calibrated in
the range that was anticipated to be collected at the low level and essen-
tially consisted of an extension of the standard curve for each of the
i1
compounds at the high level. The procedures for calibrating the gas chroma-
tographic system werepidentical to that described for the high levels employ-
ing both permeation tube and a flash unit for cross-checking the slope of
the standard curves.
r
Table 50 presents the results of instrumental calibration for Tenax
cartridge analysis (high level). Except for one case, the correlation
coefficients of the standard curves were 0.99, while for acrylonitrile it
was 0.92. A small background peak reduced the accuracy of the standard
curve for acrylonitrile.
The calibration results for the low level studies are given in Table
51. Only those compounds which were detected in the low level study are
indicated. The mass range simply was an extension of the high level calibra-
tion. The area obtained, standard deviation and coefficient of variation
for triplicate analysis at each mass are presented. Correlation coefficients
and slopes were calculated for the linear regression data. Except for
167
-------
Table 47. GROUP I COMPOUND CALIBRATION DATA FOR BAGS, BULBS AND
CANNISTERS
FEBRUARY, 1981
Compound
Vinyl chloride
Methyl bromide
Furan/acrylonitrile
Chloroprene
Chloroform
Benzene
1 , 2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2, 2-Tetrachloroethane
Bis (2-Chloroethyl) ether
1,2, 3-Trimethylbenzene
m-Di chlorobenzene
Concentrations, ppb
34,
86,
170,
6.2,
17,
64,
30,
27,
33,
19,
40,
23,
113,
22,
18,
45,
89,
3.2
8.9
33,
16,
14,
17,
10,
21,
12,
59,
11,
11
29
57
, 2.1
, 5.8
22
10
9.0
11
6.5
14
7.8
38
7.4
Correlation Coeff.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.999
.999
.998
.999
.995
.999
.999
.998
.998
.998
.998
.999
.999
.999
168
-------
Table 48. GROUP II COMPOUND CALIBRATION DATA FOR BAGS, BULBS AND
CANNISTERS
JUNE, 1981
Concentrations Correlation
Compound ppb Coefficient
Methyl chloride" 223,112,56.6 0.999
a
Propylene oxide
Vinylidene chloride 82.6,41.4,20.9 0.999
Allyl chloride 361,181,91.4 0.999
1,1,1-Trichloroethane 41.9,25.6,12.8 0.999
o
ct-Epichlorohydrin
Methyl mercaptan
Ethylbenze'ne
o-Xylene
Benzyl chloride
n-Decane
1,2, 3-Trimethylbenzene
o-Cresol
Nitrobenzene
118,59.0
29.8,14.9,7.5
68.7,42.0,21.0
116,57.9,29.2
30.8,15.4,7.8
234,143,71.5
53.9,27.2
0.998
0.999
0.997
0.998
0.999
Not detected in sample containers and poor integration in standards.
Not detected in standard or sample containers.
c
Two point calibration - non-linear response at high concentrations.
169
-------
Table 49. GROUP I COMPOUND CALIBRATION DATA WITH BAGS, BULBS AND
CANNISTERS FOR INTERFERENCE STUDIES
JULY, 1981
Concentrations Correlation
Compound (ppb) Coefficient
Vinyl chloride 52.8,38.6,27.7,13.9 0.999
Methyl bromide 114.5,83.8,60.1,30.1 0.999
a
Furan/Acrylonitrile
Chloroprene
Chloroform
Benzene 61.7,45.2,32.4,16.2 0.999
1,2-Dichloropropane 33.3,24.4,17.5,8.8 ' 0.995
Toluene' 25.9,19.0,13.6,6.8 0.992
Tetrachloroethylene0 . 17.3,12.4,6.2 0.999
Chlorobenzenec 12.8,9.2,4.6 0.997
1,1,2,2-Tetrachloroethane 35.2,25.8,18.5,9.3 0.993
Bis-(2-chloroethyl)ether 4.6,3.3,2.4,1.2 0.999
m-Dichlorobenzene 21.0,15.3,11.0,5.5 0.984
Based on furan only, acrylonitrile tube polymerized.
Not quantified due to interfering peaks..
Three-point calibration - non-linear response at high concentration.
170
-------
Table 50. GROUP I COMPOUND CALIBRATION DATA FOR
TENAX CARTRIDGE ANALYSIS - HIGH LEVEL
Methyl bromide
Vinyl chloride
Benzene
Chloroform
Toluene
1 , 2 , 3-Trimethylbenzene
Furan '
Bis (2-Chloroethyl)ether
1 , 1 , 2 , 2-Tetrachloroethane
Acrylonitrile
Tetrachloroethylene
424
851
1273
100
200
300
139
208
277
100
200
300
68
102
135
226
451
902
200
300
400
449
903
1797
417
627
835
1671
145
217
289
303
455
607
1217
10.87 + 0.94 (8.6) ^
18.73 + 0.55 (2.9) >
27.53 + 1.26 (4.6) J
5. .68 + 0.76 (13. 4n
9.10 + 0.86 (9.4) >
14.57 + 0.58 (3.9) J
37.48 + 2.94 (7.8) ^
58.03 + 1.65 (2.8) >
80.96 + 0.25 (0.3) J
2.94 + 0.15 (5.1) "^
4.79 + 0.10 (2.1) (
6.69 + 0.76 (11.4)-'
25.40 + 2.02 (7.9) ^
28.99 + 2.53 (8.7) >
37.26 + 1.44 (3.9) J
19.95 + 1.05 (5.3) "\
44.07 + 1.28 (2.9) /
90.97 + 2.42 (2.7) J
22.28 + 2.46 (11.0)^
32.55 + 2.18 (6.7) (
53.26 + 1.44 (2.9) J
0.34 + 0 (0) ""(
0.62 + 0.04 (6.8) }
1.25 + 0.12 (9.6) J
100.42 + 4.19 (4.8) ^
155.86 + 11.89 (7.6) I
213.22 + 3.71 (3.7) (
455.38 + 12.24 (2.7)J
11.46 + 0.86 (7.5) >
13.16 + 2.34 (17.8) >
33.33 + 4.76 (14.3)J
96.96 + 5.50 (5.7) "^
148.81 + 11.58 (7.8) I
198.46 + 4.66 (2.3) f
430.72 + 7.83 (1.8) J
0.99
0.99
0.99
0.99
0.99
0.99
0.99
0.99
0.99
0.92
0.99
(continued)
171
-------
Table 50 (cont'd.)
Compound
1,2-Dichloropropane
Chlorobenzene
m-Dichlorobenzene
Chloroprene
Mass
(ng)
166
250
333
666
104
208
311
166
333
498
50
100
200
Area + S.D. (C.V.) Correlation
(mean) Coefficient
16.24 + 0.17 (1.0)^
25.54 + 0.80 (3.2) I
36.54 + 2.23 (6.1) f
72.24 + 1.99 (2.7)J
17.02 + 1.20 (7.0)^
32.19 + 0.56 (1.7) >
48.88 + 2.62 (5.4) J
21.40 + 1.84 (8.6)">
38.03 + 0.89 (2.3) >
58.12 + 3.87 (6.7) J
5.02 + 0.73 (14.5)^
8.58 + 1.12 (13.0) >
15.89 + 1.34 (8.4) J
0.99
0.99
0.99
0.99
a
Permeation tubes were used.
172
-------
Table 51. GROUP I COMPOUND CALIBRATION DATA FOR TENAX CARTRIDGE ANALYSIS - LOW LEVEL
Compound
Benzene
1 , 2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2 ,2-Tetrachloroethane
Bis -(2-Chloroethyl) ether
m-Dichlorobenzene
Mass
(ng)
28
100
192
22
81
155
14
48
92
30
105
201
12
42
80
64
231
443
139
267
20
70
135
Area + S.D. (C.V.) Correlation
(mean) Coefficient
4.30 + 0.87 (20)^
20.64 -i- 2.49 (12) > 0.99
37.04 + 3.21 (9) J
1.20 + 0.05 (4) "N
4.55 + 0.46 (10) > .0.99
9.62 + 0.52 (5) J
0.72 + 2.98 (412A
9.60 + 3.82 > 0.98
27.74 + 3.89 (14) J
0.55 + 0.24 (44) "^
3.26 + 0.30 (9) > 0.99
11.83 + 1.02 (9) J
0.93 + 0.38 (41)*N
5.75 + 0.95 (17) } 0.99
11.83 + 1.02 (9) J
1.16 + 0.12 (10A
4.45 + 0.38 (9) > 0.99
7.28 + 0.61 (8) J
3.50 + 0.31 (9) "\
12.62 + 1.57 (12) J
1.77 + 0.60 (34A
8.65 + 0.67 (8) > 0.99
16.69 + 0.62 (4) J
Slope
0.200
0.062
0.310
0.037
0.150
0.017
-
0.130
Permeation tubes were used.
-------
toluene which had a correlation coefficient of 0.98, all of the remaining
correlations were 0.99.
Data for Group II compounds are given in Table 52.
QUALITY CONTROL DATA
Preparation and Calibration of Permeation Tubes
During the course of this research project permeation tubes were prepared
for the model compounds of interest, some of which are listed in Table 53
with construction materials and dimensions. In order to access the range of
permeation rates attainable, permeation tubes for each model compounds were
prepared using three kinds of plastic tubing. Surgical grade polyethylene
. ffi\
(PE) and two types of Teflon : tetrafluoroethylene (TFE) and perfluoroethyl-
enepropylene (FEP). The lengths varied from a few mm to 15 cm. The ends
were sealed with glass plugs secured by stainless steel ferrules. This
technique proved useful since a vapor pressure well above one atmosphere for
dimethylamine could be easily contained.
The selection of the appropriate plastic tubing for organics becomes a
facile process once the behavior of a few compounds for a chemical class is
®
known. In general, the non-polar liquids/gases are prepared in the Teflon
tube while the more polar, lesser volatile materials permeate at the desired
rate in polyethylene.
In previous studies it was determined that permeation systems at 20°C
resulted in severe adsorptive/condensation losses; thus all permeation tubes
were equilibrated at 30°C.
The permeation tubes were gravimetrically calibrated on a bi-weekly
schedule using a Mettler MS-5A (microgram) or Cahn balance.
Permeation tubes which were periodically gravimetrically calibrated
were used throughout all studies to synthesize air-vapor mixtures and for
calibrating instruments. Historical records of the permeation rates experi-
enced during the storage-stability and interference studies are given in
Tables 54 and 55. Each permeation rate represents a linear regression
analysis of at least five previous weighings. The mean, standard deviation
and coefficient of variation are also indicated.
The permeation rates for most of Group I compounds were stable with
coefficient of variations under 10% during the 11' month period.
174
-------
Table 52. GROUP II COMPOUND CALIBRATION DATA FOR
TENAX-CARTRIDGE ANALYSIS3
Compound
Methyl chloride
Propylene oxide
Vinylidene chloride
Allyl chloride
1,1, 1-Trichloroethane
a-Epichlorohydrin
Methyl mercaptan
Ethylbenzene
o-Xylene
Benzyl chloride
n-Decane
Mass
(M8)
583
2,915
5,830
239
1,036
1,915
294
1,273
2,351
1,116
4,830
8,925
240
1,040
1,922
907
3,919
7,235
293
1,465
2,930
194
839
1,549
174
751
1,386
238
1,032
1,907
263
1,136
2,097
Area :
Mean + S.D. (C.V.) Slope
102 + 23 (23) ^ ,
405 + 81 (2) > 0.953
504 + 174 (35)J
202 + 50 (25) "N
' 657 + 116 (18) } 0.505
1,050 + 87 (8.3)J
245 + 19 (7.7) ^
733 + 110 (15) > 0..452
1,176 + 152 (13) J
719 + 59 (8.2) •>
2,913 + 139 (4.8) > 0.520
4,790 + 210 (4.4)J
110 + 5.6 (5.1)"N
556 + 8.6 (15) } 0.424
826 + 108 (13) J
290 + 71 (24) ^
1,775 + 86 (4.8) > 0.381
2,715 + 293 (11) J
31.1 + 3.2 (10)>|
160 + 43 (27) \ 0.115
333 + 92 (28) J
246 + 48 (20) -\
894 + 15 (1.6) \ 1.015
1,620 + 144 (8.9)J
244 + 12 (4.9) "N
925 + 23 (2.5) > 1.236
1,740 + 166 (9.5)J
10.9 + 9.5 (87) "^
225 + 169 (75) > 0.236
' 362 + 213 (59) J
234 + 7.9 (3.4)>v
962 + 57 (5.9) i 0.966
2,001 + 242 (12) J
Correlation
Coefficient
0.953b
0.976
0.972
0.995
0.962
0.977
0.933
0.992
/"» s\ r\ 1
u . yyj.
0.737
0.984
(continued)
175
-------
Table 52 (cont'd.)
Mass Area: Correlation
Compound (fjg) Mean + S.D. (C.V.) Slope Coefficient
1,2,3-Trimethylbenzene 1,038 847 + 13 (1.5) "N
4,492 3,401 + 32 (0.9) V 0.788 0.995
8,299 6,568 + 479 (7.3)J
o-Cresol 582 65.1 + 23.2 (
2,518 759 + 240 (32) > 0.258 0.839
4,648 1,122 + 504 (45)
}-
.2) -N
.9) >!.!
12.6)J
Nitrobenzene 397 745 +24 (3.
