EPA-650/2-74-121
JULY 1974
Environmental Protection Technology Series
DEVELOPMENT OF METHOD
FOR CARCINOGENIC
VAPOR ANALYSIS
IN AMBIENT ATMOSPHERES
Office of Research and Development
U S. Environmental Protection Agency
Washington, DC 20460
-------
EPA-650/2-74-121
DEVELOPMENT OF METHOD
FOR CARCINOGENIC
VAPOR ANALYSIS
IN AMBIENT ATMOSPHERES
by
EdoE. Pellizzari
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
Contract No. 68-02-1228
ROAP 21 BEG, Task 05
Program Element No. 1AA010
EPA Project Officer: Eugene Sawicki
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
July 1974
-------
ABSTRACT
Analytical techniques and instrumentation were developed and evaluated
for the collection and analysis of carcinogenic and mutagenic vapors occurring
in ambient air. The areas of investigation included (a) the design and
3
testing of a cartridge sampler for concentrating trace quantities (ng/m )
of hazardous substances from air, (b) the design, fabrication and evaluation
of a thermal desorption inlet-manifold for recovering vapors trapped on an
analyte and sample transfer into an analytical system, (c) the evaluation
of thermal desorption as a technique for recovering hazardous vapors from
sorbents, (d) the development and performance of a field sampling system for
collecting trace quantities of vapors, and (e) the application of techniques
and instrumentation developed under this program to the analysis of hazardous
vapors in ambient air.
ii
-------
CONTENTS
Page
Abstract ii
List of Figures iv
List of Tables ix
Acknowledgments xi
Sections
I Conclusions 1
II Recommendations 3
III Introduction and Background 5
IV Program Objectives and Experimental Rationale 13
V Design of a Cartridge Sampler for Carcinogenic Vapors 19
VI Gas Chromatographic Inlet-Manifold for Sample Analysis 70
VII Thermal Desorption of Hazardous Vapors from Solid Sor-
bents 87
VIII Design and Performance of a Field Sampler 99
IX Application of Developed Instrumentation and Methodo-
logy 109
X References 123
XI List of Papers Submitted for Publication 129
XII Appendix 130
iii
-------
FIGURES
No.
1 Monitoring System for Hazardous Vapors in Cartridge
22
Effluents.
2 Elution Profile of Cartridge Effluent. Sampling Rate
24
was 6 1/min. Test Mixture II was used.
3 Elution Profile for Synthetic Air/Vapor Mixture I
28
from Spherical Chamber (Empty Cartridge).
4 Collection Efficiency Profile for BPL Carbon Using Test
Mixture I. Sampling Rate was 0.25 1/min; Sensitivity 29
was 6.4 x 10~U AFS.
5 Collection Efficiency Profile for SAL9190 Carbon Using
30
Test Mixture I.
6 Calculated Pressure Differentials for 12/30 Mesh Par-
ticles. Cartridge Dimensions were 1.056 cm i.d. x 38
3.0 cm in length.
7 Calculated Pressure Differentials for 35/60 Mesh Par-
ticles. Cartridge Dimensions were 1.056 cm i.d. x 39
3.0 cm in length.
8 Calculated Pressure Differential for 100/120 Mesh Par-
ticles. Cartridge Dimensions were 1.056 cm i.d. x 40
3.0 cm in length.
9 Comparison of Calculated and Experimental Pressure
/ O
Differential for a Tenax GC (60/80) Cartridge.
10 Pressure Differentials (AP) for Cartridges Containing
/ O
BPL Carbon (12/30) with a Cartridge i.d. of 0.5 cm.
iv
-------
FIGURES CONT'D
No. Page
11 Pressure Differentials (AP) for Cartridges Containing
44
BPL Carbon (12/30) with a Cartridge i.d. of 1.056 cm.
12 Pressure Differentials (AP) for Cartridges Containing
Tenax GC (35/60) with a Cartridge i.d. of 0.5 cm.
13 Pressure Differentials (AP) for Cartridges Containing
46
Tenax GC (35/60) with a Cartridge i.d. of 1.056 cm.
14 Pressure Differentials (AP) for Cartridges Containing
47
Tenax GC (60/80) with a Cartridge i.d. of 0.5 cm.
15 Pressure Differentials (AP) for Cartridges Containing
Tenax GC (60/80) with a Cartridge i.d. of 1.056 cm.
16 Pressure Differentials (AP) for Cartridges Containing
Chromosorb 101 (100/120) with a Cartridge i.d. of 1.056 49
cm.
17 Effect of Particle Mesh Range on Pressure Differential
and Relationship to Cartridge Diameter. BPL Carbon
(12/30), Tenax GC (35/60) and Tenax GC (60/80) are given 50
by A, B, and C, respectively. Packing Bed Depths were
5 cm (A), 3 cm (B) , and 3 cm (C) .
18 Effect of Particle Size on Pressure Differentials. 52
19 Pressure Differential Developed for BPL Carbon (12/30)
53
Cartridge with an i.d. of 0.5 cm During Vacuum Sampling.
20 Pressure Differentials for BPL Carbon (12/30) for Air
54
Drawn Through a 1.056 cm i.d. Cartridge.
-------
FIGURES CONT'D
No. Page
21 Pressure Differential Developed for Tenax GC (35/60)
Cartridge with an i.d. of 0.5 cm During Vacuum Sampling.
22 Pressure Differential Developed for Chromosorb 101
(60/80) Cartridge with an i.d. of 1.056 cm During Vacuum 56
Sampling.
23 Pressure Differential Developed for Tenax GC (35/60)
Cartridge with an i.d. of 1.056 cm During Vacuum Sampling.
24 Pressure Differential for Chromosorb 101 (60/80) for Air
58
Drawn Through a 1.056 cm i.d. Cartridge.
25 Thermal Desorption Inlet-Manifold. 72
26 Thermal Desorption Chamber with Annular Space. Sampling
73
Tube shown in Lower Figure.
27 Thermal Desorption Chamber. 74
28 Electronics Circuit Designed for Temperature Control on
Inlet-Manifold System.
29 Differential Heating Rate in a Glass Cartridge Con-
taining Tenax GC (60/80). 79
30 Temperature Rise Times in Sorbent Bed Using Annular
81
Spaced Chamber.
31 Comparison of Temperature Rise Times for Chambers with
82
and without an Annular Space. Chamber was at 175°C.
32 Comparison of Heating Rates for Some Sorbents. Curves
A, B, and C correspond to PCB Carbon (12/30), Oxopro-
pionitrile on Poracil C (80/100), respectively. Thermal
Desorption Chamber was Isothermal at 210°C.
vi
-------
FIGURES CONT'D
No.
33 Heating Rate for Tenax GC at Different Isothermal
85
Desorption Chamber Temperatures.
34 Gas-liquid Chromatogram of Blank Porapak-Q Cartridge. 91
35 Background During Thermal Desorption of Tenax GC Car-
93
tridge Blank. *
36 Background During Thermal Desorption of PCB Carbon
94
Cartridge Blank.
37 Gas-liquid Chromatogram of Synthetic Air/Vapor Mixture
of Hazardous Substances. Peaks A, B, C, D, and E are
300 mg of Glycidaldehyde, Butadiene Diepoxide, N-nitro-
sodiethylamine, 1,2-dichloroethyl ethyl ether, and ethyl
methanesulfonate, respectively. See test for glc
parameters.
38 Gas-liquid Chromatogram of Vapors Desorbed from Tenax GC.
Desorption Chamber at 225°C; see prior figure for Peak _7
Identity. Background from Tenax GC is represented by
Dashed Profile.
39 Relationship Between Flow Rate and Theoretical Power
102
Requirements at Various Tube Diameters and Particle Size.
40 Schematic of Universal Sampler 5-1068. 105
41 Multiport Sampling Head. 107
42 Gas-liquid Chromatograph-Mass Spectrometer Computer
(GLC-MS-COM) Outlay. 1U
vii
-------
FIGURES CONT'D
No.
43 Total Ion Current Plot During Gas-Liquid Chromatography
Mass Spectrometry of Air Sample from West Covina, CA. 114
44 Single Ion Plots of Ions Common to Aliphatic Cracking
Series. 120
45 Single Ion Plots for Ions Representative of Aromatic
Cracking Series. 121
46 GLC-MS of Bis-(chloromethyl)ether. 131
47 GLC-MS of Bis-(2-chloroethyl)ether. 132
48 GLC-MS of g-Propiolactone 133
49 GLC-MS of Vinylene Carbonate. 134
50 GLC-MS of N-diethylnitrosainine. 135
51 GLC-MS of Nitromethane. 136
52 GLC-MS of Ethyl methanesulfonate. 137
53 GLC-MS of Glycidaldehyde. 138
54 GLC-MS of Propylene Oxide. 139
55 GLC-MS of Styrene Oxide. 140
56 GLC-MS of Butadiene diepoxide. 141
57 GLC-MS of Acrolein. 142
58 GLC-MS of Methylethylketone. 143
59 Mass Spectrum of Maleic Anhydride. 144
60 Mass Spectrum of Succinic Anhydride. 145
61 Mass Spectrum of 1,3 Propanesultone. 146
62 GLC-MS of Tetramethylene Sulfolane. 147
63 GLC-MS of Aniline. 148
viii
-------
TABLES
No. Page
1 Commercially Available Materials with Chemically
10
Bonded Liquid Phases.
2 Sampling Parameters for Collecting 30 ng of Vapor
15
From Air.
3 Reagent Vapors for Evaluating Sampling Medium. 18
4 Synthetic Air-Vapor Mixtures for Cartridge Sampler
25
Evaluation.
5 Expulsion Rates from Chamber at Specified Flows. 26
6 Collection Efficiencies of Candidate Sorbents. 27
7 Approximate Collection Efficiency for Tenax-GC at
31
Various Sampling Rates.
8 Approximate Collection Efficiency for Chromosorb 101
32
at Various Sampling Rates.
9 Elution Volume Characteristics for Tenax GC (35/60) -
60
1 cm Bed Depth.
10 Elution Volume Characteristics for Tenax GC (35/60) -
61
1 and 3 cm Bed Depths.
11 Elution Volume Characteristics for Tenax GC (35/60) -
62
2 cm Bed Depth.
12 Elution Volume Characteristics for Tenax GC (35/60) -
63
3 cm Bed Depth.
13 Elution Volume Characteristics for Tenax GC (35/60) -
64
2 cm Bed Depth and 12 1/min.
14 Elution Volume Characteristics for Tenax GC (35/60) -
65
2 cm Bed Depth and 24 1/min.
ix
-------
TABLES CONT'D
No. Page
15 Elution Volume Characteristics for Tenax GC (35/60) -
66
4 cm Bed Depth.
16 Elution Volume Characteristics for Tenax GC (35/60) -
67
6 cm Bed Depth.
17 Pollutant Profile Breakthrough During Ambient Air
69
Sampling.
18 Percent Recovery of Vapors Adsorbed on Tenax GC
98
Cartridges using Thermal Desorption
19 Power Requirements to Deliver Various Sampling Rates. 103
20 Sampling Rate Characteristics for Universal Sampler with
108
Multiport Head.
21 Protocol for Sampling Ambient Air in Los Angeles, CA. 113
22 Operating Parameters for GLC-MS-COMP System. 115
23 Pollutants in Ambient Air from West Covina, CA. 116
24 Pollutants in Ambient Air from Santa Monica, CA. 118
-------
ACKNOWLEDGEMENTS
The engineering support of Mr. B. Carpenter is gratefully acknow-
ledged for the design of a field sampling unit, pressure differential
calculations for cartridge samplers and a thermal desorption chamber
with an annular flow pattern. The valuable assistance of Mr. J. Bunch
for executing laboratory and field experimentation is appreciated. Mr.
L. Retzlaff provided expert machining and construction of experimental
devices used in this research program; a sincere thanks for his support.
The design and fabrication of the temperature controller was performed
by Mr. R. L. Marguard and Mr. C. Cleary; their help is also gratefully
appreciated.
The personnel at the CHAMP stations in Santa Monica and West Covina
are thanked for their help while field samples were acquired at these
sites and Drs. G. Lauer, J. Hribar and R. Myers at Rockwell International
Science Center in Thousand Oaks, CA for making available these facilities
and their extended courtesies during the author's stay. Approval for use
of CHAMP sites was given by Mr. Ferris Benson of the Health Effects Research
at EPA, Research Triangle Park, N. C.
The helpful suggestions of Dr. M. E. Wall throughout the program
and the computer program for single ion plotting made available by Dr. D.
Rosenthal are appreciated.
-------
SECTION I
CONCLUSIONS
An analytical technique was developed for evaluating collection
efficiencies of candidate sorbents during the concentration of hazar-
dous vapors from a flowing stream. The polymeric beads - Tenax GC,
Porapak Q, Chromosorb 101 and 104 - were >_ 90% efficient in trapping
hazardous vapors such as epoxides, $-lactones, sulfonates, sultones,
N-nitrosamines, chloroalkyl ethers, aldehydes and nitro compounds from
synthetic air/vapor mixtures at 0.25 1/min. Tenax GC and Chromosorb
101 were also tested at sampling rates up to 9 1/min and efficiencies
of >_ 90% were maintained. Carbowax 400 and 600 and oxypropionitrile
coated or chemically bonded to supports and activated carbons were also
highly efficient.
A thermal desorption inlet-manifold for recovering and transferring
hazardous substances from sorbents to a gas-liquid chromatograph or a gas-
liquid chromatograph-mass spectrometer was developed. The interface con-
sisted of a desorption chamber, a six-port two position high temperature
low volume valve, a Ni capillary trap and a temperature controller. This
unit was utilized to determine the temperature rise times in the center
of cartridge samplers for a variety of sorbents under isothermal chamber
temperatures. The heating rates were: PCB and BPL activated carbons > oxy-
propionitrile and carbowax 400 chemically bonded to Poracil C > Chromosorb
104 > Tenax GC > Chromosorb 101. The heating rates were linear for all
sorbents up to 65% of the set desorption chamber (60-90 sec) but required
several minutes thereafter to reach the final temperature. Since the per-
cent recovery of several hazardous vapors adsorbed on Tenax GC using this
-------
inlet-manifold was >_ 90% at the 50 ng level, it was concluded that the
designed system was satisfactory for analysis of cartridge samplers.
Design criteria were developed for a field sampling system for col-
lecting pollutants using a cartridge containing a solid sorbent. The
relationship between the ambient vapor concentration, the total volume
of air required and the time required for sampling at various flow rates
were important considerations in the design specifications. It was con-
cluded that the power required to pump air through sorbent packed tubes
was a function of pressure differential (AP) across the tube under flow
conditions. The AP was found to be related to (1) sampling rate, (2)
diameter and length of cartridge, (3) particle size distribution, and (4)
particle shape. Based upon AP the power requirements for a pumping system
was calculated and applied to the fabrication of a field sampler.
The methodology and instrumentation developed under this program was
also applied to the analysis of air samples from the Los Angeles Basin area.
Using glc-ms-comp techniques, many aliphatic and aromatic compounds were
identified in these preliminary studies. The relative intensities of
single ion plots for ions representing aliphatic and aromatic cracking pat-
terns revealed that Tenax GC did not efficiently trap background aliphatic
constituents, a desirable feature since most hazardous vapor of interest
are semipolar/polar compounds. It was concluded that improved techniques
for resolving background pollutants occurring at high concentrations from
trace hazardous vapors need to be developed.
One oxygenated compound of significant interest tentatively identified
as styrene oxide was discovered in air from West Covina and Santa Monica, CA.