1,716 3,377 + 97(2.9) } 1.959 0.986
3,169 6,177 + 775 (12.6).
Includes cryogenic traps; permeation tubes were used.
Based on medium and low levels only.
176
-------
Table 53. PERMEATION TUBES PREPARED AND AVAILABLE DURING THE PAST YEAR0
Compound
Chloroform
1,1,2, 2-Tetrachloroethane
1 , 2-Dichloropropane
Vinyl chloride
2-Chloro-l,3-butadiene
Tetrachloroethylene
Chlorobenzene
m-Dichlorobenzene
Benzene
Toluene
1,2, 3-Trimethylbenzene
Acrylonitrile
Furan
Bis (2-chloroethyl) ether
Methyl bromide
1, 1,1-Trichloroethane
Methyl chloride
1 , 1-Dichloroethylene
Benzyl chloride
Propylene oxide
Polymeric
Material
FEP
.PE
TFE
TFE
TFE
TFE
TFE
TFE
TFE
TFE
PE
TFE
FEP
PE
TFE
TFE
TFE
TFE
PE
TFE
Dimensions
(mm i.d. x mm length)
0.476 x 107
0.317 x 5
0.476 x 77
-
0.476 x 95
0.476 x 52
0.476 x 55
0.476 x 60
0.476 x 125
0.476 x 70
0.317 x 8
0.476 x 48
0.476 x 113
0.317 x 80
0.476 x 50
0.476 x 80
0.476 x 10
0.476 x 101
0.317 x 80
0.476 x 51
Permeation Rate
Supplier ((Jg/min)
RTI
RTI
RTI
Kin-Tek
RTI
RTI
RTI
RTI
RTI
RTI
RTI
RTI
RTI
RTI
Metronics
RTI
Metronics
RTI
RTI
RTI
0.1569
0.6388
0.2242
0.2980
0.1346
0.4335
0.1617
0.2583
0.5135
0.1938
0.3668
0.4363
0.5928
1.3020
0.6700
0.0943
0.4000
0.0556
NEC
0.9364
(continued)
-------
Table 53 (cont'd.)
00
Compound
1 ,4-Dioxane
Phenol
o-Cresol
Acrolein
Dimethylamine
Di-n-butylamine
Pyridine
Aniline
t-Butyl mercaptan
Nitrobenzene
Polymeric
Material
TFE
PE
PE
TFE
FEP
TFE
TFE
TFE
TFE
PE
Dimensions
(mm i.d. x mm length)
0.476 x 102
0.317 x 96
0.317 x 109
0.476 x 100
0.476 x 70
0.476 x 52
0.476 x (?)
0.476 x 98
0.476 x 74
0.317 x 10
Supplier
RTI
RTI
RTI
RTI
RTI
RTI
RTI
RTI
RTI
RTI
Permeation Rate
(pg/min)
0.0791
0.4337
0.7897
0.2692
0.2576
0.7289
0.1339
0.0492
0.0857
NE
Listing includes those prepared through October, 1979.
3Rates are for 30°C.
"NE = not equilibrated.
-------
Table 54. HISTORICAL RECORD OF PERMEATION RATES FOR GROUP I COMPOUNDS DURING
STORAGE-STABILITY AND INTERFERENCE STUDIES3
Month
Compound
Vinyl chloride
"ethyl brtalde
Furao
Acrylonilrtle
Chloroprenr
Chloroform
§*ni*ne
l,2-Dicbloroprop*ne
Tolueoe
Tetnchloroetbrlene
ChlorobeaieQe
l.l,2,2-Tetr«cbloroetb»ae
BU-(2-chloroetbyl)etber
•-Dlcblorobeaiene
JAB (1980)
FEB
HAH
26."
6,0b
465 1
459 •
17 (3.7)
47 (10)
21 3b
154 •
418 *
241 t
196 •
437 •
163 »
608 •
-
256 *
5 (3.2)
14.6 (3.5)
60 (25)
8.4 (4.2)
19 (4.3) '
28 (17)
124 (20)
31 (12)
471
432
226
160
419
264
196
442
164
524
584
259
1 15 (3.2)
* 8.1 (1.9)
• 19 (8.4)
• 4.3 (2.7)
1 13,9 (3.3)
* 78 (29)
I 8.4 (4.3)
• 13 (2.9)
• 32 (19)
« 28 (5)
» 35 (6.0)
* 28 (11)
465
431
218
162
420
288
194
448
16]
545
580
234
• 10 (2.1)
• 10 (2.3)
t 31 (14)
* 4.0 (2.5)
» 14.2 (3.4)
• 76 (26)
« 7.5 (3.1)
« 6.5 (1.5)
« 14 (8.5)
« 66 (12)
« 26 (4.4)
« 20 (7.8)
AFR
268b
670b
465
431
216
160
423
332
200
441
173
545
584
273
• 10 (2.1)
• 8.8 (2.0)
1 31 (14)
* 4.2 (2.6)
• 2.1 (0.5)
• 43 (13)
* 13 (6.5)
• 13 (2.9)
I 1.1. (1)
• 60 (11)
« 19 (3.2)
. 46 (17)
RAY
268 • 5 (2)
670 • 5 (1)
607 • 16 (3)
755 • 34 (5)
23 • 2 (9)
160 * 5 (3)
420 • B (2)
^38 » 41 (12)
201 * 13 (6)
438 < 18 (4)
175 * 6 (3)
968 * 50 (5)
584 * 10 (2)
294 * 66 (22)
sum.
268
670
605
431
175
160
420
287
196
445
175
574
580
261
; 5 (2)
1 5 (1)
• 15 (2.5)
• 11 (2.5)
« 7 (4)
1 4 (2.5)
• 14 (3)
« 7S-(26)
« 7 (3.5)
* 9 (2)
• 7 (4)
« 99 (17)
« 14 (1.4)
• 25 (9.5)
JULY
171b'C
670 * 5 (1)
458 • 9 (21
421 * 5 (1)
68'
158 • 6 (3.8)
418 • 8 (2)
351 1 18 (5)
205 * 10 (5)
443 . 14 (3.2)
183 « 16 (9)
183 » 16 (9)
609e
314 • 61 (19)
AIKi
171
670
455
420
' 16
162
418
352
207
434
1B6
630
586
326
• 13 (2)
* 2 (0.4)
. 3 (0.7)
< 31 (194)
• 6 (4)
» 8 (2)
» 17 (5)
» B (4)
* 15 (3)
* 15 (B)
• 60 (10)
» 5 (0.9)
« 50 (15)
stp
171
670 « 13 (2)
453 > 4 (0.9)
417 • B (2)
32 • 33 (103)
161 « 5 (3)
419 • B (2)
157 » 27 (B)
108 • B (4)
439 > 25 (6)
188 * 14 (7)
562 • 16 (1)
5»0 * 16 (3)
127 < 49 (15)
OCT
171 * 3 (2)
670 • 11 (2)
449 « 11 (2)
413 > 10 (2)
32 • 30 (94)
159 « 7 (4)
421 • 7 (2)
144 * 41 (14)
206 • 16 (1)
450 « 11 (i)
189 » 11 <«>
596 * 63 (11)
111 « lit (11)
110 « 17 (IB)
•0V
171
-
649 * 10 (1.1)
411 • 9.8 (1.4)
: i 4 (in
163 • 14 (85)
422 t B.7 (2.1)
_b
107 • 16 (7.9)
443 * 42 (9.4)
181 « 10.7 (].!}
196 « 61 (10) '
531 « 111 (11)
_b
Permeation rates are derived from at least five gravimetric determinations (over prior 3 month period)
and linear regression analysis - Mean (ng/min) +_ S.D. (C.V.).
Not statistically analyzed.
New permeation tubes.
-------
CO
o
Table 55. HISTORICAL RECORD OF PERMEATION RATES FOR GROUP II COMPOUNDS DURING
STORAGE-STABILITY AND INTERFERENCE STUDIES3 -
Compound
Methyl chloride
I'ropylene oxide
Vinylidene chloride
Allyl chloride
1 , 1 , 1-Trichlo roe thane
a-Epichlorohydrin
Hethyl mercaptan
Ethy Ibenzene
o- Xy lene
Benzyl chloride
n-Decane
1.2, 3- Trimethy Ibenzene
o-Cresol
Nitrobenzene
•IAN (1981)
45,b
270 * 7.8 (2.9)C
-
1,447 * 64 (4.4)
263 «_ 17 (6.4)
-
-
136 t_ 4.9 (3.6)
-
243 * 7.2 (3.0)
244 * 44 (18)
1,099 *_ 74 (6.8)
554 +_ 74 (13)
276 «_ 17 (6.2)
451 * 3.1
267 « 7.5
-
1,413 * 81
255 » 15
-
529b
136 * 4.2
-
242 * 8.5
233 * 47
1,101 + 73
S13 * 69
'297 + 57
n
(0.7)
(2.8)c
(S.7)
(5.8)
(3.1)
(3.5)
(20)
(6.7)
(13)
(19)
Month
:B
452
264
1,377
249
135
242
216
1,101
483
292
1 3.2
*_ 5.9
-
i 82
+ n
629b
514b
+_ 5.8
-
1 8-4
+ 42
± 73
*_ 66
1 58
(0.7)
(2.2)C
(6.0)
(4.3)
(4.3)
(3.5)
(19)
(6.7)
(14)
(20)
456
264
1,343
246
513
135
240
204
1,115
458
297
» 9.0
«_ 4.8
-
+ 76
+ 7.2
696b
1 18
t 6.9
-
1 7-4
± 32
» 64
» 61
^ 55
MA
(2.0)
(1.8)
(5.6)
(2.9)
(3.4) .
(5.1)
(3.1)C
(16)
(5.7)
(13)
(19)
R
457
263
1,305
254
685
506
134
243
189
1.106
438
294
*_ 9.2 (2.0)
*_ 4.6 (1.7)
153b
*_ 58 (4.4)
+ 25 (9.8)
i si ;- s)
«_ 20 (3.9)
*_ 7.3 (5.4)
-
*. 12 (4.9)C
*_ 19 (10)
* 77 (6.9)
+53 (12)
i 57 (19)
(continued)
-------
Table 55 (Cont'd.)
Month
Compound
Apr .
May
Methyl chloride
Propylene oxide
Vinylidene chloride
Allyl chloride
1,1,1-Trichloroethane
ot-Epichlorohydrin
Methyl mercaptan
Bthylbenzene
o-Xylene
Benzyl chloride
n-Decane
1,2,3-Trimethylbenzene
o-Cresol
Nitrobenzene
459 + 9.3 (2.0)
264 + 4.4 (1.7)
316b
1,275 + 43 (3.4)
263 + 30 (11)
645 + 120 (19)d
498 + 25 (5.1)
135 + 7.4 (5.5)
243 + 12 (4.9)°
182 + 13 (6.9)
1,072 + 48 (4.4)
412 + 17 (4.1)
290 + 59 (20)
461 + 8.5 (1.9)
264 + 4.6 (1.7)
316b
1,249 +43 (3.4)
262 + 30 (12)
603 + 114 (19)d
493 + 25 (5.1)
135 + 7.3 (5.4)
301b
242 + 13 (5.3)C
183 + 12 (6.4)
1,046 +61 (5.8)
404 + 11 (2.0)
286 + 62 (22)
464 + 5.3 (1.1)
263 + 5.3 (2.0)
323 + 12 (3.9)°
1,226 + 49 (4.0)
264 + 29 (11)
557 + 43 (7.8)d
484 + 18 (3.7)
133 + 7.5 (5.6)
119b
244 + 12 (4.7)C
180 + 10 (5.3)
1,040 + 67 (6.4)c
399 + 11 (2.7)
272 + 29 (11)
465 + 3.0 (0.6)
265 + 6.6 (2.5)
326 + 12 (3.5)C
1,202 +51 (4.3)
281 + 41 (14)
578 + 11 (2.0)d
477 + 11 (2.3)
134 + 7.0 (5.2)
129b
249 + 14 (5.5)C
179 + 9.4 (5.3)
1,029 + 57 (5.6)c
395 + 9 (2.3)
271 + 33 (12)
(continued)
-------
Table 55 (Cont'd.)
Month
Compound Jun
Jul
Methyl chloride
Propylene oxide
Vinylidene chloride
Allyl chloride
1,1,1-Trichloroethane
a-Epichlorohydrin
Methyl mercaptan
Ethylbenzene
o-Xylene
Benzyl chloride
n-Decane
1, 2,3-Trimethylbenzene
o-Cresol
Nitrobenzene
464 + 2.3 (0.5)
262 + 7.4 (2.8)
327 + 10 (3.2)
1,167 + 62 (5.3)
291 + 38 (13)
750b
473 + 7.5 (1.6)
129 + 6.6 (5.1)
146b
401b
178 + 8.4 (4.7)
l,519b
385 + 21 (5.6)
56>
464 + 2.5 (0.5)
260 + 7.8 (3.0)
329 + 10 (3.1)
1,134 + 71 (6.2)
280 + 43 (15)
l,012b
464 + 16 (3.5)
130 + 6.8 (5.2)
366b
601b
180 +8.1 (4.5)
l,407b
376 + 22 (5.9)
544b
461 + 8.8 (1.9)
259 + 7.2 (2.8)
332 + 8.1 (2.4)
1,096 + 75 (6.9)
270 + 45 (16.6)
831b
464 + 16 (3.5)
128 + 6.1 (4.7)
309b
592b
175 + 14 (8.1)
l,298b
369 t 23 (6.2)
448b
451 + 20 (4.5)
259 + 7.2 (2.8)
329 + 15 (4.5)
1,039 + 114 (11)
270 + 45 (17)
739b
452 + 30 (6.6)
123 + 9.8 (8.0)
320b
414b
174 + 16 (9.0)
2,372b
355 + 30 (8.4)
265b
Permeation rates are derived from six gravimetric determinations (except for new tubes) and linear regression
analysis - Mean (ng/min) + S.D. (C.V.).