-------
SECTION II
RECOMMENDATIONS
Four major phases of research in the current program are recommended
to be expanded and pursued. These are listed as follows:
(1) Sampling -
A concerted effort should be undertaken to collect field samples
at geographical sites postulated to contain hazardous (mutagens,
carcinogens and other alkylating agents) compounds under study.
The effects of transportation and storage on adsorbents and
reliability and accuracy of analysis for collected pollutants,
and potential sources of contamination should be determined.
Experimental criteria required for optimum performance of adsor-
bents should be established. It is recommended that an alternate
backup sampling device also be examined. The sampling devices
should be improved and the system miniaturized for portability.
(2) Inlet-manifold unit -
The strengths and weaknesses undercovered in this report should
be considered for perfection of the design of the thermal desorp-
tion unit; it should interface efficiently with a glc and glc-ms.
(3) Resolution of pollutant mixture -
Techniques for the separation of hazardous substances under study
from the many hundreds of organic pollutants which are of secondary
interest at this stage should be developed and perfected. It is
recommended that the identity and relative quantities of background
organic pollutants which interfere with the major goals be established,
-------
(4) Characterization of hazardous compounds -
The unequivocal characterization of atmospheric pollutants of
interest and those hazardous to living organisms is recommended
using methodology and instrumentation developed under this re-
search program. Finally, the specifications for routine assay
of hazardous substances based on the discovered ones should be
delineated.
-------
SECTION III
INTRODUCTION AND BACKGROUND
CARCINOGENIC VAPORS IN AMBIENT ATMOSPHERE
1-3
Carcinogenic vapors have been postulated to occur in the atmosphere,
however until the present program was initiated no serious and thorough
endeavor had been made to collect and determine these substances. Epoxides,
peroxides, aldehydes, ketones, lactones, sulfonates, sultones, and nitroso
and nitro compounds have been isolated in laboratory experiments during
4
olefin oxidation, ozonization, sulfonation, nitrosation and nitration .
Some of these compounds and reaction mixtures have demonstrated carcinogenic
activity in animals . For example, the exposure of mice (strain A or
C57B) to an atmosphere containing ozonized gasoline increased the incidence
of tumors .
Because unsaturated hydrocarbons constitute a large fraction of organic
air pollutants, it is reasonable to anticipate that their oxidation products
and their production of reaction with NO and SO , whether spontaneously or
X X
photochemically induced, may also be present in ambient atmosphere . Many
types of alkylating and arylating agents are being introduced directly at a
continually increasing rate into our environment, e.g. as industrial inter-
mediates in organic synthesis, organic solvents for various chemical pro-
cesses, as cross-linking agents in manufacturing processes, as medicines, and
as antibacterial and fungistatic agents.
There is a strong indication that nitrosamines are present in the atmos-
phere; their identification in cigarette smoke has been confirmed . Alky-
lating compounds resulting from limited-aeration smouldering of plastics,
paper, and cellulose as well as auto-exhaust gases have been demonstrated to
2
occur and have been postulated as a health hazard in industrial areas .
-------
On the basis of the current knowledge of air pollution, most of the
potentially deleterious vapors which could be formed in or expelled into
the atmosphere could be therefore classified as epoxides, $-lactones,
peroxides, hydroperoxides, sulfonates, sultones, nitrosamines and a-
chloroalkyl ethers. Some of these pollutants may have a short lifetime
in the atmosphere.
The National Academy of Sciences panel in a study of the biological
effects of atmospheric pollutants has concluded and recommended in their
report on Particulate Polycylic Organic Matter "Research is needed on the
chemistry and biological activity of air pollutant cocarcinogens and tumor-
promoting agents, such as polyphenols and paraffin hydrocarbons, and on the
oxidation products of airborne olefins and aromatic hydrocarbons, including
the nature of the epoxides, hydroperoxides, peroxides, and lactones formed
and their biological properties". Recently, Van Duuren summarized a
review on the biological properties of carcinogenic vapors with the statement
"in view of the obvious importance of these aliphatic compounds (epoxides,
hydroperoxides and peroxides), it is imperative that studies be undertaken
on the analysis of volatile organic air pollutants". Once the identity of
the physiologically active vapors present in polluted atmospheres are known,
then investigators can ascertain which substances need to be routinely ana-
lyzed, studied epidemiologically and eventually controlled.
The primary mission of this research program has been to develop methodo-
logy for the reliable and accurate collection and analysis of mutagenic and
carcinogenic vapors present in the atmosphere down to nanogram per cubic
meter amounts.
-------
METHODS FOR COLLECTION AND ANALYZING POLLUTANTS FROM AMBIENT AIR
Collection Techniques
The characterization and measurement of extremely minute amounts (ppb)
of these hazardous compounds in ambient air has been seriously hampered by
the lack of a reliable sampling system and sensitive instrumentation for
direct analysis. Special systems have been developed for concentrating trace
organic vapors from large volumes of atmosphere and transferring the collec-
12—28
ted vapors to an analytical system
Many collection devices and analytes have been employed by investigators
in air pollution. In general, the concentrating techniques have utilized
. 22-27 , t. 17,28,31 . „. 13-16,18-27 .. . „,,
cryogenic , absorptive or adsorptive trapping methods.
Cryogenic (freeze-out) methods are particularly suitable for analysis of
highly volatile substances; however, if liquid nitrogen, oxygen, or solid
carbon dioxide/acetone is used as a coolant large quantities of water may
accumulate, which is a major problem during chromatographic analysis. Aerosol
formation may be also experienced with this technique reducing the trapping
efficiency. Drying the gas by passing the air over desiccants prior to cryo-
29-31
genie trapping is not feasible since some solutes may also be scrubbed
Nevertheless, the major advantage of this approach is that it is the only
technique which permits the collection of low molecular weight air pollutants.
Also, oxidation or polymerization of constituents is minimized during their
concentration.
21
Activated carbons (including activated carbon molecular sieves ) have
been shown to adsorb all of the classes of carcinogenic-type compounds to
13 14 32-34
be studied under this program ' ' . Where the adsorption is purely
physical, the compounds may be retained and subsequently released without
being decomposed. Brooman used activated carbon to adsorb acrolein
-------
quantitatively up to the point of breakthrough, then used solvent extraction
14
to recover the trapped substance . Saunders used adsorbent carbon to iden-
35
tify 23 volatile compounds in closed environmental atmospheres . Ethylene
oxide has been successively recovered from activated carbon at 30°C; acet-
35
aldehyde required 175°
The high surface activity of activated carbons also can produce arti-
facts during recovery. Formaldehyde decomposes, while methyl ethyl ketone
35
tends to form diacetyl compounds and acetic acid, on activated carbon
Epoxides and peroxides are usually destroyed, hence, activated carbon alone
has very limited potential as a sorbent for collecting highly polar and
reactive compounds.
Compounds that are sensitive to polymerization or decomposition on
20
activated carbon are generally more so on molecular sieves . For this reason
the potential role of molecular sieves is probably limited to removing and
recovering simple molecules (H»0, NH«, CH,) from air.
One unique approach which is based on gas and liquid chromatographic
principles employs liquid phases uniformly coated on solid supports (e.g.
silica, diatomaceous earth, polystyrenes) and these polymers exhibit solution
formation with trace organic vapors at ambient temperatures. Williams has
reported a collection device for organic compounds such as hexane, benzene,
toluene, aldehydes, ketones, and chlorinated hydrocarbons at the part per
hundred million level . The device consisted of a tube packed with Chromo-
sorb P coated with the stationary phase, di-ii-butylphthalate. During air
sampling the collection tube was cooled in dry ice and water was removed by
using a drying agent prior to drawing the air through the sampler.
17 9ft *}fi "^7
Application of chemically bonded stationary phases > » » to
sampling of air pollutants appears to be potentially promising because
-------
selectivity can be incorporated by matching the physico-chemical properties
of the phase with the pollutants of interest. Because the column packings
(glc and Ic) are essentially nonextractable, thermally and hydrolytically
stable, they exhibit lower backgrounds during thermal desorption than the
conventional support coated liquid phases. Permaphase^ (E. I. duPont de
Nemours & Co.) is formed by reacting silane reagents with the surface of
the porous shell of Zipax^and then polymerizing the reagents to yield the
37
desired silicone coating . Ether bonded polymeric coatings are prepared
with a variety of functional groups, ranging from very polar to nonpolar,
providing a wide range of selectivities (Table 1). The physical charac-
teristics of the stationary phase, such as concentration, film thickness,
and structure (i.e. linear or crosslinked), is controlled to obtain the
desired properties for affinity to specific classes of trace organics.
By bonding the functional moiety chemically to the core material the
vapor pressure is reduced to near zero, hence a low bleed rate is observed.
This feature makes these solvents attractive during thermal desorption of
trapped organics since a minimal background interference from the polymer-
solid support occurs.
Thermal degradation of the bonded functional moiety can result if
excessive temperatures are employed. Table I depicts some representative
commercially available materials and their physical properties.
Aue has employed a bonded silicone polymer (octadecyltrichlorosilane)
12
to collect several organic compounds from fast-flowing gas streams
Silicone coatings have been shown to concentrate and release many chemical
classes of interest to this program; however, lactones, epoxides, and per-
oxides had not been examined for quantitative analysis.
-------
Table 1. COMMERCIALLY AVAILABLE MATERIALS WITH
CHEMICALLY BONDED LIQUID PHASES
p
Type
OPN/Porasil C
Carbowax 400/Porasil C
n-Octane/Porasil C
Carbowax 400/Porasil S
Phenyl isocyanate/Porasil C
Carbowax 4000/Porasil S
article Size
(mesh)
80/100
100/120
120/150
80/100
80/100
80/100
Temperature limit
Polarity3 °C
M
N
P
N
P
P
135
150
160
200
60
200
M = medium, N = nonpolar, P - polar
Other types of bonded functional groups are the "brush"-like stationary
phases. These materials are prepared by esterification of the alcoholic
groups on silica surfaces. One certain disadvantage of esters is their
suspectibility to hydrolysis. Conceivably this might be a problem during
collection of trace organics from large volumes of polluted air, if water
vapor is also trapped during the solution process. An example of a "Brush"
type is polyethylene glycol 400 on silica.
/5\
Similarly to the Permaphases^ (or Durapak, Waters Assoc.), Poragel-P
packings are marketed. These are polystyrene resins with permanently
attached functional groups. The surface affinity of the polymer and bonded
functional group also allows interaction and thus trapping of organic solutes.
Several polystyrene type polymers are commercially available (e.g.
Chromosorb 101, 102, 104) which are used in gas-solid chromatography. Chromo-
sorb 101, a styrene-divinylbenzene polymer adsorbs hydrocarbons, alcohols,
acids, esters, aldehydes, ketones, ethers, and glycols. A polar surface is
10
-------
provided by Chromosorb 104 which is an acrylonitrile-divinylbenzene resin.
The potential utility of these polymers has not been completely exploited
for trace organic vapor analysis.
38
Leggett, et_ _al. developed a collection cartridge containing 2 g of
Porapak Q-S sorbent for concentrating and determining trace organic vapors
at ppb levels. Samples were successfully collected at 100 to 1,000 ml/min.
19
In 1966 Hollis reported the use of porous polyaromatic beads concen-
trating air samples on a chromatographic column at room temperature. The
components were subsequently eluted by temperature programming. With these
beads, water is not a major interferent since it is eluted as an early peak
without tailing. Hollis successfully adsorbed and desorbed epoxides from
these polymer beads. The use of polyalkyl styrene polymers to remove organic
39
vapors from a gas stream was patented in 1972 by Haigh
40 41 3
Dravnicks , Crittenden , and Jones have described collection systems
using Chromosorb 102. All employ thermal desorption for recovering trapped
3
vapors. Jones used this solid sorbent to demonstrate the presence of trace
quantities of ethylene sulfite a suspected carcinogen in secular atmospheres.
20
More recently Bertsch et al. described the collection of trace quan-
tities of organic volatiles from air using Tenax GC a porous polymer of 2,4-
diphenyl paraphenylene oxide with a high temperature stability. The trapped
substances were subsequently heat desorbed and the mixture resolved by high
resolution Ni capillary columns. Several hundred hydrocarbons were recognized
and the identity of about 100 was established.
Recovery and Analysis
Because of the limited sensitivity of currently available detectors
hazardous substances need to be concentrated from highly dilute samples. A
step toward the solution of this problem can be achieved when cartridges
11
-------
containing an appropriate sorbent is used and large volumes of air are
forced or drawn through the sampling device whereupon the pollutants are
trapped.
Even when the cartridge technique is used, only trace quantities of
hazardous pollutants can be expected to accumulate; thus it is imperative
that the entire sample be submitted for analysis. Recovery of trapped vapors
has been accomplished using thermal ' and vacuum desorption which
allows for direct introduction of the total sample into an analytical system
(glc, glc-ms) in the absence of a. solvent as a carrier. These methods are
subject to artifactual processes such as pyrolysis, polymerization or incom-
plete recovery. Steam desorption ' and solvent extraction of the
sorbent alleviate the above mentioned problems; however, volatile pollutants
cannot be quantitatively concentrated from dilute solutions and since gas
chromatographic (gc) analysis is limited to small aliquots of liquid samples
only a fraction of the sample can be examined. As a result the sensitivity
of the overall method is greatly reduced.
12
-------
SECTION IV
PROGRAM OBJECTIVES AND RATIONALE
PURPOSE OF PROGRAM
The general scope of this research program was to develop methodology
for reliable and accurate collection and analysis of mutagenic and carcino-
genic vapors (collectively referred to hereafter as hazardous compounds)
3
present in trace quantities in the atmosphere down to ng/m amounts. Once
the physiologically active vapors present in polluted atmospheres have been
determined, investigators can then ascertain what substances need to be
routinely analyzed, studied epidemiologically and eventually controlled.
The major objectives were:
(1) to determine, develop and fabricate a sampling device for the
collection of atmospheric vapors at ambient temperatures, i.e.
substances which are volatile solids or liquids in bulk at room
temperature and pressure.
(2) to interface the sampling device to an analytical system(s) such
as a gas chromatograph or gas chromatograph - mass spectrometer -
computer for the characterization and assay of vapors such as
N-nitrosamines, sulfones, sultones, sulfonates, epoxides, lactones,
anhydrides, aromatic amines, peroxides, hydroperoxides or any
other alkylating agents present in polluted atmospheres.
(3) to acquire performance data for the collection and analysis
methodology embodying the sampling and interfacing devices. This
was to include collection efficiency, reliability of collection
and analysis as they relate to environmental factors such as
sampling locations, temperature, possible interferences from
13
-------
other gaseous pollutants, dust load and winds. Artifacts and
interferences were to be determined and an effort made to cor-
rect or minimize them.
(4) to provide experimental support on -
a. the efficiency of pollutant collection and bleeding, and
their relationship to atmospheric variables,
b. the effect, if any, of other gaseous pollutants on the
reliability of the overall analysis methodology,
c. the qualitative composition of the ambient vapor sample in-
cluding those constituents which are of secondary importance
or interferences,
d. the characterization of those pollutants containing functional
groups common to hazardous compounds (including alkylating
agents)
and e. the estimation of the amounts of some of the atmospheric "•
vapors and especially those containing the functional groups
of interest.
GENERAL APPROACH TO PROBLEM
Although the classes of compounds which are to be studied in ambient
air under this program cover a wide range of chemical functionalities and
physical properties they can be classified as semipolar and polar. For
this reason, the reported techniques for collecting and analyzing atmospheric
pollutants need to be examined carefully prior to their acceptance as valid
methods for use in this program since the literature methods were originally
designed for aliphatic and aromatic pollutants occurring at the parts-per-
million level.
14
-------
The experimental approach adoped here was to thoroughly examine and
evaluate previously reported methods and provide appropriate modifications
which would permit the analysis of hazardous vapors for this program. In
doing so, an in depth study beyond the superficial treatment provided in
the literature was to be conducted.