Not statistically analyzed - new tube or tube with rapidly changing rate.
, One of six rates not included in estimation of mean. Excluded rate fails Q test at 90% confidence level.
Last three rates statistically analyzed - gradually changing rate.
-------
The permeation rate for chloroprene exhibited a very large coefficient
of variation and hence, the studies with this permeation tube are somewhat
suspect. However, it was not possible to control the permeation rate because
of its high instability and its propensity to polymerize. Gas chromatography/
mass spectrometry analysis was performed on Tenax GC cartridges loaded with
"old" and "new" sources of chloroprene in permeation tubes. These results
are shown in Figures 53 and 54. The relative proportions of chloroprene to
xylenes is higher for the "new" source indicating that the chloroprene in the
"old" permeation tube has probably to some extent polymerized.
Nevertheless, the permeation rates for the remaining test compounds were
ii
relatively stable.
The permeation rate for bis(2-chloroethyl)ether tube was at one point
permeating at an unacceptable rate and therefore a new permeation tube was
prepared using appropriate dimensions and materials of construction.
I
During the period between a pilot study and the storage-stability
study, the trimethylbenzene permeation tube developed a severe leak and could
not be used.. A new permeation tube was prepared.
I
Calculation of Breakthrough Volumes
Breakthrough volumes were determined for chloroprene, acrylonitrile and
@
furan on the Tenax GC sampling cartridges. The procedures have been previou-
sly described (9). Table 56 lists these breakthrough volumes.
Calibration of Ozone Monitor
On November 6, 1980, a multipoint ozone analyzer calibration was perfor-
med by personnel of RTI's Quality Assurance Department. The Bendix Ozone
analyzer, model 8002 EPA serial number 100586, was calibrated using the
laboratory's stable ozone source (ultraviolet lamp/quartz tube arrangement),
ultrapure air (Matheson), and a Dasibi ultraviolet photometer as an assay
reference. The calibration was performed in accordance with EPA recommended
procedures.
I
Just prior to the calibration, the Dasibi photometer response was
checked in the EQAD/SMD verification laboratory and was found to be in
excellent agreement with that of a CSI Photocal 3000 (S/N 10382) ozone
generator/photometer that is maintained as an instrument traceable to
183
-------
z
a —
o
TJ
Chloroprene
rylenea
Figure 53. GC/MS/COMP profile of chloroprene in vehicle carrier (xylenes) from permeation
tube (old source).
-------
s
i
I,
|o-*>
I ro re
2.-0
O
0.
Figure 54. GC/MS/COMP profile of chloroprene in vehicle carrier (xylenes) from permeation
tube (new source).
-------
Table 56. BREAKTHROUGH VOLUMES FOR TENAX GC CARTRIDGE3
Temperature (°F)
Compound
Chloroprene
Acrylonitrile
Furan
50
34
11
5
60
26
9
4
70
20
7
3
80
15
5
3
90
11
4
2
100
9
3
2
For 1.5 cm i.d. x 6.0 cm Tenax bed, values in liters.
186
-------
standards in the Quality Assurance Division of EPA's Environmental Monitoring
Systems Laboratory.
Results of the precalibration zero and span check and the calibration
itself are tabulated in Table 57. The configuration &f the equipment at the
time of the calibration is illustrated in Figure 55.
The concentrations of ozone as read from the Bendix analyzer agree very
well with those read from the photometer; the two readings agree within at
least 0.004 ppm over the range considered. Thus, readings made directly
from the front panel of the Bendix analyzer are quite acceptable.
Performance Verification of Permeation/Dilution Systems
Calibration of Mass Flow Meters and Controllers—
i
The flow rates on the portable permeation/dilutor system were checked
prior to its use in the studies using an NBS certified bubble flow meter.
The flow meters (Tylan) on the system were found to be in error by approxima-
tely 15%. The calibration of the flow meters at ambient temperatures was
not valid for gases at slightly elevated temperatures. Calibration curves
for various settings was made using the bubble flow meter in lieu of recali-
brating the flow meter readout at the elevated temperatures.
Verification of Dilution Factors--
i
An experiment was performed to check the dilution factors with the
recalibrated system. this consisted of collecting triplicate Tenax cartridges
from the manifold after dilutions were made (see Table 58) using two compounds,
benzene and tetrachloroethylene. Benzene and tetrachloroethylene were in N?
(NBS certified tanks)."
The flow capability thrpugh the permeation system was increased in an
attempt to minimize adsorption of polar organic vapors on glass surfaces in
the permeation/dilution system. A mass flow controller capable of up to 5
L/min replaced the 2 L/min controller.
A dilution ratio of 1:10 was verified by introducing benzene and tetra-
I!
chloroethylene from NBS certified cylinders and sampling the effluent (at
the manifold) with Tenax GC cartridges. Table 59 presents the results for
undiluted and 1:10 diluted primary standards by the permeation/dilution
system. ,
187
-------
Table 57. OZONE ANALYZER CALIBRATION RESULTS
00
CO
Analyzer: Bendix 8002, EPA S/N 100586
Initial Settings: Zero - 015 Span - 554
Final Settings: Zero - 015 Span - 570
Range: 0-0.5 ppm
Time Constant: 10 seconds
Ethylene Pressure: 20 psig
Calibration Photometer: Dasibi Model
1003 AH, S/N 2342
Airmass Flow Controller Settings:
#1: 3.14
Sample and Ethylene Flow: 0.5 COB, Rotameter //2: zero
Personnel :
Ozone
Generator
Sleeve
Setting, cm
Sokash, Shores, Demian
Photometer
Ozone
Concentration ,
ppm
//:
Bendix
Ozone ,
Concentration,
ppm
): 3.98
Difference
Bendix
Minus
Pho tome te r , ppm
Percent
Difference
Precalibration check, prior to span adjustments
8.0 0.529 0.485
After span adjustment
-0.044
-8.3
Closed
6.0
5.0
4.0
3.0
2.0
1.0
0.000
0.390
0.313
0.247
0.176
0.105
0.036
0.000
0.390
0.310
0.245
0.172
0.105
0.040
0.000
0.000
-0.003
-0.002
-0.004
0.000
+0 . 004
-
0.0
•'O.g
-0.8
-2.3
0.0
+11.1
As read from digital display of Dasibi photometer, less the zero offset.
As read from front panel meter of Bendix analyzer.
-------
STATION
ANALYZER
ANALYZER
TEST .
ATMOS-
PHERE
q
fe f M ft
— W " (Mfl
<7r
LL
S
PARTICULATE
^4 FILTER OZONE SAMPLE
NIFOLD -
£ERO SAMHLh
^ LAllAUol r
^ r-
i * ^ /
RO AIR
UPPLY
ZERO AIR
MANIFOLD
LINE
LINE
"1
DASIBI
BY
Figure 55. Schematic configuration used in calibrating the ozone
monitor.
189
-------
Table 58, EXPERIMENTAL PARAMETERS FOR TESTING DILUTION SYSTEM
Tenax cartridge no.
1A
IB
1C
2A
2B
2C
3A
3B
3C
Input (L/tnin)
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Dilution
Input
Aa B C
3.0
3.0
3.0
0.9 - 3.0
0.9 - 3.0
0.9 - 3.0
0.9 3.0
0.9 3.0
0.9 3.0
(L/min)
Take -out
ABC
_
_
-
0.9
0.9 -
0.9
- 0.9
0.9
0.9 -
Experiment designation.
190
-------
Table 59. OBSERVED LEVELS FOR BENZENE AND TETRACHLOROETHYLENE
FROM DILUTION SYSTEM
_
Tenax GC Benzene Tetrachloroethylene
Sample No. (ng) (ng)
1 (No dilution) 3294 3430
2 (1 -» 10 setting) 306 303
3 (!.-» 10 setting) 305 281
191
-------
Verification of GC Calibration Data
Tenax GC Cartridge Analysis—
For calibration of the thermal desorption/GC/chromatography data system,
two independent methods were employed. Since the thermal desorption GC/FID
system presently cannot be calibrated using liquid injection (solvent would
mask the region of interest) it was necessary to verify the calibration
utilizing the cartridge technique.
Calibration data were compared from two different methods for loading
test compounds on Tenax cartridges. A permeation system, a flash unit, and
NBS certified gases were the sources for comparison.
Figure 11 presented the schematic of a vaporization unit for loading
organics dissolved in methanol onto Tenax GC cartridges. This system has
been previously described (20). Several model compounds were loaded onto
®
Tenax GC cartridges by using the flash unit to obtain and compare linear
regressions (standard curves) with those developed from permeation tubes.
Vapor/nitrogen mixtures in bottled gases at known concentrations were
also used to check the calibration curves as part of the quality assurance
program.
Two compounds benzene and tetrachloroethylene were available as NBS
@
certified mixtures. These substances were loaded directly onto Tenax GC
cartridges by passing known volumes of vapor-nitrogen through the cartridge.
Thus, an' implied accuracy was obtained by comparing the three calibration
methods.
Calibration curves developed for Group I compounds utilizing permeation
tubes were also.verified using a flash unit method. Tables 60-62 present
these results. Presented are three different masses for which responses were
obtained (area as determined by a CDS 111 Varian Chromatography Data System).
Linear regressions were determined and correlation coefficients are given
(Table 60). The variation of slopes between the permeation system and the
flash unit were somewhat larger in the low level calibration than those
observed at the high level calibration. This was not unexpected since the
trace quantities which were measured were approaching the background levels
and quantifiable limits of the technique.
192
-------
Table 60. GROUP I COMPOUND CALIBRATION DATA USING FLASH UNIT METHOD - HIGH LEVELS
Compound
Benzene
1,2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2,2-Tetrachloroethane
Bis-(2-Chloroethyl)ether
m-Di chlorobenzene
Mass
(ng)
98
49
27
134
67
32
101
50
27
185
92
45
128
64
31
166
83
43
137
68
33
146
73
35
Area + S.D. (C.V.)
24.18 + 3.98 (17)
11.19 + 2.60 (23)
8.78 + 2.27 (26)
9.44 + 0.27 (3)
4.22 + 0.28 (7)
1.64 + 0.25 (15)
22.86 + 2.65 (12)
10.28 + 3.20 (31)
5.87 + 2.18 (37)
6.32 + 0.58 (9)
3.44 + 1.05 (31)
0.83 + 0.21 (25)
23.29 + 0.87 (4)
10.04 + 1.28 (13)
5.08 + 0.40 (8)
4.60 + 0.33 (7)
2.62 + 0.30 (11)
1.30 + 0.22 (17)
7.31 + 0.95 (13)
3.36 •*• 0.71 (21)
2.14 + 0.08 (4)
22.6 + 1.25 (6)
11.85 + 1.98 (17)
3.56 + 0.15 (4)
Correlation
Coefficient
0.99
0.99
0.99
0.98
0.99
0.99
0.99
0.99
Slope
0.240
0.072
0.230
0.036
0.18
0.028
0.052
0.160
-------
Table 61. COMPARISON OF CALIBRATIONS BETWEEN PERMEATION SYSTEM AND
FLASH UNIT METHODS - HIGH LEVELS OF GROUP I COMPOUNDS
Compound
Benzene
1 , 2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2, 2-Tetrachloroethane
Bis-(2-Chloroethyl)ether
m-Dichlorobenzene
Slopes
Permeation System
0.200
0.062
0.310
0.037
0.150
0.017
NDb
0.130
Flash Unit
0.240
0.072
0.230
0.036
0.180
0.028
0.052
0.160
A%a
+20
+16
-26
-3
+20
+65
ND
+23
* _ FU Slope - PS Slope v .„„
PS Slope
ND = not determined.
194
-------
Table 62. COMPARISON OF SLOPES FOR STANDARD CURVES BETWEEN PERMEATION
TUBE AND FLASH UNIT CALIBRATION METHODS - LOW LEVELS OF GROUP I COMPOUNDS
Compound
1,1,2, 2-Tetrachloroethane
Chlorobenzene
Tetrachloroethylene
Toluene
Dichloropropane
Benzene
Bis- (2-chloroethyl)ether
m-Dichlorobenzene
Chloroprene
Chloroform
Permeation tube
calibration
0.032
0.086
0.021
0.13
0.061
0.13
0.00032
0.00087
0.18
0.012
Flash unit
calibration
0.024
0.10
0.024
0.14
0.048
0.14
0.00034
0.00085
0.074
0.015
A%*
+25
-16
-14
-8
+21
-8
-6
+2
+59
-25
a „, _ Perm. Tube - Flash Unit . ft „
Perm Tube
195
-------
Sources of synthetic air/vapor mixtures that were commercially available
were procured for an independent assessment of the calibration of instrumental
systems. Where possible, NBS certified standards were used for cross-checking
the calibrations.
Comparison of the slopes of calibration curves generated by three
different techniques for benzene and tetrachloroethylene were attempted;
however, an interferent prevented a comparison of the NBS Benzene standard
with other methods. These results are given in Table 63.
Table 64 gives a comparison of percent recoveries that were calculated
for the test compounds from Tenax traps (T-) using two different calibration
techniques. These data give an implied accuracy of the methods since the
calibrations employ different approaches.
The final calibration of the instrument was carried out to verify the
response of the GC/FID to the Group II compounds. The calibration was
carried out using liquid injections of a mixture of most of the Group II
compounds in CS-. The calibration results are given in Table 65. Because of
some differences in response between the initial instrument calibration and
the final calibration, the latter was used to be applied to data generated in
the interference study to compute absolute values. The initial calibration
was used for compounds unable to be quantified using the liquid injections.
Calibration data for Group II compounds determined using permeation
tubes were compared to calibrations determined using the flash evaporative
system. A methanolic solution of the compounds which boil at 60°C or greater
was prepared and injected into a hot (250°C) helium stream and the vapors
thus generated was loaded onto Tenax cartridges downstream (Fig. 11). Only"
five determinations for each compound at two spiking levels were used. The
results are shown in Table 66. A direct comparison of the two calibrations
is shown in Table 67. The correlation coefficient of the calibrations using
the flash unit are not as high as might be expected, though this is probably
due to the small sample size.
The EPA did not provide to RTI Quality Assurance standards for use in
the storage-stability or interference studies involving Group I or Group II
compounds.
196
-------
Table 63. INSTRUMENT CALIBRATION USING DIFFERENT SOURCES OF
BENZENE AND TETRACHLOROETHYLENE3
Compound
Benzene
Tetrachloroethylene
NBS Certified
ca
0.022
Slope
Flash Unit
O.l4b
0.024
Permeation Tube
0.13
0.021
ft
Contamination.
b ®
Calibration was for analysis of Tenax GC cartridges by thermal
desorption GC with flame ionization detection.
197
-------
Table 64.ffl COMPARISON OF PERCENT RECOVERIES FOR TEST COMPOUNDS FROM
TENAX® GC TRAPS USING TWO DIFFERENT CALIBRATION TECHNIQUES
Compound
Chloroform
Benzene
1,2-Dichloropropane
Toluene
Tetrachloroethylene
Chlorobenzene
1,1,2, 2-Tetrachloroethane
Bis-(2-chloroethyl)ether
m-Di chlorobenzene
Mean percent recovery, +
Flash unit
(TQ = 0 da)
111 + 19 (17)a
59+8 (13)
126 + 17 (14)
88+9 (11)
85+6 (7)
90+6 (6)
122 + 5 (4)
102 + 5 (5)
137 + 6 (4)
S.D. (C.V.).
b._ _ Permeation Tube Calibration - Flask Unit
Permeation
(TQ = 0 da)
111 + 21 (19)
65 + 9 (14)
101 + 14 (14)
91 + 10 (11)
97+7 (7)
105 + 7 (7)
93+4 (4)
106 + 5 (5)
103 + 6 (4)
Calibration ,
-V 1 1
«»
0
+9.2
-25
+3.3
+12
-14
-31
+4
+33
nn<¥
Permeation Tube Calibration
198
-------
Table 65. CALIBRATION DATA FROM LIQUID INJECTION OF GROUP II COMPOUNDS
Compound
Propylene oxide
Vinylidene chloride
Allyl chloride
1,1,1-Trichloroe thane
a-Epichlorohydrin
Ethylbenzene
o-Xylene
Benzyl chloride
n-Decane
Mass
(ng)
1,339
6,695
10,042
1,180
, 5,900
8,850
867
4,335
6,502
880
4,400
6,600
1,100
5,500
8,250
730
3,650
5,475
Area
Mean + S.D. (C.V.)
solvent interference
solvent interference
solvent interference
241 + 49 (20)
1,143 + 96 (8)
1,586 + 7.5 (0.5)
329 + 98 (30)
1,457 + 34 (2)
2,153 + 27 (1.2)
757 + 5.2 (0.7)
4,038 + 646 (16)
5,455 + 62 (1.1)
786 + 44 (6)
4,233 + 681 (16)
5,763 + 44 (0.8)
630 + 36 (6)
3,381 + 298 (9)
3,709 + 441 (12)a
619 •!- 26 (4)
3,817 + 585 (15)
6,597 + 498 (8)
Correlation
Slope ' Coefficient
0.156 0.994
0.238 0.998
0.844 0.984
0.880 0.985
0.625 0.992
1.244 0.985
(continued)
-------
Table 65 (Cont'd.)
Compound
1,2,3-Trimethylbenzene
o-Cresol
Nitrobenzene
Mass
(ng)
894
4,450
6,675
1,027
5,140
7,710 .
1,204
6,020
' 9,030
Area
Mean + S.D. (C.V.)
697 + 41 (6)
3,343 + 379 (11)
4,581 + 210 (5)
657 + 50 (8)
3,490 + 314 (9)
5,017 + 70 (1.4)
644 + 79 (12)
2,601 + 342 (13)
3,595 + 27 (0.7)
Correlation
Slope Coefficient
0.679 0.989
0.656 0.996
0.380 0.990
3
High level calibration point was not used to obtain slope and correlation coefficient.
o
o
-------
Table 66. INSTRUMENT CALIBRATION USING FLASH UNIT METHOD
Mass (ng)
Correlation
Compound Low Level High Level Slope Coefficient
1,1,1-Trichloroethane 1,340 3,350 0.179 0.742
a-Epichlorohydrin 1,180 2,950 0.274 0.812
Ethylbenzene . 870 2,170 0.855 0.831
o-Xylene 880 2,200 0.907 0.833
Benzyl chloride 1,100 2,750 0.267 0.747
n-Decane 750 1,825 0.858 0.824
1,2,3-Trimethylbenzene 890 2,240 0.747 0.892
o-Cresol 1,030 2,570 0.391 0.883
Nitrobenzene 1,200 3,010 0.587 0.873
201
-------
Table 67. COMPARISON OF SLOPES FOR STANDARD CURVES BETWEEN
PERMEATION TUBE AND FLASH UNIT CALIBRATION METHODS -
TENTATIVE RESULTS
Compound
1,1, 1-Trichloroethane
a-Epichlorohydrin
Ethylbenzene
o-Xylene
Benzyl chloride
n-Decane
1,2,3-Trimethylbenzene
o-Cresol
Nitrobenzene
3 0/ _ Perm. Tube - Flash
Permeation Tube
Calibration
0.424
0.381
. 1.015
1.236
0.236b
0.966
0.788
0.258b
1.959
Unit v TOW
Flash Unit
Calibration
0.179
0.274
0.855
0.907
0.267
0.858
0.747
0.319
0.587
A%a
+58
-28
-16
+27
-13
+11
+5.2
-24
+70
'° Perm. Tube
Tentative value.
202
-------
Quality Control of GC Calibration Data; Containers—
The calibration gases were taken directly from the permeation/dilution
system into the gas chromatographic system described earlier. The sample
transfer line was heated and also was flushed with sample three times before
the sample was collected in the GC-cryogenic trap for measurement. This
flushing consisted of drawing sample from the permeation/dilution system by
vacuum (GC sample loading system). Until the GC system was pressurized to
one atmosphere (the GC cryogenic trap was by-passed during flushing). The GC
system calibration was not verified using other gaseous standards. This is
not considered a major problem as calibration was based on direct use of
permeation tubes which are considered as reliable as diluting a standard gas
(in a cylinder) from the ppm to the ppb level.
Blanks and Controls
Traps —
Storage and Stability Studies—Prior to initiating the collection and
®
storage study, Tenax. GC, charcoal and nickel cryogenic traps were examined
for background interference. Blanks were collected in triplicate from the
permeation/dilution system (no test compounds) for analysis at t-, t~ , and t?
during the storage study. This was to pinpoint any background interference
or false positive measurement that occurred during the analysis of samples.
Interference Studies—Sampling traps were also examined for cleanliness
prior to initiating this study. Blanks were taken of the permeation/dilution
system to determine the extent of background before proceeding with test
compounds and the introduction of inorganic gases. Control samples using
®
Tenax GC, charcoal and cryogenic traps were taken (sampling of air-vapor
mixture without presence of inorganic gases) for making a comparative analysis.
In addition, a glass fiber filter [impregnated with and without sodium
chiosulfate (1-5 mg)] was used prior to the Tenax GC cartridges.
All samples taken from the portable permeation system with the inorganic
gases present were collected in triplicate with and without a glass fiber
filter impregnated with 10% sodium thiosulfate in line between the manifold
and Tenax GC cartridges. This experimental design differentiated between
photochemical reactions occurring within the dilution bulb themselves and
®
reactions on Tenax GC. Comparison o'f the chromatograms with and without the
203
-------
filters showed a disappearance of the major peaks previously reported -
phenol, acetophenone, benzaldehyde (Fig. 56) - when glass fiber filters with
sodium thiosulfate were utilized. These results suggested that these com-
pounds were products of a reaction between the inorganic gases (probably
®
ozone) and Tenax GC itself.
During the interference experiments, solvent desorbed charcoal produced
two contaminants during analysis by GC/ECD. This background was not present
during the storage-stability studies. The two background constituents were
present in all of the charcoal tubes tested for background (i-.e. unexposed
tubes). Subsequently a new lot of NIOSH charcoal tubes was purchased from
Applied Science (Bellfonte, PA). Several of the new charcoal tubes were
solvent desorbed (methanol/carbon disulfide) and chromatographed using the
standard conditions for analysis of test model compounds. The identical
background was still present (Fig. 57). New sources of redistilled methanol
and carbon disulfide were obtained; however, the solvents used were clean
when examined by GC/ECD. Subsequently, the background components were
examined by GC/MS.
An extract of an unexposed charcoal tube was analyzed by GC/MS. These
results are given in Fig. 58 and 59. The first component was tentatively
identified as 2-ethylhexariol. The second appears to have an empirical
formula of C,H,_S_. The C.JL S contaminant interfers with measurement of
bis(2-chloroethyl)ether while 2-ethylhexanol coelutes with 1,1,2,2-tetrachloro-
ethane. The magnitude of 2-ethylhexanol, however, is small relative to the
1,1,2,2-tetrachloroethane response in the high and low level studies.
Cannisters--
Storage and Stability Studies — Bags, glass bulbs and steel containers
were checked for cleanliness after cleaning. This check consisted of filling
the containers with clean air and then analyzing this air as soon as possible
using the GC system described previously. The bags were cleaned and tested
immediately before use in the recovery studies. The containers were considered
clean if background peaks at retention times equivalent to those of test
compounds were essentially equivalent to those found with laboratory-supp-
lied, dry, clean air analyzed directly.
204
-------
•Ml
o
Ln
Air
Noise Nolle
Acetophenone
Nolae Noise
MM III! >l« lilt nil lilt ))•• ||f«
IHII IMII • IM»I l« lfl«' !•• H » CMK«tJHt*»>fM»lUK*l tllf IMCtM !•!• lit \
I ' | i I • I • • ' . • p .
III! «V««
MM i«;i
Reproduced from
best available copy.
Figure 56. GC/MS/COMP profile of background for 30 L air sample from permeation/dilution systems
with 340 ppb 03, 320 ppb NO , 200 ppb S02, and 90% humidity present. Major sources
of hydrocarbons traced to NO supply.
-------
Figure 57. Chromatogram depicting background for solvent desorbed
unexposed NIOSH charcoal tubes.
206
-------
LOZ
ig 'XcoD
a|qe|ieAe jssq I
RUN IDENTIFICATION 1712
SPECTRS FILE NflrE flfiD395
FILE POSITION £4
BACKGROUND 62
MAXIMUM INTENSITY
OUTPUT fWSS RBNGE
SCfiN SPEED 2
rwss 57
INTENSITY 1B0B0
'CHfiRCOfiL'
37 IB
40 TO
X OF TDTRL
492
ION
38.5
41
3547
43
3342
55 56 70 63 £9 98 42
2285 2876 1919 1838 1202" 754 576
RELflTIVE INTENSITY
-s
tt)
en
CO
c/i
T3
fD
n
CT
QJ
O
TV"
n
O
-s
o
QJ
(I
O
OJ
cr
ro
40
100-
200-
100-
400-
0 20 41
1
F
CO
ro
(JL>
o
OD
CO
3 6
3 B
B
Ln
txj
1
ro
•y
ro
0
-------
803
id
c
•-s
RUN IDENTIFICRTION 1712 "CHBRCOflL1
SPECTRP FILE NflME BflD395
FILE POSITION 67_
BflCKGROUND 65
MflXinuM INTENSITY B140 X OF TDTflL
OUTPUT MflSS RONGE 40 TO 492
SCRN SPEED 2
nflSS 122 75 47
INTENSITY 10000 B719 5815
ION
24.3
40
e
20
45 46 91 77 124 61 76
2766 2420 1941 1216 1083 805 803
RELflTIVE INTENSITY '
40 60 80
en
QJ
1/1
CO
too-
CD
o
o
-h
cr
O>
o
O
200-
n
4>
a
(-•
o
w
S3
r>
o
3
N
300-
O
O
-s
n
o
CT
(T)
400-
-------
Blank samples for each container type were loaded when the test
compounds were loaded. A typical analysis set was three containers of one
type containing test compounds and one container of that type containing
clean air. More control samples for each type of container were not prepared
and analyzed as each sample required typically 1 1/2 hours for analysis.