Because the concentrations of hazardous vapors sought were anticipated
3
to be in ng/ra amounts or less, it was recognized that the major problems
would arise with the development of an adequate collection system for con-
centrating sufficient quantities of material. The relationship between
the sampling parameters for collection, say 30 ng of a pollutant, from
ambient air is shown in Table 2. It was readily apparent that the sampling
time required was directly proportional to the volume of air which must be
concentrated for analysis and inversely to the sampling rate at a given
pollutant concentration. High sampling rates would be required in order to
Table 2. SAMPLING PARAMETERS FOR COLLECTING 30 NG
OF VAPOR FROM AIR
Quantity of Volume
Vapor in Air Required
u*/m
1
5
10
50
100
1000
30
6
3
0.6
0.3
0.03
0.25
2000
400
200
40
20
2
1
500
100
50
10
5
0.5
Sampling Time (hr) at
Various Flows (1/min)
4
125
25
12.5
2.50
1.25
0.125
9
56
11.1
5.6
1.12
0.56
0.056
20
25
5
2.5
0.5
0.25
0.025
30
16.7
3.33
1.6
0.33
0.17
0.017
62.5
8
4
0.8
0.16
0.08
0.008
15
-------
collect enough of a compound within a feasible period of time. For example,
a pollutant at a concentration of 1 ng/m would require a sampling volume
3
of a minimum of 30 m ; at a 1 1/min ( a rate most commonly reported by
investigators) a sampling time of 500 hr would be needed. Obviously, this
would be too long. Sampling rates of 20 1/min or greater would be more
satisfactory.
Many factors come into play when high flow rates are employed, parti-
cularly the pressure differential developed across the sampling device and
subsequently the power requirements needed to achieve the desired rates.
Such factors were considered in the development of a collection and sampling
system.
During the consideration of a collecting system, specific criteria
emerged as important in its design. These were: (1) simplicity, (2) dura-
bility, (3) reliability, (4) ability to store samples for two or more weeks,
(5) convenience during its transportation to and from sampling site, (6)
convenience of its operation and maintaining the collection system during
field trails, and (7) readily interfaced with an analytical system. The
general concept chosen was a cartridge sampler where the analyte consisted
of a solid material capable of trapping vapors of interest.
Since one of the major objectives of this program was to characterize
the collected pollutants and because they were suspected to occur in trace
amounts, the analytical instrumentation of choice which offered the most
promise for achieving these goals were gas chromatography and combined gas
chromatography-mass spectrometry. Interfacing the collection cartridge to
a gas phase analytical system was believed to best be served by using an
interface which would thermally desorb the vapors from the analyte and in-
troduce the entire sample for analysis.
16
-------
Reagent hazardous vapors listed in Table 3 were chosen to assist
in the evaluation of the developed methodology and instrumentation. A
compilation of mass spectra has been assembled in the Appendix of this
report (Fig. 46-63) for many of these compounds. Additional spectra will
be added to establish a library of hazardous vapors for future references
purposes.
This report presents the results obtained during the execution of
the described experimental plan.
17
-------
Table 3. REAGENT VAPORS FOR EVALUATING SAMPLING MEDIUM
Chemical Class
Substance
Chloroalkyl ethers
Lactones
Nitro and Nitroso
Sulfones
Sulfonates
Epoxides
Peroxides
Aldehydes and Ketones
Hydroperoxides
Acid Anhydrides
Sultones
Sulfolane
Aromatic Amines
Imino Heterocyclic
Bis-(chloromethy1)ether
alpha-dichloroethyl ether
3-propiolactone
3-butyrolactone
vinylene carbonate
parasorbic acid
diethy1 nitrosamine
nitromethane
trional
diethy1 sulfone
ethyl methanesulfonate
glycidaldehyde
propylene oxide
styrene oxide
1,2,3,4-diepoxybutane
methyl ethyl ketone peroxide
lauroyl peroxide
glycidaldehyde
acrolein
methyl ethyl ketone
phenylvinyl ketone
cyclohexene hydroperoxide
maleic anhydride
succinic anhydride
propane sultone
tetramethylene sulfone
Aniline
Aziridine
18
-------
SECTION V
DESIGN OF A CARTRIDGE SAMPLER FOR CARCINOGENIC VAPORS
The development of a cartridge containing solid material for the
collection and analysis of carcinogenic vapors from ambient atmosphere
required a study of the physico-chemical properties of several sorbents.
These investigations included (1) the examination of collection effi-
ciencies of several sorbents, (2) the relationship between cartridge
dimensions (length and diameter), sorbent particle size, sampling rate
and pressure differential, and (3) an estimation of the breakthrough
volume for hazardous vapors. During the evaluation of candidate sorbent
media, their collection efficiencies and recovery for analysis were in-
dependently determined. This was important because a sorbent may have
good trapping characteristics but exhibit undesirable effects when a
certain type of recovery method was used. The converse was also the case.
We therefore designed our experiments to allow an independent assessment
of each step.
DETERMINATION OF COLLECTION EFFICIENCIES FOR SEVERAL SORBENTS
The performance of many sorbents as to their ability to extract and
retain hazardous vapors from a moving air stream has not been adequately
studied. The parameters which are involved in determining the performance
of sorbents (collection efficiencies) can be divided into two categories.
There are those on the one hand which are related to sampling environment
such as flow rate, air temperature, and humidity, and those which are re-
lated to the physico-chemical properties of the sorbent such as surface
area, particle size and porosity, solute capacity, sorption mechanism,
degree of solute affinity, etc. Furthermore, some of these factors which
influence sorbent performance are not independent of each other.
-------
Because the collection and analysis of volatile hazardous substances
3
in ng/m amounts from ambient atmosphere requires the selection of a sor-
bent which is efficient under a variety of sampling conditions, an instru-
mental technique was designed for evaluating the collection efficiencies
of sorbents. This section presents the performance of a number of candidate
solid materials for the concentration of substances such as epoxides, $-
lactones, sulfonates, sultones, nitrosamines, chloroalkyl ethers, aldehydes,
and nitro compounds from an air stream comparable to field sampling conditions.
Experimental
Tenax-GC (60/80 mesh), Chromosorb 101 (100/120), Chromosorb 10A
(100/120) and Chromosorb W-HP (100/120) were purchased from Applied Science,
State College, Pa. A series of stationary phases chemically bonded to
supports including carbowax 400/Poracil C (100/120), oxopropionitrile/Pora-
cil C (80/100), and phenylisocyanate/Poracil C (80/100) were also obtained
from Applied Science. Stationary phases consisting of carbowax 600, didecyl
phthalate, and tricresyl phosphate and the sorbent Porapak Q were from
Supelco, Inc., Bellefonte, Pa.
Carbon derived from coke (PCB and BPL, 12/30) was acquired from Pitts-
burgh Activated Carbon Division of Calgon Corp., Pittsburgh, Pa. Cocoanut
derived carbons (SAL19190 and 580-26) were purchased from Barneby Cheney,
Columbus, Ohio.
Ethyl methanesulfonate, 3-propiolactone, N-nitrosodiethylamine, 1,2-
dichloroethyl ethyl ether, nitromethane, methyl ethyl ketone, and aniline
were from Fisher Chemicals, Pittsburgh, Pa. Glycidaldehyde and sulfolane
were obtained from Aldrich Chemicals, Milwaukee, Wis. From Eastman Organic
Chemicals, Rochester, N. Y. 1,3 propane sultone, maleic anhydride, butadiene
20
-------
diepoxide and propylene oxide were purchased. Styrene epoxide, bis-(chloro-
methyl)ether and bis-(2-chloroethyl)ether were from K&K Labs., Plainview,
N. Y.
The monitoring system shown in Figure 1 was designed and assembled
for measuring collection efficiencies. Air (breathing quality, Linde Div.
Union Carbide, East Brunswick, N. J.) from a pressurized reservoir was
passed through a scrubbing tower (5 cm i.d. x 30 cm) which contained layers
of CaCl2 dessicant and BPL activated carbon (12 x 30 mesh, Calgon Corp.,
Pittsburgh, Pa.) to remove trace contaminants. Purified air entered the
monitoring system at point A as shown in the schematic where the flow rate
was controlled with a Sho-Rate 250 flow meter (Model 1357-12F1BAA Brooks
Instruments Div., Emerson Electric Co., Hatfield, Pa.) equipped with a
teflon diaphragm regulator for compensating downstream pressure changes.
The air stream passed through a 2 1 cylindrical chamber (point B) fitted
with an injection port where known quantities of organic vapors were intro-
duced. The chamber delivered synthetic air/vapor mixtures to the cartridge
sampler which contained the sorbent under study (point C) according to the
following relationship:
C - Coe-Ft/V (1)
where C = initial concentration in chamber
o
C = concentration in.chamber after time, t, has elapsed
F = purging rate (ml/min or 1/hr)
V = volume of chamber
The effluent stream from the sampler was split and a flow of 50-100 ml/min
was directed to a flame ionization detector (Model 1200, Varian Instruments,
Corp., Walnut Creek, Ca.). Because hazardous organic vapors were used in
this study, the apparatus between points A and D was contained in a glove box
21
-------
Ni
K>
FLOW
METER/REGULATOR
EXHAUST TO
_ CKY08ENIC SAFETY
TRAPS
RECORDER
7
1
" '"AIR
AMP.
MM
•MMMM^ •
m
-
Figure 1. Monitoring system for hazardous vapors in cartridge effluents
-------
(Kewaunee Scientific Equipment, Adrian, Mich.) which was evacuated by vacuum
through cryogenic safety traps. Hydrogen and air flow to the detector were
35 and 250 ml/min, respectively. The detector output signal was amplified
(Varian Model 520) and recorded with an Omniscribe^ strip chart recorder
(Houston Instruments, Houston, Tx.). This apparatus monitored total organic
vapor in the cartridge sampler effluent.
Sorbents were packed in glass tubes (1.056 cm i.d. x 10 cm in length)
using 1 cm of silanized glass wool plugs for support. The cartridge
samplers were inserted in canisters which were constructed from tube L
3/4 in copper fitted with 3/8 in Swagelock^unions. The entrance and
exit lines at point C in the monitoring system were 3/8 in o.d. Teflon^
(Comco Plastics Corp., Raleigh, N. C.).
To synthesize known concentrations of air/solute vapors mixtures
(Table 4), microliter quantities of each organic compound were added to a
2 1 cylinderical flask. The flask was heated to 50° and the air/vapor mix-
ture was continuously stirred. An aliquot from this stock reservoir was
transferred to the chamber (point B) in the monitoring system. By con-
tinuously monitoring the cartridge sampler effluent with a flame ionization
detector the collection efficiency of each sorbent was determined. A decay
curve (Fig. 2A) which represented the concentration of the air/vapor mixture
leaving the chamber per unit time was established for each mixture and at
each purging rate by using empty cartridge samplers. The per cent collection
efficiency was estimated by comparing the areas under curves obtained for
samplers with and without sorbent (Fig. 2A and 2B).
Results and Discussion
At a sampling rate of 0.25 1/min, over 90% of the synthetic air/vapor
mixture is purged through the cartridge sampler in less than 20 min (Table 5).
23
-------
to
(no sorbent in cartridge)
B (Tenax GC - 1.0 cm i.d. x 3.0 cm)
Figure 2.
4 0
TIME ( WIN.)
Elution profile of cartridge effluent. Sampling
rate was 6 1/min. Test mixture III was used.
-------
Table 4. SYNTHETIC AIR-VAPOR MIXTURES FOR
CARTRIDGE SAMPLER EVALUATION
Mixture Componentsa
I ethyl methanesulfonate
3-propiolactone
N-nitrosodiethylamine
1,2-dichloroethyl ethyl ether
nitromethane
methyl ethyl ketonec
II styrene epoxide
%-nitrosodiethylamine
butadiene diepoxide
glycidaldehyde
sulfolane
propylene oxide
III aniline
bis-(2-chloroethyl)ether
bN-nitrosodiethylamine
Bis-(chloromethyl)ether
maleic anhydride
1,3 propane sultone
Q
Approximately 300 ng of each component was used for synthesizing the
air/vapor mixture.
^Internal standard.
cAs decomposition product from MEK peroxide.
At 9 1/min it takes approximately one minute. Since the decay process
for moderate concentrations of vapor (~0.05 ng/s) can be monitored with
a flame ionization detector, the ability of a sorbent to extract vapors
from a flowing gas stream was determined. This phenomenon was defined
as collection efficiency, i.e. the fraction of solute vapor in the polluted
gas which was retained by the sorbent bed.
The collection efficiencies for several sorbent media are given in
Table 6. All of the polymeric bead sorbents were relatively effective
in extracting vapors from a flowing stream of 0.25 1/min. The activated
25
-------
Table 5. EXPULSION RATES FROM CHAMBER
AT SPECIFIED FLOWS
Concentration (ng/1) remaining in chamber
Time elapsed, min
0
1
2
3
4
5
10
20
30
0.25 Urn
1800
1588
1402
1238
1092
964
516
148
42
4 l/m
3000
406
54.9
7.4
1.0
0.14
0.056
6 A/m 8 £/m 9 fc/m
3000 3000 3000
149 54.9 33.3
7.4 1.0 0.4
0.37 0.02 0.004
0.02
carbons were highly efficient in trapping the constituents in mixture I
(see also figures 3-5); however, they were relatively ineffective when
tested with mixtures II and III. In separate experiments it was dis-
covered that propylene oxide and butadiene diepoxide were not effectively
trapped by two of the cocoanut type carbons and thus accounted for the
low collection efficiencies for synthetic air/vapor mixture II. A com-
parison of several liquid phases coated on solid supports revealed a
considerable decrease in the trapping efficiency when the polarity of the
phase was increased (carbowax 600 vs tricresyl phosphate). These results
suggested that the more non-polar vapors are not forming a "solution" with
the polar liquid phases (absorption via "like dissolves like") as readily
as might be expected if polar vapors were tested. A discrimination in the
26
-------
Table 6. COLLECTION EFFICIENCIES OF CANDIDATE SORBENTS
S-4
(!) CO
S T3
rH 0)
0 pq
P-i
co
pi
o
f>
H
n)
CJ
0]
0)
CO
cd
P-4
13
•H
3
0"
•H
Sorbent
Tenax GC
Porapak Q
Chromosorb 101
Chromosorb 104
Activated carbons
Chemical Type
2,6-diphenyl-p-phenylene oxide
polyalkyl styrene
Styrene-divinyl benzene
acrylonitrile-divinyl benzene
PCB cocoanut (Pittsburgh Act.)
BPL coal (Pittsburgh Act.)
SAL9190 cocoanut (Barneby Cheney)
580-26 Cocoanut/pecan (Barneby-
Cheney)
20% Carbowax 600 on Chromosorb W(HP) 100/120 mesh
bCarbowax 400/Poracil C 100/120
bOxypropionitrile/Poracil C 80/100
25% Didecyl phthalate on Chrom P 100/120
20% Tricresyl phosphate on Chrom W(HP) 100/120
A
Percent Efficiency
Mixture
I
95
90
95
98
90
90
90
95
90
90
98
50
20
Mixture
II
90
95
95
90
95
90
30
30
90
90
96
80
20
Mixture
III
80
90
95
80
90
—
—
—
90
95
90
50
20
Sampling rate was 0.25 1/min, packing bed dimensions were 1.056 cm I.D. x 3.0 cm in length.
exponential dilution flask was maintained at 50°C.
^Chemically bonded phases.
The
-------
Sensitivity?