Background peak areas found in the control samples for each container type
on Days 0, 3 and 7 were subtracted from the areas determined for the test
compounds on Days 0, 3 and 7 respectively.
Interference Studies--The model compound conentrations were generated
with the portable permeation/dilution system•which has been described in
earlier reports. The organic vapor concentrations were all less than 100
ppb. Interference gases were generated at two different levels. The high
level concentrations' were 360 ppb ozone, 360 ppb nitrogen dioxide, 200 ppb
sulfur dioxide, and 90% relative humidity at 20°C. The low level interference
concentrations were 75 ppb ozone, 100 ppb nitrogen dioxide, 10 ppb sulfur
dioxide, and 26% relative humidity at 20°C. Clean air was humidified by
passing it thorugh a fritted bubbler containing water and then irradiated
with ultraviolet light to generate ozone. The ozone/water vapor in air
mixture was mixed with nitric oxide and sulfur dioxide from certified gas
cylinders to generate the required concentrations of each component. The
entire interferent gas mixture was mixed with the model compounds in the
clean air stream within a mixing bulb contained in the permeation/dilution
system oven which was heated to approximately 200°C.
The relative humidity was monitored with a YSI Dew Point Sensor. Ozone
and nitrogen dioxide concentrations were monitored with a Bendix ozone
monitor. Ozone concentrations were measured directly whereas nitrogen
dioxide concentrations were measured by the difference in ozone concentration
after gas phase titration. Sulfur dioxide concentrations were not monitored.
Chronology of Quality Control Practices
Steps taken to provide quality control and assurance of the data
generated was discussed above. Table 68 presents a summary chronological
record of QA/QC procedures performed during the storage-stability and inter-
ference studies.
209
-------
Table 68. CHRONOLOGICAL RECORD OF QUALITY CONTROL AND QUALITY ASSURANCE PROCEDURES
PERFORMED DURING STORAGE-STABILITY AND INTERFERENCE STUDIES
Experiment
Dates
Quality Control and Assurance Performed
ho
I—I
o
Storage-Stability
(High Level)
(Group I Compounds)
(Sorbents and Traps)
Storage-Stability
(Low Level)
(Group I Compounds)
(Sorbents and Traps)
1/25/80
2/5/80
2/20/80
3/30/80
3/31/80
4/16/80
4/22/80
4/25/80
5/8/80
5/12/80
5/20/80
6/18/80
Calibration of TD/HRGC System (Varian Model 3700, CDS-111
Chromatography Data System)
Calibration of GC/FID (Fisher Victoreen Model 4400)
Recalibration/Verification of TD/HRGC System
Verification of Flow Meters and Controllers on Permeation/
Dilution System
High Level Storage-Stability Study Initiated
Calibration of GC/ECD (Fisher Victoreen)
Verification of TD/HRGC Calibration Using Independent Method
(Flash Unit)
Verification of TD/HRGC Calibration Using Independent Method
(NBS certified gases)
Low Level Storage-Stability Study Initiated
Breakthrough Volumes for Chloroprene, Furan and Acrylonitrile
Determined for Tenax GC Cartridge
Verification of Flow Meter Calibrations for Permeation/
Dilutor System
Q
Low Level Storage-Stability Study Continued
(continued)
-------
Table 68 (cont'd.)
Experiment
Dates
Quality Control and Assurance Performed
Interference Study
(High Level)
(Group I Compounds)
(Sorbents and Traps)
Interference Study
(Low Level)
(Group 1 Compounds)
(Sorbents and Traps)
7/2/80 • Verification of TD/FRGC Calibration Using Independent Method
(Flash Unit)
7/7/80 • Verification of TD/HRGC Calibration Using Independent Method
(NBS certified gases)
7/11/80 • Analysis of Chloroprene/Xylene Mixture (of Permeation Tube)
by GC/MS to Verify Chloroprene Concentration
8/6/80 • Calibration of NO Monitor
8/6/80 • Calibration of 0 Monitor
8/6/80 • Verification of Flow Meter Calibrations for Permeation/
Dilution System
8/11/80 • Interference Study Inititated (Tenax GC Traps)3
8/11/80 • Identification of Background on Exposed Tenax GC and Unexposed
Charcoal Traps by GC/MS
9/6/80 • Verification of NO Monitor Calibration
9/6/80 • Verification of 0., Monitor Calibration
9/6/80 • Verification of Flows on Permeation/Dilution System
9/12/80 • Interference Study Continued (Charcoal and Cryogenic Traps)
9/12/80 • Calibration of GC/ECD
9/22/80 • Calibration of NO Monitor
X
9/22/80
Calibration of 0» Monitor
(continued)
-------
Table 68 (cont'd.)
Experiment
Dates
Quality Control and Assurance Performed
t-0
9/25/80 • Interference Study Initiated (Tenax GC and Charcoal Traps)'
10/6/80 • Verification of NO Monitor Calibration
10/6/80 • Verification of 0 Monitor Calibration
10/6/80 • Verification of Flows on Permeation/Dilution System
10/8/80 • Interference Study Continued (Cryogenic Traps)
10/16/80 • Verification of TD/HRGC Calibrations
10/16/80 • Verification of GC/ECD Calibrations
11/6/80 • Audit of NO and 0 Monitor Calibrations
" A J
5/19/81 • Calibration of Thermal Desorption - High Resolution
GC/FID System for Tenax GC Cartridge Analysis
5/27/81 • Recalibration of TD/HRGC System
6/2/81 • Purity Check of Propylene Oxide
6/4/81 • Verification of Calibration of TD/HRGC System (Flash
Loading Technique)
6/5/81 • Calibration of Nutech Sampling Pumps
6/8/81 • Storage Study on Group II Compounds
6/24/81 • Calibration of TD/HRGC System for Methyl Chloride and
Methyl mercaptan
Storage-Stability
(High Level)
(Group II Compouds)
(Sorbents and Traps)
(continued)
-------
Table 68 (cont'd.)
Experiment
Dates
Quality Control and Assurance Performed
Interference Study
(High and Low Level)
(Group II Compounds)
(Sorbents and Traps)
Storage-Stability
(High Level)
(Group I Compounds)
(Containers)
7/3/81
7/15/81
7/17/81
8/3/81
8/25/81
.8/26/81
4/28/80
5/27/80
5/28/80
10/15/80
12/3/80
1/2/81
1/6/81.
1/8/81
2/4/81
Background Check of Permeation/Dilution System
High Level Interference Study
Low Level Interference Study
Calibration Check of Group II Compounds on TD/HRGC System
Calibration Verification for TD/HRGC System by Injection
of Liquid Standards
Calibration Check for TD/HRGC System
Initial Calibration of GC/FID (Perkin Elmer 3920)
Improved temperature control of permeation/dilution system.
Permeation system flow controls evaluated and calibrated.
Checked and recalibrated permeation/dilution system flow
controls.
Tested new SCOT SE-30 column in order to improve resolution.
House clean air generation system replaced in order to
improve moisture removal and increase output pres-
sure.
Storage stability study initiated .
Checked permeation/dilution system flows.
Recalibration of GC/FID
(continued)
-------
Table 68 (cont'd.)
Experiment
Storage-Stability
(Low-Level)
(Group I Compounds)
(Containers)
Interference Study
(High Level)
(Group I Compounds)
(Containers)
Storage-Stability
(High Level)
(Group II Compounds)
(Containers)
Interference Study
(High Level)
(Group II Compounds)
(Containers)
Dates
6/12/80
6/18/80
6/23/80
7/30/81
7/28/81
8/4/81
6/19/81
6/24/81
6/26/81
6/8/81
7/1/81
7/8/81
7/15/81
7/17/81
7/14/81
Quality Control and Assurance Performed
• Containers cleaned
• Calibration of GC/FID (Perkin Elmer 3920)
• Recalibration of GC/FID. Low level study discontinued
because of inadequate detection limit with volume of
sample taken from container.
• Initiated measurements with low-level interferents.
• Initiated measurements with high-level interferents.
• Calibration of GC/FID.
• Calibration of GC/FID.
• Calibration check
• Calibration check
• Initiated storage-stability study.
• Calibration check.
• Calibration check.
• Initiated measurements with high-level interferents.
• Initiated measurements with low-level interferents.
• Calibration check (extension of calibration check from
Group II storage-stability study).
(continued)
-------
Table 68 (cont'd.)
Experiment
Bag Contamination
Study
Dates
7/28/80
7/28/80
7/29/80-
8/9/80
9/9/80-
9/23/80
9/29/80
11/3/80
Quality Control and Assurance Performed
• Initiated study of sources of bag contamination
• Bags leak-tested
• Bag loading, storage and analysis experiments
• Repeated bag loading; storage and analysis experiments
• Initiated evaluated of bag cleaning methods
• Continued bag loading, storage and analysis experiments
With the initiation and continuation of each set of experiments "background" controls were taken of
the permeation/dilution system in addition to the normal trap blanks.
-------
SECTION 8
DESIGN AND FABRICATION OF AN AUTOMATIC SAMPLER
DESIGN FEATURES
Efforts were made to design, a sampler that would be reasonably inex-
pensive but which adequately satisfied the most important requirements of
collecting air samples on sorbent cartridges. A summary of the design
features are as follows:
(1) The design of the sampling heads was to accommodate 12 samples
with six in each sampling head and with provisions for one blank
in the center position.
(2) A sampling head with bolt-on cap to be quickly accessible.
(3) Sample head mounting in a down position with a cover and slipjoint
hinge.
(4) Twelve 1/8 inch lines to the console approximately 10 feet long.
(5) Provisions for shipping cartridges in sampler head.
(6) Electrical and mechanical switching and flow control on front
panel of console.
(7) Variable orifices for low flow rate settings and range.
(8) Manual flow adjustment of each channel through variable orifices
(valve).
(9) A mass flow meter which is switchable to each channel with the
ability to obtain total flow.
(10) The capability to check the mass flow meter or calibrate.
(11) Solid-state times preferred over electromechanical with also
provisions for a clock integrator, printer, and manual mode.
(12) Seventy-two hour maximum sample time for six samples; 24 hours
maximum single sample with provisions also for automatic or manual
setting available for any time period.
216
-------
(13) Reset capability for flow integrator after each sample but not
after power failure.
(14) Options on the console for series (1-12) samples and duplicate
parallel sampling.
(15) Sampling rates settable from 7 mL/min to 1.5 L/min.
(16) The automatic sampler was to operate on alternating current.
MECHANICAL DESIGN
Figure 60 depicts a schematic of the automatic sampler and its speci-
fications are given in Table 69. It consists of two sampling heads which
allows for duplicate sampling. Each sampling head housing six sampling
cartridges plus one blank. Figure 61 depicts the control panel for the
automatic sampler and Table 70 lists the control settings. The major compo-
nents shown are a flow meter (the flow may be monitored through either
sampling head independently or in combination), and a printer (records the
total volume of air sample/unit of time). The sampling periods available
are of 15, 30, 45 and 60 min and 1/2, 2, 3, 6, 8 and 12 hr. Duplicate or
single cartridge sampling is possible (in a serial fashion for collection of
up to 12 sampling cartridges) for a maximum of 12 hrs/sampling cartridge.
Flow control for each sampling head is achieved with a variable orifice.
Figures 62 and 63 depict the flow measurement processes and the console
flow diagram for the automatic sampler.
Figures 64-66 depict the schematic of a sample cartridge collector.
The cartridge collector is constructed of aluminum with accommodation of up
to 7 cartridges (Fig. 67), of which 6 would be used for sampling and the 7th
a blank. Figure 68 gives the sample holder set in a sample cover. Figure
69 depicts the heated sample cover for the sample cartridge collector.
FLOW DIAGRAM AND ELECTRONIC CONTROLS
The overall flow diagram was shown in Figure 63. An expanded view of
the manifold is given in Figure 70 and of the muffler in Figure 71.
The control and timing functions of the sampler are shown in the electro-
nic schematic (Figs. 72 and 73). The timing circuit generates a series of
pulses at a frequency of 100 KHz (Fig. 73). This frequency is reduced by
means of divide counters to the desired sampling frequency, ^.e., 15 minutes
to 24 hours. The time cycle switches on the front panel pack up the time
217
-------
Sampling Head
Sampling
Head Block
6 Gas Flow Lines
Heater Supply Line
Thermocouple Wire
Tripod
Main Control Unit
Figure 60. Automatic sampler.
218
-------
Table 69. SPECIFICATIONS OF THE AUTOMATIC AIR SAMPLER
Item
Specification
Input voltage
Fuse
Flowmeter (Model AFSC 500):
minimum flow rate
maximum flow rate
accuracy
operating temperatures
response time
Flow integrator
minimum volume
maximum volume
Maximum number of samples
Maximum number of controls
Total number of cartridges
Sampling Periods
Number of sampling periods
per sampling sequence
Printer paper
120 volts AC, grounded
3 amp, 250 volt
0 seem
500 seem
+1% @ maximum flow rate
0-40°C
25 sees to 90% of reading
0.01 liters
9999 liters
12
2
14
15, 30, and 45 rain; 1, 1.5,
2, 3, 4, 6, 8, 12, and
24 hrs
6 (parallel mode); 12
(serial mode)
electrosensitive (100 ft
rolls)
United Systems Corp. P/N
19-17210, or Radio
Shack P/N 26-1412
219
-------
Not Shown:
Sampling Head
Block Heater
Controls
Printer.
Printer.