No Sorbent
Mixture I
Parameters
Sampling Rate: 0.25 1/min
Split Ratio: 5/1
6.A x 10"11 AFS
20 40
TIME (MIN)
60
Figure 3. Elution Profile for Synthetic Air/Vapor
Mixture I from Spherical Chamber
(Empty Cartridge)
28
-------
a
<
o
_
u.
u.
o
UJ
o
oc.
50 r
40
30
20
10
20 40
TIME ( MIN )
60
Figure 4. Collection efficiency profile for BPL carbon usjng
test mixture 1. Sampling rate was 0.25 1/ir.in;
sensitivity was 6.A >: 10 AFS .
29
-------
so r
Parameters
40
Sampling Rate: 0.25 1/min
Split ratio: 5/1
Sensitivity: 6.4 x 10"11 AFS
Sorbent: SAL9190 carbon
Mixture I
O
_J
_l
Z>
u>
o
30
BASELINE
_ .XL
20
40 60
TIME (MIN )
80
100
Figure 5. Collection Efficiency Profile for SAL9190
Carbon Using Test Mixture I
-------
In the affinity and therefore trapping of chemical classes may be pos-
sible with these media; a desirable feature since the final resolution
would be somewhat simplified.
Upon reviewing the collection efficiencies for candidate sorbent
media, and the results from thermal desorption experiments (Section VII)
two polymeric beads, Tenax GC and Chromosorb 101, were selected and
further tested at higher flow rates. As shown by the data in Tables 7
and 8 high collection efficiencies were also obtained at rates up to
9 1/min.
The described monitoring system can be also used to determine the
effects of continuous sampling when no additional solute vapors are pre-
sent in the air stream. This represents the extreme case encountered
during field sampling. When polluted gas enters a sorbent bed an equili-
brium zone is established near the point of entry. As more pollutant is
introduced this zone may expand through the packing length until the
Table 7. APPROXIMATE COLLECTION EFFICIENCY FOR
TENAX-GC AT VARIOUS SAMPLING RATES
Sampling Rate (1/min)
Test Mixture3
I
II
III
4
95b
95
>95
6
90
95
95
8
90
90
95
9
90
90
95
Approximately 1500 ng/component/mixture was used to determine efficien-
cies. Packing dimensions were: 10.5 mm x 60 mm, Mesh 60/80. The
exponential dilution flask was maintained at 50°C.
values are average of duplicate runs.
31
-------
Table 8. APPROXIMATE COLLECTION EFFICIENCY FOR
CHROMOSORB 101 AT VARIOUS SAMPLING RATES
Sampling Rate (1/min)
Test Mixture8
I
II
III
A
95b
>95
>95
6
95
95
95
8
90
95
90
9
95
90
80
Approximately 1500 ng/component/mixture was used to determine efficien-
cies. Packing dimensions were: 10.5 mm x 30 mm, 100/120. Exponential
flask was at 50°C.
All values are average of duplicate runs.
capacity of the sorbent is exceeded. However, if after an initial period
of time no additional polluted vapors are introduced and purging of the
packing bed continues, the zone of vapors may move through the packing
bed. When the mass zone moves to the end of the available packing bed
and the vapors begin to leave, breakthrough has occurred. Furthermore,
the elution volume (EV) can be calculated if the time required for the
zone to traverse and elute from the sorbent bed and sampling rate are
known. In an ideal system, EV has an infinite value. An investigation of
EV is presented later in this section.
The amount of vapor adsorbed to a given quantity of adsorbent depends
54
on the pressure, temperature and concentration of the solute vapor . The
higher the pressure or concentration, the greater the amount adsorbed. When
an adsorbent and solute vapor are placed in contact with each other, an equili-
brium is reached between adsorption and desorption. If the concentration of
the vapor increases or decreases, the mass of adsorbed vapor also increases
32
-------
or decreases to establish a new equilibrium value. This is an important
phenomenona during the collection of vapors from a flowing stream; the
binding affinity of the adsorbent for the solute must be very high in
order that the sampling rate to be relatively independent of collection
efficiency. The data presented in this section supports this viewpoint.
The attractive forces of atoms or molecules responsible for adsorp-
tion on the surface of a solid are attributed generally to two phenomena:
physical and chemisorption. Physical adsorption is primarily due to
van der Waal's forces which is similar to the condensation of a gas. The
magnitude of the heat evolved during adsorption and gas condensation are
similar. The quantity adsorbed may be several monolayers. When the pres-
sure of the vapor or its concentration is lowered, a subsequent decrease
in physical adsorption is observed. In contrast, chemisorption is not
readily decreased and only a monolayer of solute vapor is adsorbed on the
surface. Also, the heat evolved during chemisorption is much larger than
in physical; a surface compound is formed.
Langmuir ' advanced a model theory of adsorption which regards the
surface of a solid to be homogeneous, i.e., that the sites on the surface
all exhibited equivalent affinity for the solute vapor, and each site was
independent of another. However, surfaces are known to be heterogeneous.
The Langmuir equation is represented as follows:
c/q - I/Kb + c/b (2)
where c = concentration of solute (m/1)
q = moles of solute adsorbed per gram of adsorbent
b = moles of solute vapor adsorbed per gram of adsorbent
K = equilibrium constant for V + S —*• VS
33
-------
The complex VS represents the fraction of adsorbent sites occupied by vapor
molecules (V), and V is the fraction of free sites. At a given value of
q, V and VS are constant and equation (2) reduces to
K' = 1/c
Thus, the enthalpy of adsorption, AH°, may be calculated by plotting
1/c vs 1/T at a given q value:
AH° - 2-303RTlT2 (iog K' - log K' ) (3)
12~11
The Clausius-Clapeyron equation (3) can be used to calculate the heat of
desorption. A larger amount of heat is required to desorb a mole of gas
when only a small fraction of the surface is covered than when a large
fraction of the surface is covered; indicating inhomogeneity of the surface.
Since adsorption affinity constants are strongly temperature depen-
dent, the effects of temperature should be considered in any collection
efficiency and breakthrough study. We are currently examining the influence
of temperature and humidity on the collection efficiency of Tenax GC and
Chromosorb 101 using the monitoring system described here. The monitoring
system is also amenable to studies of humidity and temperature and their
effects on sorbent trapping performance. Calibrated humidity levels
can be introduced into the chamber or can be heated with a heating mantel
for simulation of high temperature and humidity as observed in field condi-
tions .
With respect to bed dimensions (length and i.d.) and particle size,
20
it has been suggested that they play no major role on breakthrough ; how-
ever our preliminary studies have shown that an increase in both collection
34
-------
efficiency and EV can be expected. These relationships need to be defined
more closely.
Displacement chromatography may occur during sampling of polluted
air leading to early breakthrough. Because each substance has a specific
affinity for the sorbent, the quantity adsorbed is characteristic for each
substance; furthermore one compound can be displaced by another, if the
latter has a higher adsorption affinity. Thus, breakthrough studies should
be performed in the field under the full complexity of polluted air rather
than in the laboratory. These results are discussed below.
The capacity of Tenax GC for various compounds such as alkanes, alco-
hols, and amines have been reported to be higher than for aldehydes, ketones,
and phenols . Bertsch et al. also reported that all sulfur compounds
examined were trapped in a narrow zone at the cartridge entrance. Vola-
tile hydrocarbon compounds containing less than five carbon atoms are
not efficiently trapped by Tenax GC while aromatics were .
RELATIONSHIP BETWEEN SAMPLING RATE AND CARTRIDGE PRESSURE DIFFERENTIAL
Most of the design principles that have been developed for the effi-
cient use of sorbents have been prescribed for the removal of vapors from
air to improve its quality. For example, Jones discusses the most
efficient physical arrangement of a charcoal bed to remove benzene vapor
from air. His design presumes a steady flow of air of uniform concentration,
with the gradual accumulation of vapor until the sorbent becomes saturated
and breakthrough occurs (frontal elution).
However, additional criteria were considered in this problem during
the design of a sampling system which is based upon a cartridge concept
for the collection and analysis of trace organic vapor pollutants. The
most important factors are revealed in the aerodynamic features of the
35
-------
cartridge, specifically the pressure differential, AP, produced across a
specified sorbent bed. The pressure differential is related to (1) the
air flow-rate, (2) the cartridge shape (diameter and depth of packing),
(3) the particle size distribution, (4) the particle shape, and (5) to a
lesser extent upon the air temperature and humidity. Since AP is an
important ingredient for relating the power requirement of a field sampler
(Section VIII) to sampling rate and duration, it was carefully investigated
in this program.
Experimental
An equation was derived from the mechanical energy balance and Leva's
53
correlation to predict AP which was valid at isothermal flow:
2 2 _ 2SRG2T
1 2 = g M
V
In ~ +
2 2fm
Vl
s
(4)
2
where p. = pressure at inlet to sampler, N/cm
2
p« = pressure at outlet of sampler, N/cm
2 = gas compressibility factor (air = 1)
R = gas constant, 831467 N x cm/kg moles x °K
2
G = fluid mass velocity, kg/sec x cm
T = absolute temperature, °K
g = conversion factor, force to mass, kg x cm/N x sec
M = molecular weight of gas flowing, kg/mole
3
V1 = specific volumes of gas at inlet, cm /kg
3
V,- = specific volume of gas at outlet, cm /kg
fm = modified friction factor for flow through packed solids
e = bed void fraction
2/3
= particle shape factor = V /0.205 x surface area
36
-------
D = particle diameter of packing, cm
L = depth of sampler packing, cm
The particle diameter (Dp) can be calculated as I/(Ex/dp), where x is the
mass fraction with size d and
P
. ,, , xO.5
dp = (dl x d2)
The minimum and maximum openings of adjacent sieves are designated as
d.. and d». In the above equation, the mass-force conversion factor, gas
compressibility factor, and upstream and downstream specific gas volumes
are represented by g , 2, V- and V2, respectively.
For the computations, Z and § were set equal to unity, n and fm were
evaluated as functions of the Reynolds number (particle diameter x mass air
rate/air viscosity), and E was evaluated as a function of the ratio particle
diameter/tube diameter. Because of the relatively small magnitude of
ln(V-/V-), it was omitted. All calculations were programmed and executed
on an IBM 360/70 computer at the Triangle Universities Computation Center,
Research Triangle, N. C.
Results and Discussion
Figures 6-8 depict computer generated pressure differentials for
three different sorbent particle ranges (12/30, 35/60, and 100/120) and one
set of cartridge dimensions. It was evident from these data that AP in-
creased exponentially as the flow rate was increased. Recalling that 10.03
2
Newtons/cm is equivalent to 760 mm Hg, then a flow rate of 5 1/min produced
a AP of 11.4 mm Hg (Fig. 6) for a mesh range of 12/30. On the other hand,
to achieve the same flow rate with mesh ranges of 35/60 and 100/120
(Fig. 2 and 3) a AP of 68.4 and 228 mm Hg, respectively would be produced.
37
-------
Ol
3.92
2.99
2*6
2.33
1.99
1.66
z
a. 133
Q998
Q666
Q334
0.0029
0.167 2.65 5.13 762 10.1 12.6 15.1 17.6 20.0 22.5 25.0
FLOW RATE IN L/MIN
Figure 6. Calculated pressure differential for 12/30 mesh particles.
Cartridge dimensions were 1.056 cm i.d. x 3.0 cm in length.
38
-------
8.33
7.50
6.67
5.84
cT &01
E
z 4.17
o.
< 3.34
2.51
1.68
0.850
0.0194
0.167 1.82 3.47 5.12 6.77 842 10.1 11.7 13.4 15.0 16.7
FLOW RATE IN L / MIN
Figure 7. Calculated pressure differentials for 35/60 mesh particles.
Cartridge dimensions were 1.056 cm i.d. x 3.0 cm in length.
39
-------
U
Q.
<
3.48
3.14
2.81
2.47
2.13
1.79
1.45
I.II
0.771
0.432
0.0929
j I
i i
0.167 0.650 1.13 1.62 2.10 2.58 3.07 3.55 4.03 452 5.00
FLOW RATE IN L/MIN
Figure 8. Calculated pressure differential for 100/120 mesh particles,
Cartridge dimensions were 1.056 cm i.d. x 3.0 cm in length.
40
-------
The calculated pressure drop curves for each mesh size allows some
judgement to be made with respect to the practical attainable flow rates
for each of these mesh sizes, and the most suitable bed packing dimensions
which will allow the attainment of desired field sampling rates. Further-
more, large AP values were undesirable since vacuum desorption ("stripping")
of vapors initially trapped on a sorbent would occur if a sampling system
was used whereby the air sample was drawn through the cartridge. This situa-
tion would thus seriously affect the collection efficiency and no doubt
breakthrough (elution volume decreased) would occur prematurely. A pumping
system which forces the air through the cartridge would be preferrable for
these reasons.
In addition to deriving a function which would allow AP to be pre-
dicted under a variety of conditions, AP was experimentally determined to
establish the validity of equation (1). The pressure differential was
measured for air either drawn or forced through the cartridge containing
sorbent. Figure 9 compares AP values which were calculated assuming an
average particle diameter, D , of 0.0211 cm to the experimentally determined
P
for a sorbent with a 60/80 U.S. mesh range. This data demonstrates a good
correlation between AP values calculated using equation (1) and those ex-
perimentally measured; however, for exact calculation of pressure drop the
particle size and shape must be known.
The effect of (1) bed packing diameter, (2) bed length, (3) sampling
rate, and (4) particle mesh range on the pressure differential developed
across a sampling'cartridge was examined. These data are depicted in
Figure 10-17 for air forced through the sampler. High AP values were
experienced when cartridge diameters of 0.5 cm were tested (Figures 10, 12,
and 14) at high sampling rates, even when coarse sorbent particles
41
-------
700r
600-
900-
E
E
a.
g
a
ui
ui
4OO-
300
200-
100-
A = experimental
O = calculated
D = 0.0211 cm
P
Cartridge = 1.056 cm i.d. x
3.0 cm
Figure 9.
10
FLOW RATE , liters / min
Comparison of calculated and experimental pressure
differential for a Tenax GC (60/80) cartridge.
20
42
-------
1200
1000
800
I 600
400
200
B
2 4 6 8 10 12 14 16 18
FLOW RATE ( l/min )
1200,-
1000-
800-
600-
400-
200
9 l/min
l/min
5 i/min
4 l/min
2468
PACKING DEPTH (cm )
Figure 10. Pressure differentials (AP) for cartridges containing BPL carbon (12/30) with a
cartridge i.d. of 0.5 cm.
-------
200
160
E
E
120
80
40
6 10 14
FLOW ( l/min)
200
160
120
80
40
B
18 l/min
15 l/min
9 l/min
2468
RUCKING BED DEPTH (cm)
Figure 11. Pressure differentials (AP) for cartridges containing BPL carbon (12/30)
with a cartridge i.d. of 1.056 cm.
-------
6cm
1400
1200
1000
— 800
o»
X
E
E
600
Q.
<
400
200
ft/ f
• AA Ji '
I I I I I
4 6 8 K) 12 14 16 18
FLOW l/min
1400
1200
1000
800
600
400
200
B
2 4
PACKING BED DEPTH (cm)
6
Figure 12. Pressure differentials (AP) for cartridges containing Tenax GC (35/60) with a cartridge
i.d. of 0.5 cm.
-------
1400 r
1200
1000
E
E
800
600
400
200
8 cm
•
5 cm
3cm
cm
.D-
j L
2 4 6 8 10 12 14 16 18
FLOW l/min
KOO
1200
1000
800
600
400
200
O
B
2468
PACKING BED DEPTH (cm)
Figure 13. Pressure differentials (AP) for cartridges containing Tenax GC (35/60) with a
cartridge i.d. of 1.056 cm.
-------
4cm
E
E
1400
1200-
1000 -
800 -
B
600-
400 -
200 -
6 10 14
FLOW l/min
1400
1200
1000
800
600
400
200
5 l/min
l/min
2 4
PACKING BED DEPTH (cm)
Figure 14. Pressure differentials (AP) for cartridges containing Tenax GC (60/80)
with a cartridge i.d. of 0.5 cm.
-------
1400 r
00
1400
6 10 14
FLOW ( l/min )
1000
600
200
B
2468
RACKING DEPTH (cm)
Figure 15. Pressure differentials (AP) for cartridges containing Tenax GC (60/80)
with a cartridge i.d. of 1.056 cm.
-------
8 cm
l4CX)r
1000-
E
E
Q.
<
600-
200-
I
5cm
3cm
Icm
1
4 8 12 16
FLOW ( l/min )
20
1400,-
1000-
600-
200-
B
18 l/min
15 l/min 9 |/min
D o
5 l/min
j
2468
PACKING BED DEPTH (cm)
Figure 16. Pressure differentials (AP) for cartridges containing Chromosorb 101 (100/120)
with a cartridge i.d. of 1.056 cm.
-------
1400
1000
E
E
a.
<
600
200
18 l/min
1000
600
200
0.5 1.0 1.5
CARTRIDGE DIAMETER
(I.D.), cm
B
1400
1000
600
200
A 9 l/min
5 l/min
0.5
CARTRIDGE
1.0 1.5
DIAMETER
( I.D.) . cm
0.5 1.0 1.5
CARTRIDGE DIAMETER
(ID.), cm
Figure 17. Effect of particle mesh range on pressure differential and relationship to
cartridge diameter. BPL carbon (12/30), Tenax GC (35/60) said Tenax GC (60/80)
are given by A, B, and C, respectively. Packing bed depths were 5 cm (A),
3 cm (B) , and 3 cm (C) .
-------
(12/30 mesh) were used. Similarly, a mesh range of 100/120 in 1.056 cm
diameter cartridges yielded large AP values (Figure 16). Figure 17 and
18 summarizes the effect of cartridge diameter and particle size on pres-
sure drop respectively. At a flow of 10 1/min AP doubles when 35/60 and
60/80 mesh ranges are compared. It increases almost an order of magnitude
from mesh 12/30 to 60/80. These data taken collectively indicated that
cartridge diameters of 0.5 cm and containing mesh ranges of >^ 35/60 pre-
cluded their use in any studies which would require field sampling rates
> A 1/min; likewise, a mesh of 100/120 and bed diameter of £ 1.056 cm would
not be suitable.
Pressure differentials developed across cartridges while drawing air
through the sampler were also measured (Figures 19-24). The general trends
observed in previous AP experiments were also apparent in these studies.
However, the restrictions on sampling rates were even greater.
ESTIMATION OF BREAKTHROUGH DURING FIELD SAMPLING
On the basis of results obtained for collection efficiencies, thermal
desorption (Section VII) and pressure drop measurements, Tenax GC and
Chromosorb 101 were selected for further study as possible candidate ma-
terials for collecting hazardous vapors. Field sampling experiments were
designed to provide relative breakthrough data for volatile vapors, i.e.
to determine when the mass transfer zone has moved to the end of the
available packing bed. The estimation of elution volume required for break-
through to occur provided a further assessment of sorbent materials and
their utility for collecting carcinogenic vapors from ambient atmosphere.
Furthermore, -the packing bed depth required to minimize breakthrough yielded
information useful for designing reliable cartridge samplers.
-------
700r
600
500
E 400
E
a.
o
I
LU
o:
Q.
300
200
100
TUBE DIAMETER, 1.06 cm
PACKING DEPTH, 5 cm
TENAX GC 60/80 U.S. MESH
TENAX GC 35/60
U.S. MESH
BPL CARBON 12/30
U.S. MESH
FLOW RATE , liters /min
Figure 18. Effect of particle size on pressure differential.
52
-------
Ul
OJ
X
£
E
a.
<
500
400
300
200
100
2 4 6 8 10 12
FLOW RATE (l/min)
5OOr
400 -
300-
200-
6 l/min
23436
PACKING DEPTH (cm)
Figure 19. Pressure differential developed for BPL carbon (12/30) cartridge with an i.d. of
0.5 cm during vacuum sampling.
-------
600
400
o>
x
E
E
200
7 cm
&
/** 6 cm
.0
A /
.,0 5 cm
A' o^ *S 3cm
2 6 10 14 18
FLOW RATE ( l/min )
600
400
200
y
14 l/min
10 l/min
2468
PACKING BED DEPTH (cm)
Figure 20. Pressure differentials for BPL carbon (12/30) for air drawn through a
1.056 cm i.d. cartridge
-------
600 r
600r-
400
c*
X
E
E
o.
<
200
3 cm
2 cm
cm
4 8 12
FLOW RATE ( l/min )
400-
200-
5 l/mln
01234
PACKING DEPTH ( cm )
Figure 21,
Pressure differential developed for Tenax GC (35/60) cartridge with an i.d,
of 0.5 cm during vacuum sampling.
-------
600 r
Ul
400
o>
X
E
E
200
600,-
2 cm
D
1.5 cm
11
4O I cm
0246
FLOW RATE ( l/min )
400
200
4 l/min
O
01234
PACKING DEPTH (cm)
Figure 22. Pressure differential developed for Chromosorb 101 (60/80) cartridge with
an i.d. of 1.056 cm during vacuum sampling;..
-------
800
600
E 400
< 200
7 cm
6 cm
m
/ 0 | 5cm
26 10 14 18
FLOW RATE (l/min)
800
600
20°
B
7 l/min
6 l/min
4 l/min
2468
PACKING BED DEPTH (cm)
Figure 23. Pressure differential developed for Tenax GC (35/60) cartridge with an
i.d. of 1.056 cm during vacuum sampling.
-------
600 r
400
oo
X
E
E
o.
< 200
7 cm
600
cm
4 8 12
FLOW RATE ( l/min )
16
400
200
6 l/min
5 l/min
l/min
2468
RACKING BED DEPTH (cm)
Figure 24. Pressure differential for Chromosorb 101 (60/80) for air drawn through
a 1.056 cm i.d. cartridge.
-------
Experimental
To determine ET and EV a dual tandem cartridge arrangement and the
sampling system described in Section VIII were used. At the beginning
of each sampling period a known quantity of vapor (methyl ethyl ketone,
phenyl methyl ether and/or nitrobenzene) was introduced directly into the
entrance corridor of the front cartridge. Periodically (15 and 30 min
intervals) the back cartridge was replaced with a virgin one and the back-
up cartridge was examined by thermal desorption-glc (Section VII). At the
end of the experiment the front cartridge of the duo-series was also analyzed.
Peak areas for each standard were calculated to determine the amount of vapor
disappearing from the front cartridge and appearing in the second, backup
cartridge. The ET and EV values of four unidentified constituents (to be
identified by glc-ms) present in ambient air were also determined.
Breakthrough on total "pollutant profile" was also evaluated by using
the duo-tandem cartridge arrangement. The area under the chromatographic
peaks developed from a glc program run (75-165°C, 10°/min, 12 ft 2% DECS
column) for each cartridge was estimated with a planimeter and expressed
as a percent of the total area for each set of tandem cartridges.
Results and Discussion
Table 9 presents the results observed for packing bed dimensions of
1.056 cm i.d. x 1 cm in length (front and back). No methyl ethyl ketone
(MEK) or nitrobenzene (NB) could be detected in either the front or backup
cartridge at any time during sampling. On the basis of these observations
it was concluded that breakthrough (and possible low collection efficiencies)
had occurred in less than 15 min. On the otherhand PI, P2, P3 and P4 was
detected and appeared to remain relatively constant in the backup cartridge
throughout the entire sampling period. Since P1-P4 were endogenous interfering
59
-------
Table 9. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a - 1 CM BED DEPTH
Sampling Volume
Time (Min)
Front
15
45
75
105
165
195
225
255
315
Cartridge
Air Sampled
(1)
180
540
900
1260
1980
2340
2700
3060
3780
Peak Area
MEK
NDC
ND
ND
ND
ND
ND
ND
2.8
ND
ND
PI
1.
1.
2.
1.
3.
1.
2.
2.
-
9.
7
8
2
9
0
9
0
7
2
P2
1.
1.
2.
4.
4.
3.
1.
1.
-
11.
9
9
5
0
1
4
9
6
5
(Cm2)b
P3
2.
1.
3.
6.
4.
3.
4.
3.
-
13.
7
9
0
5
8
7
0
7
5
P4
1.9
0.7
1.0
1.9
2.6
1.8
2.2
2.2
-
10.0
NB
ND
ND
ND
ND
ND
ND
4.3
ND
ND
ND
lacking bed Dimensions - Front Cartridge: 1 cm dia. x 1 cm in length,
Backup Cartridge: identical to front sampler.
Sampling rate - 12 1/min; sampling location - Res. Tri. Prk.
bPl, P2, P3, P4 are unidentified peaks appearing at 100, 115, 135, and
165°C, respectively in chromatogram of air sample.
CND = not detected.
constituents in ambient air and were probably present during the entire
sampling period (frontal analysis), it could not be concluded whether
these results reflected low collection efficiencies, early breakthrough or
both.
Because early breakthrough was suspect in the previous experiment,
another study was conducted which utilized a 1.056 i.d. x 1 cm long and
a 1.056 x 3 cm long front and back cartridges, respectively. Table 10
shows these results. In this case early breakthrough for MEK and NB was
60
-------
Table 10. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a - 1 AND 3 CM BED DEPTHS
Sampling Volume
Time (Min)
15
30
60
90
120
150
180
210
240
Front Cartridge
Air Sampled
(1)
180
360
720
1080
1440
1880
2160
2560
2880
MEK
NDC
6.6
ND
ND
d1.2
2.0
ND
ND
ND
ND
Peak Area
PI P2 P3
1.8 5.3 7.3
1.6 3.7 4.8
2.6 2.7 5.3
2.6 5.3 6.3
3.2 5.1 8.0
3.0 6.3 8.5
_
_
_
4.9 6.0 5.4
(Cm2)b
P4
2.5
1.5
2.0
1.4
2.7
3.1
-
-
-
26.0
NB
12
13
14
19
25
26
30
24
20
2.5
Packing bed dimensions - Front Cartridge: 1 cm dia. x 1 cm in length
Backup Cartridge: 1 cm dia. x 3 cm in length
Sampling rate - 12 1/minj sampling location - Res. Tri. Prk.
bSee Table 1 for description of PI, P2, etc.
CND = not detected.
be a contaminant peak.
detected which is consistent with the previous conclusions. Since the air
sample was drawn through the cartridge, a vacuum gradient was created
across the packing bed. This situation probably leads to vacuum desorp-
tion or "stripping" of the analyte and therefore early breakthrough of
solute vapors. In effect a smaller EV occurs for the backup cartridge
than for the front cartridge (the pressure differential across increasing
packing bed lengths are shown to be nonlinear) .
Tables 11 and 12 present EV characteristics for packing bed lengths
3
of 2 and 3 cm. Breakthrough occurred after approximately 1.5 m of air
61
-------
Table 11. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a - 2 CM BED DEPTH
Sampling Volume Air Sample
Time (Min) (1)
15
45
75
105
135
165
195
225
255
180
540
900
1260
1620
1980
2340
2700
3060
id
PI
2.0
2.9
2.2
2.6
4.8
3.9
3.7
2.4
4.7
P2
0.8
3.1
1.6
3.3
5.2
3.6
3.2
6.5
6.0
Peak
P3
3.0
5.1
6.6
4.0
8.4
6.6
8.2
7.9
9.6
Area (Cm2)b
P4
2.2
1.9
1.6
1.4
3.5
2.8
3.3
3.2
5.0
NB
NDC
ND
ND
ND
75.6
47.2
40.0
40.0
35.0
lacking bed dimensions - Front Cartridge: 1 cm dia. x 2 cm in length
Backup Cartridge: same as above
Sampling rate - 12 1/min; sampling location - Res. Tri. Prk.
bPl, P2, P3, P4 are described in Table 6.
CND = not detected
was sampled for a 2 cm bed length of sorbent. Increasing the bed depth
to 3 cm prevented breakthrough; all of the NB was detected in the front
3
cartridge. The EV for MEK was about 2 M for a 1 x 3 cm cartridge.
Two sampling rates 12 and 24 1/min were compared to determine whether
the elution volume characteristics were flow dependent. The results shown
in Tables 13 and 14 indicated that breakthrough is relatively independent
of sampling rate when phenyl methyl ether (PME) and nitrobenzene (NB) were
used as the test vapors. Analysis of the front cartridge revealed the
presence of PME and NB and therefore breakthrough had not occurred.
62
-------
Table 12. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a - 3 CM BED DEPTH
Sampling Volume Air Samp]
Time (Min) (1)
15
45
75
105
135
165
195
225
255
Front Cartridge
180
540
900
1260
1620
1980
2340
2700
3060
Led
MEK
NDC
ND
ND
ND
ND
18.0
25.0
15.0
ND
ND
2
Peak Area (Cm )
PI
1.8
1.2
1.6
1.5
1.7
1.7
1.6
0.8
1.0
9.6
P2
2.5
2.4
2.2
4.2
4.5
3.8
3.9
3.4
3.7
11.5
P3
2.6
1.4
3.7
7.6
4.3
5.0
5.4
4.0
4.7
20.4
,b
P4
1.2
0.8
1.3
1.5
1.9
1.9
1.8
1.5
2.2
5.7
NB
ND
ND
ND
ND
ND
ND
ND
ND
ND
8.4
i'acking bed dimensions - Front Cartridge: 1 cm dia. x 3 cm in length
Backup Cartridge: same as above
Sampling rate - 12 1/min; sampling location - Res. Tri. Prk.
bSee Table 6 for description of PI, P2, etc.
CND = not detected
Using a flow of 24 1/min, 4 cm and 6 cm beds of Tenax GC (35/60) were
evaluated and breakthrough did not occur for NB even after 15 m of air
had been drawn through the front cartridge and analysis of the front car-
tridge by glc confirmed the presence of NB (Tables 15 and 16). On the
3
otherhand, PME appeared to breakthrough after about 9m of air had been
sampled through a 6 cm bed of Tenax GC packing but not for a 4 cm bed.
This was attributed to the higher vacuum experienced with a 6 cm bed to
maintain a 24 1/min flow rate and thus "vacuum stripping" of PME probably
occurred.
63
-------
Table 13. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a - 2 CM BED DEPTH AND 12 L/MIN
Sampling Volume Air Sam]
Time (Min) (1)
3 led
Peak Area (Cm^
PI
15
30
60
90
120
150
180
210
240
270
Front Cartridge
170
350
710
1060
1420
1770
2120
2480
2830
3190
3190
1.
1.
3.
3.
2.
3.
3.
1.
1.
1.
3.
3
3
2
1
5
2
4
9
2
4
6
PME
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
7.2
P2
1.9
1.9
2.3
4.9
4.1
3.0
2.9
0.8
1.1
2.6
6.4
P3
4
5
2
10
11
3
3
1
1
6
12
.1
.5
.1
.5
.0
.8
.0
.8
.6
.8
.6
P4
1.9
1.5
1.5
3.6
5.6
3.0
3.3
0.9
1.1
3.7
So.o
)b
P5
0
0
0
1
2
1
1
0
1
1
9
.9
.6
.6
.3
.8
.8
.8
.9
.0
.7
.6
NB
NDC
ND
ND
ND
ND
ND
ND
ND
ND
ND
10.8
Packing bed - Front Cartridge: 1.5 cm i.d. x 2.0 cm in length.
Backup Cartridge: 1.5 cm i.d. x 3.0 cm in length.