On/Off
Calibration
Ports
Pump
On/Off/Auto
Quick Connects
to Sampling Head
Not Shown:
Sampling Head Block
Thermocouple Outlets
Sampling
Period
Selector
Reset Switches
Reset
Step
Serial or
Parallel Mode
Indicator Lamps
Run-Ready/RST
Channel
iolenoid
Switches
Flow Rate Readout
(mL/min)
Main Power Switch
Flow Control Valves
Heater Power Plug
for Sampling Head
Block
Cooling Fan Vents
Figure 61. Main control unit.
220
-------
Table 70. SWITCH POSITIONS AND CONTROL SETTINGS AT
POWER UP OF AUTOMATIC SAMPLER
Switch or Control
Position or Setting
Sampling Period Selector
Printer On/Off
Head Heater Controls
Serial/Parallel Switch
9 Sec/Prog. Switch
Run/Ready-Reset Switch
Pump
Channel Solenoid Switches
Flow Control Valves
Any
Off
Lowest Setting (full CCW)
Any
Non-functional
Run
Off
Up
Any Setting
221
-------
zero
adj.
gain
adj..
DIGITAL .
DISPLAY
(Flow rate)
MASS FLOWMETER
INTEGRATOR
PRINTER
(Flow volume]
Figure 62. Functional diagram of flow measurement processes.
222
-------
7-Port Manifold
B—
Calibrate A
7-Port Manifold
Calibrate B
Valve A (VO) (VO) Valve B
So" A x^S /^SSoJA
Flow
Sensor
|! Check Valve
4-Port
Manifold
Quick Connects (12)
'12
Figure 63. Flow schematic of the main control unit.
223
-------
Bottom Retainer Plate
Teflon Restrict or
Teflon Spacer
Glass Fiber Filter
Viton O-Rings
(Size 114)
Sampling Head Block
(Aluminum)
Viton O-Rings
(Size 010)
'Teflon Spacer
Top Retainer Plate
1/8" Brass Swagelok Unions
1/8" Tubing
Studs and heater wells
not shown for simplicity
Figure 64. Exploded view of sampling head block for sampling mode
(upside down).
224
-------
Bottom Cap Plate
Teflon Restrictor
Teflon Spacer
Viton O-Rings
(Size 114}
Sampling Head Block
Viton O-Rings
(Size 010)
•Teflon Spacer
Top Cap Plate
Studs and heater wells
not shown for simplicity
Figure 65. Exploded view of sampling head block for transportation/
storage mode (upside down).
225
-------
Threaded Stud
Indentation for O-Ring
Port for Tenax Cartridge
Thermocouple Well
Heater Cartridge Wells
Figure 66. View of top of sample head block - approximately to scale.
226
-------
16 mm O.D.
Glass Wool
Glass Wool
10.1-10.2 cm
(10.15 nominal)
1 mm Wall Thickness of Glass Tubing
%" O.D.
Figure 67. Tenax cartridges for automatic sampler.
227
-------
SAMPLE HOLDER
V8" TFE (6)
H H
TEMPERATURE AT
BLOCK SET AT
30°C
Figure 68. Sample holder.
228
-------
-HINGED WEATHER COWER
Figure 69. Heated sample cover for sample cartridge collector.
-------
•1,"
-
Mat!: stainless steel
Figure 70. 4-Port manifold.
230
-------
FELT
-1/4" NPT
?
•GASKET
• ALUMINUM
Figure 71. Muffler.
231
-------
OJ
110 VKC
YSI, VKLVl SEQUENCE P/WEl WD4CAIOBS
5 VOC
lOUC
Figure 72, Auto sampler electronic schematic.
-------
N3
CO
00
15 min. 30 min. 1 nr.
1 1
V/, hr.
1
3 hr.
1 J
6 hr.
I 1
12 hr.
1 1
24 hr.
Figure 73. Time pulse generator.
-------
pulse at various points along the pulse dividing network. The time pulse
initiates the Print Routine Generator, advances the Valve Cycle Controller
and resets the Total Flow Integrator.
The Total Flow Rate Indicator (Fig. 74) consists of a voltage to frequency
converter with an output of 1000 pulses per second per volt input. The
calibration of the flow sensor will typically be 11 amp per volt. Based on
this ratio the pulse output is divided by 600 to give an output of 0.01
liter in the least significant and 1000.00 liter in the most significant
digit. These six digits are fed into the print buffer. In addition the
flow rate senor dives a panel meter to indicate the instantaneous flow rate.
The Valve Cycle Controller (Fig. 75) is basically a shift register
(Fig. 76) which advances one position with each time pulse. It is arranged
in two units corresponding to the duplicate channels A and B so that they
can be advanced in parallel (with one line on each channel open) or in
series (with the first line of channel B following the completion of sampling
on the last line of channel A). Upon receiving the signal from the shift
register (Fig. 76), relay drivers provide power to solenoid valves controlling
flow thru appropriate sample lines. An open valve is indicated by LED on
the front panel .and also transferred in BCD format to the print buffer (Fig.
77).
At each time pulse a print routine is initiated. The printer prints
time of day and date followed by the information in the print buffer, i.-e^- ,
sample volume and sample lines being sampled. At the conclusion of sampling
the Valve Cycle Controller deactivates the pumping system and turns the
sampler off. Remaining pertinent circuits are depicted in Figures 78-85.
The fabrication of the automatic sampler also included the following
additional features:
(1) A cooling fan was installed behind the pump to reduce potential
overheating.
(2) A manual channel stepping switch and 9 sec stepping function was
installed on the front panel. This feature aids the checking of
all solenoid valve switching operations and instrument flow calibra-
tion.
234
-------
(-0
01
.OOlitn
Ifcm/V
1
tte'V
Stai
= 60(1(1
.01 litw = 600 p
ANALOG DEVICES
VOLTAGE TO FREQUENCY
MO
>fi
GENERATOR
FLOW o w^-
SENSOR V,, R,
OUTPUT O
ANALOG
GROUND
455 j
-01
3 o—
40—
5
6
70—
1—0*15 V
^"? 0
V
Vrr
S
7490
DIGITAL
GROUND
"cc
|s
vcc
3 1,4
7490
06
II!
1
-J 3
r
6 12
+ 10
VCC
3 M4
7490
OS
his
s
7490
I
?
1
V0
3
14
7490
04
In
3
C
S 3
14
7490
03
|
+ 10
1
— y
S
1
7492
INPUT
BC
»cc
11 5
3
7490
3
14
02
+J
V
1"
X
5
= 60p
= 8p
= 1 P
3
7490
4
II.
>TJr TO PRINT BUFFER MAX
-------
Panel LED
N>
W
ON
Power Relay
w n **
| 1—|
| B2-
1
Ready
Channel B
3
I 1
" «
11
Encoder
4
r
~
J,
~l
i
g
-|
1
1 |
Vcc-t
L
sw
Parallel
19
1 CLR
74164
2 14
_J "
10
I
3
«
te. Is
2
13
2
13
13
75417
75417
75417
i
i
;
!
V7
V9
V*
VID
V12
Relay Driver
Figure 75. Valve sequencer.
-------
To register B
or terminate
cc
input flip flop
|
H ^1 F
VCC QS 0
A B
IN IN G
^-QJ-2-p
1 1
VCC
/
\ j
1 I
i i
*i i"i F°I r*i
7 Qg 05 Clear (
1 Q2 03 04
Ll LJ UJ UJ
• !
I
^ A x
"I
:iock
GND
7J
Time Pulse
To To
valve encoder
relay
driver
Figure 76. 74164 8-Bit parallel-out serial shift register.
237
-------
CHMACTHIS
f, i, • i, t, is i, » « », SP A
PC,
1
0
0
1
1
0 I
0 1
) 0
1
0
0
0
0
0
1
0
0
0
0
1
0
0 1
1 0
D 0
0 0
0 0
0 0
0 1
0
0
0
1
0
1
0
COMPONEKTS
M.
DESQUPTNM
1312 64WUT MUITBT.EXERS
54/74 257 (UJAO 2 UK TO t LME MUITIPUXER
64/74 9SB 4«T PARALLEL ACCESS SHIFT REGISTER
5*74 32 (MAO HOW OR GATE
Mf?4 21 DUAL WIPUT AID GATE
Figure 77. Printer interface buffer.
238
-------
Perf board, 0.1"
110VAC
3V4"
Figure 78. Flow sensor power supply.
239
-------
6 in
6 out
'cC'
14
13
R ra m Hi
Vcc 6A
1A
6Y 5A 5Y 4A 4Y
1Y 2A 2Y 3A 3Y GND
TiJTiJ -lil LU Lh
A — line in
Y — line out
Figure 79. 7404 Hex inverter.
240
-------
V,
cc
Ac1
I Serial/parallel switch
H M
K Q
CLK PR
N liJ
14 J13J I |l2J |l1 | |lo| 9
Q GND !
1
1
1
1
1
1
1
1
1
1
1
1
1
CLR J 1
i
N Hi 111 H N 3
output
from
74164
V
cc
1
I Ready switch
Figure 80. 7476 Dual JK flip flop with preset and clear.
241
-------
SOUS
1 *i h
2.4 to 5 V
* COUNT IN
pi] fl3| fl
ri ^
2] |n| 10
9
8
A NC A D GND B C
IN
BD
IN Ro(1) R0(2) NC Vcc Rg(1) Rg(2)
N N LJ
1 Res
on
•J liJ 5
et
1
6
i
7
+ 10 out
Figure 81. 7490 Decade counter as divided by 10;
242
-------
Count in »•
-^•Next Decade
R R R
A IMC A
IN
D GND B C
BD
IN R0(1) «o(2)
Vcc Rg(1) Rg(2)
T
6
T
1 Reset
on 1
Possible alternate 8285
Figure 82. 7490 Decade counter in cascade.
243
-------
v,
Input lines
cc
R R R R
5 6
11
Vcc or GND
\
H
'cc
E0 GS 3 2
1 0 A
4 5 6 7
A2
GIMD
liJ LiJ LiJ Id
\i
2 1
Input lines
A3 A2
Figure 83. 74148 8-Line to 3-line encoder.
244
-------
Q2 from 74164-
Valve #2
Sol. V1
114 I 113
9 8
2A
2Y Clamp
S 1A
1Y Gnd
N N HI liJ liJ
vcc
(for low output
to 1A)
Value =1
Q1 from 74164
Figure 84. ' 75417 Dual peripheral driver.
245
-------
LED Panel
a-
L_M_J-LJ"LJLJLJ
C
•i n rr i-
74123
=)
D1P 1 D1P2 D1P3
Valve Sequencer
Printer Interface/
Flow Rate Integrator
Zero
-Gain
(Not occupied)
Time Generator
Front Panel
Bottom
Figure 85. Arrangement of P.C. boards on edgecard frame (looking
into connectors).
246
-------
(4) The sampler was equipped with two calibration ports with 0.25 in
quick-connect assess. They were tied into the flow at each channel
before the manual flow control valves. This feature facilitates
presetting the flow rates as well as calibrating the flow meter.
and (5) Addition of noise suppression steps to improve the timing and valve
realiability.
247
-------
SECTION 9
PERFORMANCE OF AUTOMATIC SAMPLER
EVALUATION OF STEPPING SEQUENCER AND SAMPLER CALIBRATION
The Nutech automatic sampler was tested in both parallel and serial
modes to determine whether the sampler was stepping through each channel in
the proper order. The sampler was tested for sampling periods at 15 min, 30
min, 45 min, 1 hr, 2 hr, 12, 24 hr and was found to be stepping through the
channels correctly at the proper time intervals.
The samplers digital flow meter was calibrated to read the actual flow
rate being sampled by adjusting zero and gain controls to the digital flow
meter. The sampler was then tested to determine if all selonoids were
functioning properly to allow for sampling through the 12 sampling ports.
Tests revealed all sampling ports were functioning properly. The sampler
was then tested to ensure that the flow set on the digital flow meter
corresponded to the actual flow (as measured by an NBS bubble meter). The
results of these tests are given in Tables 71 and 72. These experiments
revealed that the sampler set flow, and actual flow agree within 2%. The
integrator was then checked to see if the data printout agreed with the
actual volume being sampled. Initially, the integration was found to deviate
20% in the serial mode and 3-5% in the parallel mode. Table 73 lists the
calibration results. Table 74 gives the comparison of set point and actual
flow vs_ the dynamic range.
The sampler was then also tested to reveal if the digital flow meter
would drift from the actual flow rate being sampled over long periods of
time. The test revealed that no significant drift was present (Tables 75
and 76).
EVALUATION OF SAMPLING HEADS
Initial experiments included determining the sealing capabilities of
two prototype sampling heads. One head utilizes a phenolic screw cap to
248
-------
Table 71. FLOW METER CALIBRATION OF AUTOMATIC SAMPLER -
SERIAL/MANUAL MODE
Port No.
1
2
3
4
5
6
% Dev. from
Channel
Set Point
406
406
406
406
406
406
Mean
S.D.
% S.D.
Set Point
A
Actual
398.9
399.6
400.7
400.0
398.7
400.0
399.6
0.75
0.19
1.56
Channel B
Set Point
421
421
421
421
421
421
Mean
S.D.
I S.D.
Actual
418.5
418.5
417.4
418.8
419.1
417.7
418 '.2
0.80
0.19
0.66
Checked with NBS bubble meter, mL/min.
249
-------
Table 72. FLOW METER CALIBRATION OF AUTOMATIC SAMPLER -
PARALLEL/MANUAL MODE
Port No.