Sampling rate - 12 1/min; sampling location - Res. Tri, Prk.
bPl, P2, P3, P4, P5 were peaks eluting at 81°, 95°, 110°, 135°, and
165°, respectively; see text for glc conditions.
CND = not detected
A comparison of the "pollutant profile" from Los Angeles, CA. for
the front and backup cartridges were also made to evaluate overall break-
through for Tenax GC and Chromosorb 101 sorbents during field sampling.
These results are given in Table 17. Early breakthrough from the front
cartridge resulted when Chromosorb 101 was used as the collection media.
In contrast, more than 90% of the "profile" collected on Tenax GC was
3
still in the front cartridge after 5.6 m of air had been sampled.
64
-------
Table 14. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a - 2 CM BED DEPTH AND 24 L/MIN
Sampling Volume Air Samp
Time (Min) (1)
led
Peak
PI
15
30
60
90
150
180
Front Cartridge
350
710
1420
2120
3540
4250
4250
1
1
2
1
1
2
1
.1
.7
.4
.4
.2
.5
.5
PME
ND
ND
ND
ND
ND
ND
13.0
P2
2.0
2.8
3.8
2.4
2.1
3.4
4.6
P3
3
7
8
4
5
10
8
Area
(cm2)b
P4
.4
.2
.8
.4
.7
.0
.8
2
3
4
4
4
2
9
.1
.1
.9
.8
.5
.1
.6
P5
1.1
1.1
1.9
2.0
1.6
0.9
5.0
NB
NDC
ND
ND
ND
ND
ND
3.4
T'acking bed - Front Cartridge: 1.5 cm i.d. x 2.0 cm in length
Backup Cartridge: 1.5 cm i.d. x 3.0 cm in length
Sampling rate - 24 1/min; sampling location - Res. Tri. Prk.
bPl, P2, P3, P4, P5 were unidentified peaks eluting at 80°, 95°, 110°,
135°, and 165°, respectively; see text for glc conditions.
CND = not detected
-------
Table 15. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a- 4 CM BED DEPTH
Sampling Volume Air Samp
Time (min) (1)
30
60
90
120
150
180-
210
240
270
300
330
360
390
420
450
480
Front Cartridge
710
1420
2120
2830
3540
4250
4960
5660
6370
7080
7790
8500
9200
9910
10620
11330
11330
led
Peak
PI
2
1
2
0
3
2
2
3
1
3
2
2
2
2
2
2
1
.0
.7
.6
.8
.8
.9
.8
.4
.2
.5
.1
.0
.9
.0
.1
.8
.6
PME
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
7.4
P2
2.4
1.3
3.7
1.4
4.5
3.4
4.5
3.3
1.0
4.5
2.4
2.6
3.1
2.8
3.2
3.3
2.0
P3
5
2
7
3
7
7
10
7
2
10
5
5
7
7
6
8
4
Area (cm )
.8
.1
.0
.1
.7
.5
.5
.0
.4
.5
.2
.2
.5
.6
.7
.5
.6
P4
4.5
1.3
4.3
1.8
4.5
8.5
14.0
2.8
1.6
8.1
2.6
5.5
9.1
10.5
9.0
5.4
5.8
P5
1
0
2
0
1
3
3
1
0
3
1
2
4
3
4
2
4
.3
.7
.8
.5
.5
.7
.8
.8
.7
.5
.3
.4
.0
.6
.3
.1
.8
NB
NDC
ND
ND .
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3.4
racking bed - Front Cartridge: 1.5 cm i.d. x 4 cm in length
Back Cartridge: 1.5 cm i.d. x 3 cm in length
Sampling rate - 24 1/min; location - Res. Tri. Prk.
See previous Table for explanation of PI, P2, etc.
CND = not detected.
66
-------
Table 16. ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a- 6 CM BED DEPTH
Sampling Volume Air Samp]
Time (rain) (1)
30
60
90
150
210
270
300
330
360
390
420
480
540
570
600
660
720
750
810
870
930
960
350
710
1060
1770
2480
3190
3540
3890
4250
4600
4960
5660
6370
6730
7080
7790
8500
8850
9560
10270
10970
11330
.ed
Peak Areas
PI
0
0
0
1
1
2
0
1
1
2
3
2
3
3
1
3
.84
.68
.50
.40
.68
.64
.55
.35
.20
.40
.38
-
-
-
-
-
-
.53
.00
.18
.62
.00
PME
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
10.8
3.0
ND
ND
ND
P2
2.5
1.5
1.4
2.9
3.0
4.9
0.9
0.9
1.2
1.9
1.5
4.8
1.2
2.1
3.2
2.2
1.7
1.0
3.1
3.2
1.7
3.0
P3
4
2
1
4
5
5
2
2
2
5
3
10
3
6
9
4
4
2
8
8
4
6
(cm
V
P4
.1
.8
.9
.5
.4
.5
.0
.0
.3
.5
.8
.5
.1
.5
.3
.8
.9
.3
.4
.2
.5
.2
2
4
1
2
1
2
1
0
3
2
1
6
1
3
7
1
2
1
5
7
4
6
.3
.6
.5
.4
.9
.8
.8
.9
.6
.6
.7
.3
.1
.7
.2
.8
.2
.9
.2
.5
.7
.8
P5
0.4
0.7
0.4
0.6
0.6
0.6
0.6
0.6
1.3
1.0
0.8
2.2
1.0
0.8
2.4
1.5
1.5
1.0
1.7
2.6
1.7
3.5
NB
NDC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
67
-------
Table 16 (continued). ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
(35/60)a - 6 CM BED DEPTH
Sampling Volume Air Sam]
Time (min) (1)
1020
1080
1140
1200
1260
Front Cartridge
12040
12740
13450
14160
15000
15000
)led
PI
2.40
1.75
1.75
0.84
1.00
1.20
fy u
Peak Areas (cm )
PME
ND
ND
ND
ND
ND
ND
P2 P3 P4 P5 NB
2.0 4.1 2.5 3.0 ND°
2.4 3.8 6.0 3.2 ND
2.2 4.9 8.2 2.8 ND
1.7 4.0 2.1 2.0 ND
- ND
- 4.8
lacking bed - Front Cartridge: 1.5 cm i.d. x 6 cm in length
Back Cartridge: 1.5 cm i.d. x 3 cm in length
Sampling rate - 24 1/min; location - Res. Tri. Prk.
bSee"Table 14 for explanation of PI, P2, etc.
CND = not detected
68
-------
Table 17. POLLUTANT PROFILE BREAKTHROUGH DURING AMBIENT AIR SAMPLING3
vo
Sorbent Cartridge
Chromosorb 101 (60/80)
Front
Backup
Front
Backup
Front
Backup
Front
Backup
Front
Backup
Front
Backup
Tenax GC (35/60)
Front
Backup
Front
Backup
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Volume Air
Dimensions Sampling Time (Min) Sampled (1)
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
3
3
3
3
3
3
1
1
1
1
1
1
3
3
3
3
.0
.0
.0
.0
.0
.0
.5
.5
.5
.5
.5
.5
.0
.0
.0
.0
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
30
30
60
60
90
90
30
30
60
60
120
120
60
60
120
120
450
450
1340
1340
3970
3970
400
400
1400
1400
5400
5400
1260
1260
5600
5600
Percent of
Combined Areab
75
25
40
60
40
60
90
10
55
45
40
60
90
10
95
5
Experiments were performed in Santa Monica, CA.
bThe total area of the chromatogram between 75-160° (lp°/min) on a 12 ft 2% DECS column was
determined using a planimeter.
-------
SECTION VI
GAS CHROMATOGRAPHIC INLET-MANIFOLD FOR SAMPLE ANALYSIS
Because of the limited sensitivity of currently available detectors
hazardous substances need to be concentrated from highly dilute samples.
A step toward the solution of this problem is achieved when cartridges
containing an appropriate sorbent is used and large volumes of air are
forced or drawn through the sampling device whereupon the pollutants
are trapped. The performance (collection efficiency, pressure differen-
tial, elution volume) characteristics for a selected number of sorbents
was described in Section V.
Even when employing a cartridge technique, only trace quantities of
hazardous pollutants are accumulated; thus, it is imperative that the
entire sample be submitted for analysis. The levelo of carcinogenic and
mutagenic vapors (e.g. epoxides, nitrosamines, sulfonates, sulfites, sul-
tones, aldehydes, ketones, 6-lactones, chloroalkyl ethers and nitro com-
pounds) to be collected and identified from ambient air under this research
3
program are anticipated to occur at ng/m amounts or less. These severe
restrictions on sampling and analysis have required the development of a
technique(s) such as thermal desorption as a means of transferring the
entire amount of trapped vapors from the cartridge to the analytical system.
This section describes (1) the design and fabrication of an inlet-
manifold for effecting desorption of vapors and efficient transfer of pol-
lutants to a glc, and (2) the heating characteristics of selected sorbents
using the inlet-manifold.
DESIGN AND FABRICATION OF INLET-MANIFOLD
Thermal desorption systems of varying designs have been des-
cribed ' ' for recovering trapped vapors and their subsequent
70
-------
46
analysis by glc and glc-ms. Duel's solvent-free sample inlet system
was designed so that with the sample fraction "in-line" with the sample,
the entire mixture could be introduced directly onto the glc column.
This concept was retained for the development in this research program
of an interface system for analyzing carcinogenic vapor samples by glc
and glc-ms.
The described manifold design evolved from a consideration of
several criteria. These were: (1) the heat transfer characteristics of
cartridge samplers containing sorbents of various physical dimensions,
(2) the flow rate requirements for purging desorbed vapors from the cart-
ridge sampler and for packed and glc capillary columns, (3) the tempera-
tures necessary for effecting thermal desorption of trapped carcinogenic
vapors from sorbents (from Section VII), and (4) the ability to con-
veniently interchange and assemble on any standard glc or glc-ms instrument,
The fabricated inlet/manifold system (Fig. 25) consisted of four main
components: a desorption chamber, a six-port two position high temperature
low volume valve (Valco Inc., Houston, Tx.), a Ni capillary trap, and a
temperature controller. One of the three prototype brass thermal desorp-
tion chambers constructed for use in this program is shown in Figure 26.
Brass was chosen for its good conductive properties. The configuration
of the chamber was designed to allow an inert gas to enter through a side-
arm near the bottom and sweep up an annular space between the chamber and
glass cartridge. This permitted the purge gas to be preheated to the
chamber temperature prior to passing down through the sorbent bed. Two
additional prototype brass desorption chambers were designed so that the
purge gas entered near the top and passed directly through the glass
cartridge to the valve proper (Fig. 27). The overall chamber lengths were
71
-------
COMPRESSION SPRING
HEATING CARTRIOE
CARRIER GAS
TO GLC CAPILLARY
HEATING AND COOLING BATH
Nl CAPILLARY TRAP
VALVE POSITION A
(SAMPLE oeaonrriOM)
CARRIER
GAS
1
WOVE POSITION 8
(
CARRIER-.
S >"
Figure 25. Thermal desorption inlet-manifold
-------
i
10.5 cm
10.56 mm. 1.0.
1
r
13.0 mm.O.D.
Figure 26. Thermal desorption chamber with annular space. Sampling
tube shown in lower figure.
73
-------
OESORPTION CHAMBER
T
15.0 cm
13.30 cm
il
TEFLON INSERT
COMPRESSION SPRING
D< PURGE GAS
GLASS CARTRIDGE SAMPLER
QT " ""_ Q I3.0mm ] 10.56mm
i~T~
< 10.5 cm M T
Figure 27. Thermal desorption chamber
74
-------
13.3 cm which accommodated a pyrex sampling cartridge 10.5 cm in length.
Two of the desorption chambers (Fig. 20 and 21) were designed to accept
cartridges 10.56 mm i.d. x 13.0 mm o.d. While the third accepted a
larger cartridge of 15.6 mm i.d. x 16.5 mm o.d.
An aluminum sandwich served as a heat sink (Fig. 25) which accepted
any one of the three desorption chambers. Two 150 w, 115V heating
cartridges (Varian Part No. 22-0000-18-00) were used to heat the aluminum
sandwich and the temperature was monitored and controlled with a platinum
sensor probe (100ft, Varian Part No. 64-000009). The desorbed vapors passed
via a short insulated capillary line through a six-port two position valve
which was also encased in an aluminum heating bath. Temperature control
was identical to the thermal desorption chamber. The electronics circuit
which controlled the temperatures on each heat sink is shown in Figure 28.
The temperature was monitored directly on a pyrometer; control was + 1°C.
A nickel capillary (0.020 i.d. x 0.032 x 0.5 m in length) constituted
one loop of the valve proper which was cooled with liquid N~ or solid
carbon dioxide/isopropanol and served as a trap for collecting and concen-
trating desorbed vapors for their introduction into high resolution glc
columns. The vapors were released from the capillary trap by rapidly
heating to 175° using a wax bath.
The multiport valve used on the described inlet/manifold was chosen
for its polyimide internal stem to minimize the contact of desorbed trace
vapors with reactive metal surfaces, therefore, mimimizing contamination
or decomposition of sample constituents.
In a typical thermal desorption cycle a sampling cartridge was placed
in the preheated (ca. 225°C) chamber, and N0 gas was purged through the
/.
cartridge (ca. 20 ml/min) to purge the vapors into the liquid N9 cooled
-------
I COW
HEATER
"*
Figure 28. Electronics circuit designed for temperature control on inlet-manifold system,
-------
Ni capillary trap; this constituted valve position "A" (Fig. 25). After
the thermal desorption step was complete, the six-port valve was rotated
to position "B" (Fig. 25) and the temperature on the capillary loop was
rapidly raised (> 10°/min) whereupon the carrier gas carried the vapors
onto a glc column.
As designed, the prototype thermal desorption chambers which were
easily interchangeable, accommodated cartridges of two different diameters
with up to 8 cm of packing (sorbent) depth. Thus, comparisons of dif-
ferent cartridge sizes with respect to sorbent background during thermal
desorption could be made. This inlet-manifold configuration also allowed
the desorbed vapors from one or more cartridges to be accumulated in the
capillary trap prior to analysis by glc or glc-ms.
The inlet/manifold system described here was employed in studies on
thermal desorption of vapors from cartridges. Its performance characteris-
tics are further discussed in Section VII.
HEAT TRANSFER CHARACTERISTICS FOR SELECTED SORBENTS AND THERMAL DESORPTION
CHAMBER
An investigation was made of the heating rates for candidate sorbent
media. Using the measured heat transfer coefficients as a guideline, the
required heating period and temperature for effecting quantitative desorp-
tion was selected (Section VII). Furthermore, two prototype desorption
chamber designs with and without an annular space (Fig. 26 and 27) were
compared with regard to the rate of heat transfer to the sorbent bed.
Experimental
Tenax-GC (60/80 mesh), Chromosorb 101 (100/120), Chromosorb 104
(100/120) and Chrouosorb W-HP (100/120) were purchased from Applied Science,
State College, Pa. Stationary phases chemically bonded to supports which
77
-------
included carbowax 400/Poracil C (100/120) and oxopropionitrile/Poracil C
(80/100) were also obtained from Applied Science. Carbowax 600 and didecyl
phthalate stationary phases and the sorbent Porapak Q were from Supelco,
Inc., Beliefonte, Pa.
Carbon derived from coke (PCB, 12/30) was acquired from Pittsburgh
Activated Carbon Division of Calgon Corp., Pittsburgh, Pa. A cocoanut
activated carbon (580-26) was purchased from Barneby Cheney, Columbus,
Ohio.
The temperature rise times in the sorbent bed of the cartridge was
measured with a calibrated thermocouple (mV ys °C); signal output was
directly recorded on a strip chart recorder (Varian Model A-25, Varian
Instruments, Walnut Creek, CA.). The rate of temperature increase in
the sorbent was determined using preset and isothermal desorption chamber
temperatures. The rate of temperature increase was also monitored on the
inner glass wall of the cartridge.
Results and Discussion
Prior to defining the parameters (temperature, heating time) for ef-
fecting thermal desorption of vapors trapped on sorbents (Section VII), the
rate of heat transfer from the thermal sink to the sorbent bed was measured
using the inlet-manifold interface depicted in Figure 25. Operating the
desorption unit under isothermal conditions, the temperature increase in
the sorbent bed was monitored immediately after inserting a cartridge.
Figure 29 depicts the thermocouple response time, and the heating rate
in the center of a cartridge containing Tenax GC using one of the previously
described thermal desorption chambers (Fig. 27). Because the response of
the thermocouple significantly contributed to the temperature rise profile
(8 sec required to reach 63% of the upper temperature limit), its contribution
78
-------
o
o
UJ
-------
was subtracted from all measurements. These temperature rise times
(Fig. 29) indicated a cross-sectional gradient was produced immediately
after inserting the cartridge into the chamber. The periphery of the
sorbent bed reached the temperature maximum in approximately 4 min, while
the center of the sorbent bed required an additional 2 min. Furthermore
increasing the cartridge diameter from 1.0 to 1.5 cm increased the tem-
perature differential by a factor of 1.3.
The heat transfer coefficient for sorbents and similarly their observed
differential temperature gradient, varied considerably. The relative tem-
perature rise times for several sorbents which previously were shown to
have good collection efficiencies (Section V) were compared. The order of
their heating rates were observed to be: PCB and BPL carbon (12 x 30)
> oxopropionitrile and carbowax 400 chemically bonded to Poracil C (100/120)
> Chromosorb 104 (100/120) > Tenax GC (60/80) and Chromosorb 101 (100/120).
These results are depicted in Figure 30 which were obtained with a thermal
desorption chamber designed to have an annular space between the chamber
wall and glass cartridge. The largest differences in heating rates (in the
center of the sorbent bed) were exemplified by the activated carbons and
Tenax GC (Fig. 30).
Because inspection of the heating rates for Tenax GC (60/80) with and
without an annular space indicated relative differences (Fig. 29 and 30),
a more detailed study was conducted comparing these two thermal desorption
chamber designs. Figure 31 presents this comparison. This data clearly
shows that a 1 mm annular space in the chamber reduced the heating rate of
the inner glass cartridge wall as well as the center of the packing by at
least a factor of two. Furthermore, a larger heating gradient was observed
between the center and periphery of the sorbent bed with the annular spaced
80
-------
2461-
oo
25
O= Chromosorb 104
A= Chromosorb 101
D= Tenax GC
O= BPL carbon
A= PCB carbon
D-A= oxypropionitrile/Poracil C 80/100
• « 20% carbowax 600 on Chrom W (100/120)
01 234 56
TIME ( MIN )
Figure 30. Temperature rise times in sorbent bed using
annular spaced chamber.
-------
197 r
00
NJ
,-D-r£T.-A-
—. • A
temperature in center of cartridge,
with (O) and without (A) annular space
B « temperature of inner glass wall of
cartridge, with (/\) and without
annular space
Thermocouple response given by0
0 1234567
TIME ( MIN )
Figure 31. Comparison of temperature rise times for chambers with and
without an annular space. Chamber was at 175°C.
-------
chamber than the one providing direct contact between the brass wall and
glass cartridge.
In view of the results obtained with the prototype desorption chamber
with an annular space (Fig. 26), the heating rates for candidate sorbent
media were also investigated with a chamber providing direct contact be-
tween the glass cartridge and chamber wall (Fig. 27). In general the
trend was the same as previously observed (compare Fig. 30 and 32) except
all heating rates were significantly greater in this case.
Since the heating rate was generally 1.7 times faster when the cart-
ridge was in direct contact with the brass desorption chamber wall and a
smaller temperature gradient was observed through the packing bed, it was
concluded that a chamber incorporating a 1 mm annular space 90% of the
length of the cartridge was not a feasible design for application to desorp-
tion of vapors. Thus, the prototype desorption chamber which introduced the
carrier gas near the top of the chamber (Fig. 27) was employed in thermal
desorption of vapors trapped on sorbents.
The temperature rise time for Tenax GC was also examined using three
isothermal conditions on the desorption unit (Fig. 33). The heating rate
(slope) was only slightly increased by increasing the desorption unit
temperature; however, it was evident that the rate was linear upto 65% of
the final temperature or during the first 75 sec. Thereafter, an additional
several minutes was required to reach a plateau. These data, therefore,
indicate that the desorption unit should be set at a temperature which allows
the attainment of the required desorption temperature in 60-90 sec after
insertion of the cartridge sampler.
When a thermal desorption chamber design permits rapid heating of the
cartridge sampler with a minimum of temperature gradient across the packing
83
-------
220 r
3 4
TIME (MIN.)
Figure 32. Comparison of heating rates for some sorbents. Curves
A, B, and C correspond to PCB carbon (12/30), oxopro-
pionitrile on Poracil C (80/100), respectively. Thermal
desorption chamber was isothermal at 210°C.
84
-------
230 r
205 -
A = 180°C
B = 205°C
C = 228°C
0 24 6
TIME (MIN)
Figure 33. Heating rate for Tenax GC at different isothermal
desorption chamber temperatures.
85
-------
bed then the desorbed vapors may be introduced directly onto a conven-
tionally packed glc column. The carrier gas flow rate was satisfactory for
efficiently purging the desorption chamber and maintaining column resolu-
tion. Thus, the desorption chamber can be "in-line" with the glc column
for a time sufficient to desorb substituents with the greatest adsorption
affinities and then returned to the "by-pass" mode during the remainder
of the chromatographic period. Under these conditions the background intro-
duced during heating of the sorbent can be minimized as well as artifactual
processes resulting from decomposition, polymerization, etc. of solute vapors.
The use of cartridges of <_ 1.0 cm i.d. and the prototype desorption chamber
shown in Figure 27 allowed the inlet-manifold to be operated in the manner
described.
With larger diameters, the temperature gradient in the sorbent bed was
too great. In this case, a cartridge "in-line" with the glc column during
the desorption cycle produced excessive solute band spreading and sample
resolution was decreased. It was concluded that desorption was not uniform
across the sorbent bed because of the large temperature gradient which pro-
bably accounted for the loss in glc resolution.
When sampler cartridges of > 1.0 cm i.d. were subjected to thermal
desorption or high resolution capillary columns were employed, the vapors
were concentrated in a small carrier gas volume in order to prevent the
excessive band spreading and decreased column efficiency. This was achieved
as previously described by retrappiiig desorped vapors in a Ni capillary
(0.020 in i.d. x 0.5 m length) using liquid N • as the coolant. After the
desorption period, the carrier gas was routed through the capillary trap
(Fig. 25) and the trap rapidly heated.
86
-------
SECTION VII
THERMAL DESORPTION OF HAZARDOUS VAPORS FROM SOLID SORBENTS
The recovery of vapors adsorbed or absorbed on various sorbents by
"I ft 01 OQ / *% 7 /A
thermal ' and vacuum desorption has been reported. Duel
employed a combination of vacuum-thermal stripping to remove pollutants
adsorbed to cocoanut charcoal. The sample was heated from ambient to
170°C at 10°/min; partial fractionation was accomplished prior to glc
32
analysis. Damico also used cocoanut charcoal to trap glc fractions
for ms analysis. Desorption of propionaldehyde and 2-nonanone occurred
at room temperature and 70°C, respectively, when loaded capillaries were
introduced into a high vacuum of the mass spectrometer. Other investiga-
tors also reported the thermal desorption of vapors from charcoal; * ' '
however, analysis of pollutants had been restricted primarily to aliphatic
and relatively nonpolar aromatic compounds.
The high surface activity of activated carbons has been reported to
produce artifacts during recovery. Formaldehyde decomposes, as does methyl
ethyl ketone which forms diacetyl compounds and acetic acid. Furthermore,
compounds sensitive to polymerization or decomposition on carbon are also
20 21
generally sensitive to carbosieve '
Desorption of semi-polar and polar compounds by thermal means has been
., i j r i * ^ j 15,17,20,21,38,40 „.,,, 17
successively achieved from polymeric beads . Williams
used temperature programming up to 210°C to elute trace contaminants from
38
Porapak columns. Leggett, e_t^ al_. reported significant amounts of con-
taminants from Porapak was produced if the temperature exceeded 110°C
during analysis. Similar results were obtained by Krumperman when
Porapak Q cartridges were heated above 170°C.
87
-------
20 21
In contrast, Zlatkis, e£ jal. ' were able to desorb many volatile
polar urine metabolites as well as atmospheric pollutants from Tenax GC
at 300°C. The background from this polymer was extremely low.
Thermal desorption of trace vapors absorbed on liquid phase coated
31
beads was extensively employed by Williams for analyzing atmospheric
pollutants.
Although there are many reports on the use of thermal desorption
as a means for recovering and introducing the vapors into a glc, a thorough
study has not been made on the quantitative aspects of this method for
semi-polar and polar chemical classes of compounds such as carcinogens.
This section discusses (1) the background contribution from sorbents,
(2) the desorption parameters (temperature, time) for effecting recovery
for sorbents, (3) the quantification of the thermal desorption step, and
(4) the percent recovery of hazardous substances of interest to this research
program from Tenax GC.
EXPERIMENTAL
Tenax-GC (60/80 mesh), Chromosorb 101 (100/120), Chromosorb 104 (100/120)
and Chromosorb W-HP (100/120) were purchased from Applied Science, State
College, Pa. Stationary phases chemically bonded to supports which included
carbowax 400/Poracil C (100/120) and oxopropionitrile/Poracil C (80/100)
were also obtained from Applied Science. Carbowax 600 and didecyl phthalate
stationary phases and the sorbent Porapak Q were from Supelco, Inc., Belle-
fonte, Pa.
Carbon derived from coke (PCB, 12/30) was acquired from Pittsburgh
Activated Carbon Division of Calgon Corp., Pittsburgh, Pa. A cocoanut
activated carbon (580-26) was purchased from Barneby Cheney, Columbus,
Ohio.
88
-------
All sorbents were thermally conditioned 10°C below the maximum recom-
mended temperature limit for at least 12 hr under approximately 20 ml/min
of He flow. After sorbents were packed into glass cartridges they were
conditioned again for 15 min in the thermal desorption unit prior to use.
The standards-ethyl methanesulfonate, 3-propiolactone, N-nitrosodiethyl-
amine, 1,2-dichloroethyl ethyl ether, nitromethane, methyl ethyl ketone,
and aniline - were from Fisher Chemicals, Pittsburgh, Pa. The source of
glycidaldehyde and sulfolane was Aldrich Chemicals, Milwaukee, Wise. The
supply of 1,3 propanesultone, maleic anhydride, butadiene diepoxide and
propylene oxide was from Eastman Organic Chemicals, Rochester, N. Y.
Styrene epoxide, bis-(chloromethyl)ether and bis-(2-chloroethy1)ether were
acquired from K&K labs., Plainview, N. Y.
In order to demonstrate the efficiency of collection plus thermal
desorption of trapped vapors, synthetic air/vapor mixtures were prepared
using N-nitrosodiethylamine (100 ng) as an internal standard; the quantity
of the other vapors tested was varied from 50 to 300 ng. An aliquot of
this mixture was introduced directly through the thermal desorption cham-
ber which contained a glass cartridge packed with only glass wool and
the mixture resolved by glc. The peak areas for each solute in this
calibration mixture (cm) was measured by triangulation and the relative
response ratios were calculated:
^cm = Ap/Ail x n* i? (5)
where A and A were the areas of the solute peak and of N-nitrosamine,
respectively. An identical aliquot of the synthetic air/vapor mixture was
purged (4.0 1/min) through a cartridge containing a sorbent using the moni-
toring system described earlier (Figure 25). The trapped vapors were
89
-------
desorbed in the thermal desorption chamber followed by glc analysis. The
RR for each constituent was calculated and the percent recovery was deter-
s
mined as a ratio of the RR values to those for RR .
s cm
RR
, /RR x 100% = Percent recovery (6)
sorbent cm
Gas-liquid chromatography (glc) was conducted on a Perkin-Elmer 900
series chromatograph (Perkin Elmer Corp., Norwich Conn.) equipped with dual
flame ionization detectors. A 2.5 mm i.d. x 3.6 m silanized glass column
containing 2% DECS on Chromosorb W(HP) 80/100 mesh was used for resolving
synthetic air/vapor mixtures. The column was programmed from 55 to 200°C
at 10°/min with an initial and final isothermal period of 2 and 10 min,
respectively. Carrier gas (N«), hydrogen and air flow rates were 45, 30
and 250 ml/min, respectively. The injection port, manifold and detector
temperatures were maintained at 250°C.
RESULTS AND DISCUSSION
Prior to evaluating the desorption step of trace organics from sorbents,
the candidate sorbents were subjected to thermal desorption temperatures;
the background contribution from each was examined. The polymeric beads-
Porapak Q, Chromosorb 101 and 104-all exhibited significant background even
when conditioned using the manufactured recommended procedures. Of these
three, Porapak Q was the worst offender (Fig. 3A). These results confirm
1 f\ *^ft
those previously reported by other investigators ' . Polymeric beads were
extracted in a Soxhlet for 18 hr with acetone, methanol, or benzene and then
thermally conditioned; all attempts to reduce the background of Porapak Q
were unsuccessful. Background was reduced significantly for Chromosorb 101
when extracted with methanol but this procedure was less successful for
90
-------
Parameters
VO
20
Col - 55° - 185°
initial T° - 2 min delay
program - 13°/min
Flow rate - 40 ml/min
2% DECS Chrom W (HP) - 100/120
Thermal Desorption - 6' @
By-pass Loop - open 2'
Sorbent-Porapak Q
I
I
18
16
14
12 10 8
TIME (MIN)
90
80
70
60
50
<
o
40
30
20
10
u.
o
CJ
-nrrT-c'- of blank Porarjak-Q cartridge.
-------
Chromosorb 104. Although Porapak Q exhibited high collection efficiencies,
its background precluded any studies requiring trace analysis.
On the otherhand background contamination from Tenax GC beads was
very low (Fig. 35). Pre-extraction with methanol for 18 hr followed by
thermal conditioning at 325°C produced cartridges which allowed nanogram
quantities of hazardous vapors to be detected and quantitated.
Sampling cartridges packed with any one of the activated carbons
also exhibited low background (Fig. 36).
In contrast, inert glc supports coated or chemically bonded with liquid
phases gave relatively high contaminant peaks; the latter was somewhat better.
Procedures for treating these candidate sorbents to yield cartridges with low
background was not exhaustively investigated. Further studies in this area
are warranted, especially with phases chemically bonded to supports.
Because the desorption experiments indicated that the background con-
tribution from Tenax GC was least of the sorbents tested, and it exhibited
excellent collection efficiencies for selected hazardous substances (Section
V), the thermal recovery of vapors adsorbed to this polymer was examined.
Ten compounds which represent a broad spectrum of chemical properties and
are of particular interest in air pollution studies were chosen. Each sub-
stance was introduced through a cartridge of Tenax GC (1.0 cm i.d. x 3.0 cm
in length) at levels of 50, 100, 200, and 300 ng as a synthetic air/vapor
mixture.