1
2
3
4
5
6
Mean
S.D.
%S.D.
Channel
Set Point
301
301
301
301
301
301
% Dev. from Set
Point
A
Actual
295.2
296.0
295.9
301.5
294.9
293.3
296.1
2.80
0.95
1.62
B
Set Point Actual
306 295.4
306 296.3
306 298.4
306 295.2
306 303.1
306 309.9
299.7
5.79
1.93
2.05
A + B
Set Point Actual
587 590.6
587 592.3
579 594.3
586 596.7
595 598.0
599 603.2
595.8
4.52
0.76
1.12
Checked by NBS bubble meter, mL/min.
250
-------
Table 73. FLOW INTEGRATOR CALIBRATION OF AUTOMATIC SAMPLER -
PARALLEL/MANUAL MODE
Port No.
1
2
3
4
5
6
Channel
Set Point
587
587
579
586.
595
599
A + B
Actual
590.6
592.3
594.3
596.7
598.0
603.2
Integrator Set Point
Total (£) Total (£)
-
2.88
2.87
2.88
2.95
2.97
-
2.94
2.90
2.93
2.98
2.99
Actual
Total (£)
-
2.96
2.97
2.98
2.99
3.02
Checked with.NBS, bubble meter, mL/min.
251
-------
Table 74. CALIBRATION OF DYNAMIC RANGE FOR FLOW METER
ON AUTOMATIC SAMPLER
Set
8
38
98
210
318
417
510
607
Channel A
(8/8/80)
Point Actual
9
38
93
209
314
410
507
599
Channel A
(8/19/80)
Set Point Actual
32
94
179
294
394
496
592
678
33
90
175
292
391
493
586
675
Channel B
(8/19/80)
Set Point Actual
55
94
195
316
421
506
616
716
54
91
193
315
418
507
620
722
252
-------
Table 75. AUTOMATIC SAMPLER OUTPUT-PARALLEL MODE
15 min
Time Period
30 min
1 hr
6ft 7B 8003.19
89/11 10:04
5fl 7B 0063.19
89/11 09.'49
4Fi 7B 8663.19
09/11 09.'34
3ft 7B 0663.21
09/11 89: 19
2fi 7B 6083.25
09/11 69:04
](R 7B 0005.01
89/11 88M9
6R &B 6012.47
09/11 13:13
5fi 5B 8612.51
09/11 12:43
4fi 4B 0612.32
09/11 12:13
3ft 3B 0312.35
09/11 1K43
2R 2B 06 12. 46
•09/n 11:13
Ifl IB 0612.46
09/H 18M3
6R 6B 6625.62
6?'/ll 23ISv
5fi 5B 6S25.C7
89/ii 22: ee
4F. 4& 6824.67
69/11 21165
3ri 3B 8624. tfc
89/ii 2e: ee
2fi 2B 662^.77
65/11 19:30
1R IB SB24.7v
09/n is: ee
Reproduced from
best available copy.
253
-------
Table 76. AUTOMATIC SAMPLER OUTPUT-SERIAL MODE
Time Period
15 min
7ft 65 6663.26
69/68 17:84
7ft 5B 6663.20
85/83 16M9
7fl 4B 8663.15
69/08 16Z34
7ft 3B 0003.26
69/08 16: 19
7ft 2B 0663.26
09/08 16:04
7fi IB 6863.23
69/65 15:49
6R 7B 6083.17
09/08 15:34
5fl 7B DG83.17
09/68 15: 19
4ft 7B 6883.17
89/08 15:04
3fi 7B 0883.17
89/08 14:49
2R 7B 6863.17
69/03 14:34
Ifl 7B 8603.54
89/68 14.' 19
30 min
7fi 6B 6666.19
69/69 IS: 45
7fi 5B 6606.15
69/69 IS.' 15
7fl 4B 6686.18
69/69 17:45
7fi 3B 6666.19
09/69 17.' 15
7fi 2B 6666.28
69/69 16.M5
7fl IB 6666.24
69/69 16: 15
6fi 7B 6666.34
69/89 15:45
5ft 7B 8066.34
69/09 15: 15
4fl 7B 0086.35
65/69 14:45
3fl 7B 6686.36
69/69 14: 15
2ft 7B 6686.37
69/09 13M5
Ifl 7B 6832.87
69/89 13: 15
30 min
7fi 6B 6636.20
69/10 63.'42
7ft 5B 6886.21
09/16 63: 12
7ft 4B 6666.21
69/18 62:42
7fi 3B 6666.21
69/16 62: 12
7fi 2B 6666.21
69/16 0i:42
7fl IB 6606.25
69/18 0i:i2
6ft 7B 6666.37
69/16 88:42
5ft 7B 6666.37
89/16 68:i2
4R 7B 6886.37
"85/63 23.'42
3ft 7B 6806.37
89/89 23:12
2fi 7B 6686.38
69/09 22:42
Ifl 7B 6886.68
69/09 22:12
Reproduced
1 hr
7Fi 6B 8S12.31
69/18 28:31
7n 5£ 8512. 2J
65/13 19:53
7fi 4B 6812.26
69/18 IS: 55
7fi 3B 6812.26
69/16 17:55
7fi 2fc 8622.27
89/16 16:53
7ft IB ££12.31
83/18 13.' -3
6ft 7B 6812.61
69/16 14155
5fi 7E 8312.63
69/18 13:55
4ft 7B 8812. £;•
0r, • < r. • "•• ' C —
y •• A I' i i . •-• •-•
3fi 7B 8cl2.£i-
85/16 i::35
2ft 7B 6812.76
89/18 18:55
1ft 7B 6812.95
09/18 89155
from JM%
best available copy. %fl^
254
-------
seal the Teflon spacers to the sampling cartridges while the other sampling
head design utilized a bolt-type arrangement.
Pressure Tests
Pressure leak tests were conducted with both sampling head designs.
The heads were sealed and a pressure source from a He tank reservoir was
applied. In line was a flow meter and pressure gauge for monitoring leaks.
A 10 psi pressure was applied and then turned off from the source, the
ability to hold the applied pressure was monitored on the pressure gauge.
Both types of heads (screw-cap and bolt-type) were found to leak pro-
fusely in these initial tests. Table 77 gives this leakage rate. The
@
leakage in both cases was attributed to the noncompressability of the Teflon
spacer between the block and end-plates. Eventhough the bolt-on end plate
could be tightly secured, warping of the end-plate aggravated the leakage
problem.
©
Thus the Teflon spacers were omitted at both ends of the sampling
head, permitting the Viton o-rings (size 114-Ace Lab Glass and size 010-
Cajon) to seal around each end of the sampling cartridge and the end plate.
Pressure leak tests to 10 psi with He revealed no leaks on the bolt type and
screw-cap heads. A helium leak detector set at "high" sensitivity was. used
to monitor around all fittings.
®
Background of Stored Tenax GC Cartridges
The initial background studies began with determining whether clean
Tenax cartridges placed in the sampling head could be maintained background
free during periods of storage. The initial experiment employed 24 hr
storage and then subsequent experiments were 45 and 95 hrs long.
Tenax cartridges examined throughout the storage study were subjected
to TD/HRGC analysis with the integration of the background area using a CDS
111 Varian Chromatography Data System. The chromatographic parameters used
in the storage study are given in Table 78.
Table 79 presents the results obtained for the background level for
Tenax cartridges stored in the prototype sampling heads. In general, the
background increased above that observed for a cartridge stored for the same
period of time in a Kimax culture tube. Best results were obtained with
the'phenolic screw-cap utilizing Teflon O-rings for a storage period of 41
hours.
255
-------
Table 77. INITIAL LEAK TEST RESULTS
Sampler Head
Type
Screw-cap
Bolt-on
Delivered Pressure
(psi)
10
10
Final Pressure
(psi)
2
2
Leak Rate
(L/min)
3
2.5
Measured by regulator at tank source.
Measured by in-line pressure gauge.
Measured by in-line rotameter, max. initial rate achieved.
256
-------
Table 78. CHROMATOGRAPHIC PARAMETERS USED IN STORAGE STUDY
Parameter
Conditions
Column Temperature
Capillary
Injector Temperature
Detector Temperature
Attenuation (AUFS)
GC
Data System
40°C (2 min) 4°C/mil1 »210°C (2 min)
SE-30 WCOT/BaC03; 85 m
220°C
270°C
X64, 10"11
Varian 3700
CDS-111
257
-------
00
Table 79. BACKGROUND LEVEL STUDY FOR TENAX® GC CARTRIDGES STORED IN PROTOTYPE SAMPLING HEADS
USED FOR THE AUTOMATIC SAMPLER
Bolt
Teflon
Storage:
Initial Control
Final
Sample
Control
: Mean
S.D.
C.V.
45
6
5
28
9
34
hr
.21
.82
.42
.68
.06
Type/
0-Rings
91
6.
6.
68.
18.
27.
Screw Cap/
Teflon 0-Rings
hr
43
17
09
73
52.
45
17
15
27
6
23
hr
.96
.10
.35
.48
.69
95 hr
2.62
1.83
108.55
13.81
12.73
42
1
2
56
18
32
Screw
Viton
hr
.87
.74
.26
.43
.75
Cap/
0-Rings
Area counts of cartridge background determined immediately after preparation.
b ®
Area counts of cartridge background stored in Kimax culture tube sample period of time as
"sample".
j-«
Cartridge stored in sampling head of automatic sampler, mean of 3 determinations.
-------
Additional research was conducted to determine the merits of a screw-
cap vs. bolted head for sealing cartridges. Cleaned and vacuum pumped Viton
0-rings were used. The background study was conducted over a 2 week period.
The results in Table 80 clearly indicates the ineffective sealing using the
screw-cap type closure. Bolt-on end plates preserved the sampling cartridges
as well as culture tubes.
With the subsequent omission of Teflon spacers in the sampling heads
(see above) the background on Tenax GC cartridges was further investigated.
All components were cleaned as follows. The sampling heads and end plates
were washed twice in methanol (B&J) and once in pentane (B&J), air dired and
then placed in a vacuum oven for 2-4 h at 110°C. Viton o-rings washed in
warm Isoclean solution with sonication for 1 min, rinsed 3-6 times with
warm water and sonicated for 1 min, 3 times in distilled water, air dired
and placed in a vacuum oven for 2-4 h at 40°C.
Tenax GC cartridges were prepared as previously described (Section 6)
using glass tubes designed to fit the sampling head.
Storage experiments were conducted using both sampling head designs.
The background on each cartridge was determined by TD/HRGC as previously
described, with the area integrated under all the peaks occuring in the
chromatogram for a specified time. These results are given in Table 81.
These data indicate that the Tenax cartridges'stored in the bolt-type head
was similar to culture tube storage. However, when each cartridge was
removed from the sampling head, the remaining cartridges were potentially
exposed to atmospheric background thus confounding the subsequent results.
The storage experiment was repeated, but the heads were opened only once, 7
days after storage. These results are given in Table 82.
All of the experimental data indicates that the bolt-type sampling head
is a better design and thus it was incorporated into the automatic sampler.
259
-------
Table 80. BACKGROUND LEVELS OF TENAX CARTRIDGES STORED IN
SAMPLING HEADS COMPARED TO CARTRIDGES STORED IN CULTURE TUBES
Storage Vessel
Culture tubes
Sample head with bolt-on end caps
Sample head with screw-on end caps
Slope3
69
77
790
Correlation
Coefficient
0.744
0.672
0.936
•a
Based on y = mx + b where y is the total instrumental response, and
x is the time in days of storage. Therefore, the greater the slope,
the greater the level of contamination over time.
260
-------
Table 81. BACKGROUND OF TENAX GC CARTRIDGES STORED
IN SAMPLING HEADS
Storage
Mode
Culture tubes
Bolt-type head
Screw-cap head
Storage Time
(days)
0
2
7
14
0
2
7
14
0
2
7
14
15
Area
(Arbitrary Units)
0
0
3,823
862 + 182
0
1,696
1,772
1,288 + 162
0
3,397
4,593
14,732
10,128
Area was summed for all chromatographic peaks occurring in the 62
min of the run.
261
-------
Table 82. BACKGROUND OF TENAX® GC CARTRIDGES STORED IN
BOLT-TYPE SAMPLING HEAD
Storage
Mode
Culture tubes
Bolt-on Type Head
Storage Time
(days)
0
7
0
7
Area
(Arbitrary Units)
0
758 + 191
0
1,856 + 924
Area was summed for all chromatographic peaks occurring during the
first 12 min of the run.
262
-------
SECTION 10
PRELIMINARY DEVELOPMENT OF DIFFUSION TUBES FOR LOW VAPOR-PRESSURE COMPOUNDS
INTRODUCTION
In addition to the preparation of permeation tubes for selected test
compounds, a parallel effort included an investigation into a method for
synthesizing a flowing stream of air/vapor mixture of low vapor pressure
organics (b.p. > 215°C). There is no literature report for accomplishing
this.
DELINEATION OF A DIFFUSION TUBE SYSTEM
Chromatographic Method
A new approach was devised and tested for delivering constant levels of
polar and/or non-volatile compounds for the purpose of synthesizing air/vapor
mixtures. This method employed the use of vessels containing the organic
compound to which was attached a short chromatographic column (Fig. 86).
The concept examined was the use of a GC phase coated onto a support to give
controlled diffusion of vapor from the vessel when the container and.chroma-
tographic column are maintained at a constant temperature. The selection of
the GC phases were based upon the McReynold's No. and the polarity of diffus-
ing compound. Thus the rate of diffusion is controlled by temperature, •
phase loading, and length of chromatographic column. A model system which
would allow the control on a rate of diffusion to within a factor of 5 for a
group of model compounds with vapor pressures that are significantly apart
was desired.
To test this system diffusion tubes for those compounds listed in
Table 83 were prepared. For some compounds permeation tubes have been
successfully prepared; however, they served as cross-checks.
A series of chromatographic supports (Chromosorb-W, HP) coated with
varying percents of OV-17 (1,3,5 and 10%) in chromatographic columns-of
varying dimensions (1.25 mm, 2.5 mm i.d.) and packing bed depths of various
263
-------
Teflon Union
GC packing
Figure 86. Schematic of diffusion tube and GC column.
-------
Table 83. COMPOUNDS EXAMINED IN MODIFIED DIFFUSION TUBES
Compound
Pyridine
Phenol
Naphthalene
Anthracene
Quinoline
1-Naphthylamine
1 , 2-Dihydroxybenzene
N-Nitrosodiethylamine
GC Packing
OV-225
OV-225
OV-17
OV-17
OV-17
OV-225
OV-225
DECS
Column Dimensions
i.d. x length (mm)
(bed)
2.0 x 6.0
2.0 x 6.0
2.0 x 6.0
2.0 x 2.0
2.0 x 3.0
2.0 x 2.0
2.0 x 3.0
2.0 x 6.0
-------
lengths (0.5, 1.0 and 3.0 cm) was prepared and attached to a reservoir of
pyridine and phenol. The reservoir with the chromatographic column was
placed in an oven bath at 70 and 80°C with the exit of chromatographic
column attached to a manifold through which nitrogen gas was passed. An
injection port downstream of the manifold served for the withdrawal of a gas
aliquot for analysis by GC/FID to determine the concentration of pyridine
and phenol in the synthesized air/vapor mixture.
Examining the various variables of length, packing and phase loading,
the results indicated that some control on diffusion could be achieved to
generate different levels by selecting these parameters; however, a difference
of within an order of magnitude could not be achieved. It was desired that
the parameters would provide enough variability for model compounds of
varying vapor pressures and that they could be made to diffuse near the same
rates so that a synthetic air/vapor mixture might be synthesized as a multi-
component mixture at one particular bath temperature.
The encouraging aspect of this approach was the rather constant level
of compound which was sparged into the nitrogen gas stream after equilibration
had been achieved. The levels of the synthetic air/vapor mixture for pyridine
and phenol appeared to be controllable to +2% over several hours.
Gas-solid sorbents - Spherocarb, Tenax GC, Chromosorb 102, Porapak N
and XAD 2 - were also evaluated. Modified diffusion tubes (pyridine and
phenol) were placed in an oven bath at 80° with a 1.25 mm i.d. x 2.0 cm bed
length of each of the sorbents which led to a manifold whereby the air/vapor
mixture generated was examined by GC/FID. The equilibration to a constant
level of pyridine vapor was examined vs. time for each of the solid sorbents.
®-
The maximum difference observed between Tenax GC and Spherocarb was about a
factor of 2. This difference was deemed unacceptable for the work in this
project since the vapor pressures for the model compounds to be incorporated
into such a method are greater than an order of magnitude. Thus, constant
levels into a single synthetic air/vapor mixture stream were not possible
with this approach.
Non-Chromatographic Method
Another modified diffusion tube system was assembled as shown in
Figure 87. By equalizing the vapor pressures via selecting appropriate
266
-------
GLASS MANIFOLD
ho
ON
TEFLON TUBING
\
HELIUM PURGE
GAS
INSULATED TUBULAR HEATER
CONNECTED TO VARIAC (•)
SAMPLE OUT
INSULATED TUBULAR
HEATER CONNECTED
TO VARIAC (*)
VALCO ZERO DEAD
VOLUME SAMPLING
VALVE
SAMPLE LOOP
CARRIER GAS
TO COLUMN
•I THERMO-COUPLES MONITOR TEMPERATURE (°C)
HEATING TAPE IS WRAPPED AROUND
ALL TUBING IN THIS REGION TO
AVQID COLD SPOTS (*>
Figure 87. Modified diffusion tube system for generating vapor/gas mixture.
-------
temperatures on the diffusion tube, it was possible to deliver a constant
and relatively equilivalent value of each of the test model compounds in
question. The initial studies determined whether a constant level could be
maintained for a synthetic air/vapor mixture over a relatively long .period
of time (8 hrs) in order to be useful for the evaluation of collection
devices. By using a vapor pressure data and variable temperatures, the
equilibration times and constancy of delivery as well as accuracy of the
individual compounds were examined.
Initially a 1 mL sample loop was installed. However, broad chromato-
graphic peaks resulted on the capillary column and a 150 (JL sample loop and
a 5 sec injection time was then employed and proved successful. A minor
constructural modification was made to the diffusion tube system to eliminate
back pressure. A union T was placed in line with the transfer line to the
sampling loop to eliminate back pressure and the exhaust was connected to a
charcoal trap. Flows from the exhausts of the loop and vent lines were
measured with bubble devices to obtain total flow and calculated concentra-
tions of air/vapor mixtures.
Using the system shown in Figure 87 the temperatures on the diffusion
tubes were adjusted so than an equivalent amount of each compound was
delivered into the gas stream to provide a synthetic gas/vapor mixture.
This device will permit the replacement of a permeation system on the portable
diluter when model compounds with very low vapor pressures are used for
evaluating sampling devices.
The modified diffusio'n tube system was tested for its stability over 4
and 7 hour time periods. Peak heights were measured as the compounds were
held at the same operating temperature during this entire period. No attempt
to quantify the concentrations of the compounds by comparison with liquid
injection was made; however, future studies should be conducted to establish
equivalency.
Table 84 presents the results obtained for the stability of the diffusion
tube system for naphthalene analyzed over a 4 hr period. Indicated are the
operating parameters and the mean standard deviation and coefficient variation
of the peak height for this compound. Likewise, a 7 hr stability study for
naphthalene, quinoline and o_-chloronitrobenzene was also conducted and these
268
-------
Table 84. DIFFUSION TUBE STABILITY FOR NAPHTHALENE -
4 HOUR STUDY
Parameter
Set Point
Diffusion tube Temp.
Manifold Temp.
Transfer line Temp.
Valve Temp.
Manifold flow
Injection period
Injection volume
Detector Temp.
Column Temp.
Column
Capillary flow
200°C + 1°C
200°C
200°C
270°C
1.0 L/min
10 sec
150 (jL
300°C
160°C
Carbowax CP-Wax 20
(50 m x 0.5 mm I.D.)
3.4 mL/min
Mean
S.D.
% S.D.
Results
131.2 mm (peak height)'
6.5 mm
4.9
Represents 48 data points.
269
-------
results are shown in Table 85. In general the stability, i-e. the variation
as expressed by the percent relative standard deviation was small for the
time period investigated which implies that this technique may be suitable
for the synthesis of air/vapor mixtures of the group of compounds which
cannot be prepared in permeation tubes.
270
-------
Table 85. DIFFUSION TUBE STABILITY FOR NAPHTHALENE, QUINOLINE,
o-CHLORONITROBENZENE - 7 HOUR STUDY
Parameter
Set Point
Diffusion tube Temp.
Naphthalene
Quinoline
o-Chloronitrobenzene
Manifold Temp.
Transfer line
Valve Temp.
Sample Volume
Injection Time
Sample -Loop Flow
Tee Flow
Total Flow
Detector
Column
Capillary Flow
99°C
154°C
166°C
200°C
210°C
270°C
150 |JL
5 sec
24 mL/min
3,505 mL/min
3,529 mL/min
300°C
190°C
3.4 mL/min
Results
Naphthalene
Quinoline
o-Chloronitrobenzene
Mean + S.D. (C.V.)'
120.9 + 7.0 (5.8)
92.9 + 5.1 (5.4)
99.9 + 9.9 (9.9)
Represents 42 data points.
271
-------
REFERENCES
1. The National Research Council. Vapor-phase Organic Pollutants.
National Academy of Sciences. Washington, DC, 1976. 411 pp.
2. The National Research Council. Air Particulates. University Press.
Baltimore, MD, 1979. 343 pp.
3. Fishbein, L. Atmospheric Mutagens. In: Chemical Mutagens, Principles
and Methods for Their Detection, Vol. 4, Plenum, New York, 1976.
pp. 219-319.
4. Duce, R. A. Speculations on the Budget of Particulate and Vapor-phase
Nonmethane Organic Carbon in the Global Troposphere. Pure Appl. Geophys.,
116:244-273, 1978.
5. Bursey, J. T., D. Smith, J. Bunch, R. Williams, R. Berkley and E.
Pellizzari. Application of Capillary GC/MS Computer Techniques to
Identification and Quantification of Organic Components in Environmental
Samples. Am. Lab., Sept.:35-43, 1977.
6. Kaiser, R. E. Gas Chromatography in Environmental Analysis. J. Chroma-
tog. Sci., 12:36-39, 1974.
7. Pellizzari, E. D. Development of Method for Carcinogenic Vapor Analysis
in Ambient Atmospheres. EPA-650/2-74-121, U.S. Environmental Protection
Agency, Cincinnati, OH, 1974. 148 pp.
8. Pellizzari, E.D. Development of Analytical Techniques for Measuring
Ambient Atmospheric Carcinogenic Vapors. EPA-600/2-75-076, U.S. Environ-
mental Protection Agency, Cincinnati, OH, 1975. 186 pp.
9. Pellizarri, E.D. Analysis of Organic Air Pollutants by Gas Chromatog-
raphy and Mass Spectrometry. EPA 600/2-77-100, U.S. Environmental
Protection Agency, Cincinnati, OH, 1977. 104 pp.
10. Pellizzari, E.D. The Measurement of Carcinogenic Vapors in Ambient
Atmospheres. EPA 600/7-77-055, U.S. Environmental Protection Agency,
Cincinnati, OH, 1977. 228 pp.
272
-------
11. Pellizzari, E.D. Measurement of Carcinogenic Vapors in Ambient Atmos-
pheres. EPA 600/7-78-062, U.S. Environmental Protection Agency,
Cincinnati, OH, 1978. 224 pp.
12. Pellizzari, E. State-of-the-Art Analytical Techniques for Ambient
Vapor-phase Organics and Volatile Organics in Aqueous Samples from
Energy-related Activities. In: Application of Short-term Bioassays in
the Fractionation and Analysis of Complex Environmental Mixtures. EPA
600/9-78-027, U.S. Environmental Protection Agency, Cincinnati, OH,
1978. 227 pp..
13. Pellizzari, E.D. Analysis of Organic Air Pollutants by Gas Chromatog-
raphy and Mass Spectrometry. EPA 600/2-79-057, U.S. Environmental
Protection Agency, Cincinnati, OH, 1979. 229 pp.
14. Grimsrud, E.P., and R. A. Rasmussen. The Analysis of Chlorofluorocarbons
in the Troposphere by Gas Chromatography-Mass Spectrometry. Atmos.
Environ., 9:1010-1013, 1975.
15. Tyson, B. J. Chlorinated Hydrocarbons in the Atmosphere-Analysis at
the Parts Per Trillion Level by GC/MS. Anal. Lett., 8:807-813, 1975.
16. Seila, R. L., W. A. Lonneman and S. A. Meeks. Evaluation of Polyvinyl
Fluoride as a Container Material for Air Pollution Studies. J. Environ.
Sci. Health, All:121, 1976.
17. Kopczynski, S. L. W. A. Lonneman, F. D. Sutterfield and P. E. Barley.
Photochemistry of Atmospheric Samples in Los Angeles. Environ. Sci. &
Technology, 6:342, 1972.
18. Sickles, J. E. II, R. S. Wright, C. R. Sutcliffe, A. L. Blackard, and
D. P. Dayton: Smog Chamber Studies of the Reactivity of Volatile
Organic Compounds. EPA Contract No. 68-02-3352, Final Report, Res.
Triangle Park.
19. Lonneman, W. A., J. J. Bufalini, R. L. Kuntz and S. A. Meeks. Contami-
nation from Fluorocarbon Films. Environ. Sci. Tech., 15_:99 (1981).
20. Pellizzari, E.D. Collection and Analysis of Purgeable Organics Emitted
from Wastewater Treatment Plants. EPA 600/2-80-017., U.S. Environmental
Protection Agency, Cincinnati, OH, 1980. 203 pp.
21. Pellizzari, E.D. Electron Capture Detection in Gas Chromatography. J.
Chromatog., 98:323-361, 1974.
273
-------
22. Pellizzari, E.D., T. D. Hartwell, C. Leininger, H. Zelon, S. Williams,
J. J. Breen and L. Wallace. Human Exposure to Vapor-Phase Halogenated
Hydrocarbons: Fixed-Site vs_ Personal Exposure. Proced. Nat. Symp. on
Recent Adv. in Measurement of Pollutants from Ambient Air and Station-
ary Sources, Raleigh, NC, May, 1982.
23. Simmonds, P. G. and E. Kerns. Direct Aqueous Injection Gas Chromatogra-
phy for the Analysis of Trace Organics in Water, in Advances in Chromato-
graphy, ed. A. Zlatkis. Proc. 14th Symp., U. of Houston, September,
1979. p. 864.
274
------- |