Initial experiments were designed to determine optimal recovery of
solutes using temperatures. The principle reason was to use sufficient
temperature for vaporizing the trapped constituents while minimizing sor-
bent background and decomposition of labile compounds or inter-compound
92
-------
GLC Parameters
FID-GLC
Col. - 12 ft 2% DECS Chrom W-HP
Program - initial hold 2 min, 70-180° @ 10°/min
Detector - 200°
Attenuation - 6.4 x 10~ AFS
Thermal Desorption Parameters
Chamber - 175°
6-port Valve - 70°
Cartridge Preheat Period - 5 min
By-pass Open - 2 min
90
80
70
60
ui
fe
I-
LJ
50 flc
40g
0.
UJ
tr
30 oc
UJ
o
tr
o
o
20 £
10
16
14
12
10 8
TIME (min.)
Figure 35. Background during thermal desorption of Tenax GC
Cartridge Blank
93
-------
GLC Parameters
FID-GLC
Col - 12 ft 2% DECS Chrom W-HP
Program - initial hold 2 min, 70-180° @ I0°/min
Detector - 200° ...
Attenuation - 6.4 x 10 AFS
Thermal Desorption Parameters
Chamber - 175°
6-port Valve - 70°
Cartridge Preheat - 5 min
By-pass Open - 2 min
14
12
10 8
TIME (min.)
Figure 36. Background during thermal desorption of PCB carbon
cartridge blank.
94
-------
reactions. At a temperature of 125° relatively little amounts of vapors
were desorbed from Tenax GC; 30-40% recovery was obtained at 175°C. When
the desorption chamber was raised to 200°C, approximately 80-90% recovery
was observed for Mixtures I, II and III (Section V).
Quantitative thermal desorption was achieved at 225°C in 90 sec. A
comparison of the glc analysis for an aliquot of the synthetic air/vapor
mixture used for loading a cartridge and vapors desorbed from Tenax GC is
shown in Figures 37 and 38. It was concluded that none of the vapors
studied had decomposed during the thermal desorption step since the chroma-
tograms were essentially identical to those obtained for mixtures directly
injected into the glc.
The percent recoveries are given in Table 18. Except for nitromethane,
all of the substances examined were quantitatively recovered. Accuracy for
duplicate analysis was + 2%.
In contrast to the results obtained for Tenax GC, attempts to desorb
these vapors from the activated carbons was not achieved even when tempera-
tures up to 330°C were used; higher temperatures began to exhibit chromato-
graphic peaks with retention indices different from the parent compounds
suggesting that decomposition was occurring.
-------
B
ao
6.0
s
X
u
4.0
a:
cc
P
a
2.0
8
6
TIME ( MIN)
145
135
125
115
105
r
95
85
75
TEMPERATURE (°C )
65
Figure 37
Gas-liquid chromatogram of synthetic air/vapor mixture of
hazardous substances. Peaks A, B, C, D, and E are 300 ng
of glycidaldehyde, butadiene diepoxide, N-nitrosodiethyi-
amine, 1,2-dichloroethyl ethyl ether, and ethyl methane
sulfonate, respectively. See text for glc parameters.
96
-------
4
TIME (WIN)
145
135
125
115
105 95
TEMPERATURE (°C)
85
75
9.0
8 JO
7.0
6.0
,0 2
4.0
3.0
2.0
1.0
in
CO
a:
ui
o
65
Figure 38. Gas-liquid chromatogran of vapors desorbed from Tenax GC.
Desorption chamber was 225°C; see prior figure for peak
identity. Background from Tenax GC is represented by
dashed profile.
-------
Table 18. PERCENT RECOVERY OF VAPORS ADSORBED ON
TENAX GC CARTRIDGES USING THERMAL DESORPTION3
Quantity Adsorbed (ng)
Compound
£
N-nitrosodiethylamine
B-propiolactone
ethyl methanesulfonate
nitromethane
glycidaldehyde
butadiene diepoxide
styrene epoxide
aniline
Bis (chloromethyl) ether
Bis- (2-chloroethyl) ether
50
100
105
105
-
100
100
100
95
100
95
100
100
100
100
-
100
100
100
95
100
90
200
-
100
95
70
95
100
105
95
100
90
300
-
100
100
70
80
100
90
60
90
-
Tenax GC cartridge - 10.5 mm i.d. x 30 mm in length. Synthetic air/
vapor mixtures were introduced onto a Tenax GC bed @ A 1/min. Desorp-
tion Unit was at 225°C.
^Represents theoretical amount in synthetic air/vapor.
cValues, an average of duplicate runs, were calculated on basis of a ratio
of peak areas for calibration mixture and from thermal desorption.
98
-------
SECTION VIII .
DESIGN AND PERFORMANCE OF A FIELD SAMPLER
Collection of pollutants by other investigators has been performed
at modest flow rates (20-2000 ml/min) because the concentrations of the
substances sought were relatively high. In contrast, the sampling rates
and time required in this research program were much greater since the
3
hazardous vapors were anticipated at the low ng/m levels. This relation-
ship was previously shown in Table 2 (Section IV).
In order to collect sufficient quantities of each atmospheric carcino-
gen for instrumental analysis, the field sampling unit had to meet several
3
requirements. These were: (1) a sampling rate adjustable from 0-3 M /hr
at the pressure drop encountered with a sampler cartridge in-line, (2) a
capability of multiple cartridges on-line during a sampling period, (3)
uninterrupted 24 hr operation, and (4) the opportunity to "push" or "pull"
air samples through the cartridge sampler. All of these factors ultimately
determine the power (milliamps/1/min) required for sampler operation.
Also, these and additional factors were not independent of each other.
Other limitations were imposed by: (1) the collection efficiencies of the
packing, (2) sorbent breakthrough characteristics, (3) the lowest detectable
concentrations of the carcinogenic compounds and (4) the size and shape of
the cartridge sampler and how pressure differential increases with increased
flow through it. Because contamination of a sample was to be avoided, the
pump design was also important. Of these criteria considered in designing
a field sampling unit, the pressure differential developed across a cartridge
at a specified flow rate was the most important.
99
-------
Questions regarding multiple sampling, continuous or pulsed flow were
also considered. Flexibility, portability and durability were features
sought in the unit.
The specifications for the design of a field sampling unit consisting
of a pump, multiport manifold, cartridge samplers, valves and flow indica-
tors were guided by all of the stated criteria.
This section presents a discussion of these factors individually and
combined since many are interdependent.
SAMPLE VOLUME
For the purpose of developing a sampling system it was assumed that
the amount of a compound required for its identification by a technique
such as gas chromatography-mass spectrometry is 30-50 ng and that the sor-
bent was quantitative in collecting the vapors at low concentrations. The
relationships between the ambient vapor concentration and the total volume
of air that must be pumped through the cartridge, and thus, the time re-
quired for sampling, at various flow rates were given in Table 1. The total
power required for sampling was directly related to sampling rate and dura-
tion.
POWER REQUIREMENT
The power required to pump air through sorbent-packed tubes depends
upon the pressure differential across the tube under flow conditions, which
in turn depends mostly upon the air flow-rate, the shape of cartridge
(diameter and depth of packing), the particle size distribution, and the
particle shape of the sorbent, and to a lesser extent upon the air tempera-
ture and humidity. A method for calculating power requirements from pres-
sure drop values was developed for this study.
100
-------
From the values of p and ?„ in equation (4) , the theoretical power •••?$.
(watts) required for compression or expansion of the air (compression if
the sampler were downstream from the pump) and its delivery through the
sampler was estimated using a formula derived from the Moss and Smith
equation for adiabatic horsepower. The formula employed was:
k-1
Power (Watts) = SL/m x 5.968 x [ (pp k -1] (7)
where H/m = liters per tnin, p. and p_ are the high and low pressures across
the sampler, and k is the ratio of specific heats for air, c and c . This
formula is based upon air at 14.7 psi, 23°C and 36 percent relative humidity;
3
its density was taken at 0.075 Ib/ft ; k was set at 1.3947.
Figure 39 shows the theoretical power required to pump air through
5-cm depths of sorbent packing. For a flow-rate of 20 liters/min, the
change from 60/80 mesh particles to 18/20 in the 1.06 cm tube reduces power
requirements by almost a factor of 5 (30.4 watts to 6.4). Increasing the
tube diameter from 1.06 cm to 1.82 cm decreases power for the 60/80 mesh
packing from 30.4 to 21 watts. From these data it was concluded that tube
diameters of approximately 1.5 cm, particles of about 35/60 mesh, and bed
depths of about 5 cm will allow sampling rates of 20 liters /minute for a
theoretical power consumption of 15 watts (0.02 horse power) . Practical
power requirements are expected to be at least twice the theoretical values
to allow for power losses in the pump itself. For example, a cartridge of
Tenax GC, 35/60 mesh, in a 1.82-cm tube would require twice the theoretical
7,4 watts, or 14.8 watts to sample air at 20 £/tn.
While the power requirements can be supplied by almost any 115 V,
50-60 Hz source, the power ratings of portable battery powered samplers
would have to be increased substantially to permit their use at low
101
-------
lOOr
o
NJ
O.I
D (Cm) * U. S. Mesh
P
Tube i.d. (cm)
A 0.0211
O 0.0211
A 0.035
80.093
0.035
60/80
60/80
35/60
18/20
35/60
1.06
1.82
1.82
1.06
1.49
10
100
POWER ( WATTS )
Figure 39.
Relationship between flow rate and theoretical power requirements
at various tube diameters and particle size.
-------
concentrations and sampling rates. The required 6-volt battery ratings for
a sampler with a 5-cm depth of Tenax GC 30/65 in a 1.8 cm tube are given
in Table 19.
With the exception of nickel-cadmium batteries, which provide up
to 45 mA for 10 hours, most of the sampling would therefore involve the
use of several batteries, possibly, in parallel, with recharging at regular
intervals, determined by the specific case. The use of battery operated
3
samplers in detecting ng/m concentrations of organic vapors depends upon
the effectiveness and capacity of sorbents with particle sizes in the 12/30
mesh range. Pressure-drops at high flows (20 &/m) must be kept low to
avoid vacuum desorption if the air sample is drawn through the cartridge.
FLOW VECTOR REQUIREMENTS
Whether the pumping unit would best serve in the capacity of pulling
or pushing the atmosphere through the cartridge also depended upon the
pressure differential developed. For instance, if a large pressure drop
Table 19. POWER REQUIREMENTS TO DELIVER VARIOUS SAMPLING RATES
Sampling Rate
A/min
Power Needed
Watts
Six-Volt Battery Capacity
Required Per Sample
millamp-hr
ng/M3
1
4
9
20
0.022
0.34
2.06
14.8
1
1800
7500
19150
61500
10
180
750
1915
6150
100
18
75
192
615
1000
1.8
7.5
19.2
61.5
103
-------
(> 100 mm Hg) was required to attain the desired sampling flew rates
then a "pull" system may significantly decrease the collection efficiency
profiles for each sorbent (vacuum stripping). Since collection efficien-
cies in our studies have been acquired under positive pressure ("push")
then a "push" sampling system would deliver the best correlation between
field and laboratory trials.
A continuous flow pumping unit was preferred to a pulse system since
the latter may have disturbed the packing material in the cartridge and/or
differ in trapping characteristics from those observed by our continuous
flow system in laboratory experiments. A separate examination of collec-
tion efficiency and breakthrough volumes would be necessary in order to
determine the merits of pulse sampling. Such a study was not conducted.
REPLICATE SAMPLING
The uncertainties inherent in sampling for the hazardous organic vapors
imposed a need for a versatile system for multiple sampling, i.e., taking
several simultaneous samples. For example, acquisition of duplicate samples
simultaneously circumvented the problem associated with diurnal fluctuations
of vapor concentrations occurring when duplicates were obtained sequentially.
Furthermore, different collection media could be compared by sampling the
same atmosphere concurrently. The ability to collect with several car-
tridges at one time was conducive to overall shorter field sampling periods
as well as comparison of sample duration ys breakthrough (Section V).
To draw 25 Jt/m of air through two sets of two samplers in parallel,
each containing a 3-cm depth of 0.035-cm diameter adsorbent in a 1.06-cm
2
diameter tube, the pump must draw 50 £/m at a pressure drop of 8.34 N/cm
(626 nan Hg). On the other hand it was also desirable to use cartridges
in tandem to determine whether breakthrough has occurred. The Universal
104
-------
Sampler 5 1068 (Research Appliance Co., Allison, Pa.) met these requirements
and has been used in our field sampling studies. Its operation required
approximately 1/2 h.p. at 110 V, 50-60 Hz. This unit is shown in concept
in Figure 40.
VACUUM GAUGE
CRITICAL ORIFICE SET
C3-
GAS VACUUM
METER RECORDER
RUNNING TIME
METER
Figure 40. Schematic of Universal Sampler 5-1068
The "pull" sampler would go at "A"; the push sampler, if employed,
at "B". This system consists of those elements that will provide control
of the sampling at desired rates. The system was quantified for a range
of flow rates up to 25 liters per minute, using the sampler employed in
testing adsorbents (a 3 cm deep packing of 0.035 cm diameter adsorbent
in a 1.06 cm diameter tube). Calculations showed that, to draw (or push)
25 1/m of air through two sets of two samplers in series, the pump must
2
draw 50 1/m (1.76 cfm) at a pressure drop of 8.34 N/cm (12.1 psi).
105
-------
In order to accommodate the use of multiple cartridges during a sam-
pling period, a multiport chamber was designed and fabricated (Fig. 41).
The entire chamber was constructed of Teflon^. Six ports were located
on the chamber equidistant from one another (60°) so that any multiple of
1,2,4, and 6 sampling cartridges could be employed simultaneously without
experiencing different drawing rates since a symmetrical geometry was main-
tained. The air was drawn through a glass fiber filter (to remove "particu-
lates") and the cartridge, into the multiport head and then through the
pumping system. The sampling rate through the cartridges was regulated
by a bleed valve located on the multiport chamber. A vacuum/pressure gauge
on the chamber was used to monitor the pressure differential and by con-
sulting the pressure drop curves, a AP was selected and imposed in the cham-
ber which produced the desired sampling rate through each cartridge.
The multiport chamber was utilized with the Universal sampling pump.
Table 20 depicts the maximum sampling rates which this system is capable
of achieving under various conditions (bleed valve closed). When four
cartridges of Tenax GC (60/80) are compared to the same number of Chromo-
sorb 101 (100/120), the vacuum required to draw equivalent rates per cart-
ridge (and therefore total flow) was increased by 75 mm Hg. This observation
corroborates the experimentally determined AP for various particle diameters.
The magnitude of the vacuum required to achieve a prescribed sampling rate
may ultimately influence the performance of the sorbent. Decreased collec-
tion efficiencies and/or elution volumes may be experienced.
The design parameters and systems described in this and previous sec-
tions were applied to the collection of trace quantities of hazardous
pollutants in ambient atmospheres.
106
-------
CROSS- SECTION
VACUUM PRESSURE
GAUGE
TEFLON -
CHAMBER
GLASS FIBER FILTER
'S/SS/S//W//////A
'////S//////, I 'MSS///
PRESSURE CONTROL
VALVE
GLASS COLLECTION CARTRIDGE
CARTRIDGE HOLDER
TO PUMP
TOP VIEW
Figure 41. Multiport sampling head,
107
-------
Table 20. SAMPLING RATE CHARACTERISTICS FOR UNIVERSAL SAMPLER WITH MULTIPORT HEAD
Sorbent
Tenax GC
(35/60)
Tenax GC
(60/80)
Chromosorb 101
(100/120)
b
No. of Cartridges
1
2
4
6
1
2
4
6
1
2
4
6
Sampling Rate
1/min/cartridge m^/hr/ cartridge
77.0
41.5
21.5
14.7
76.5
41.2
21.6
14.6
74.7
40.5
21.4
14.5
4.62
2.49
1.29
0.88
4.59
2.47
1.29
0.88
4.49
2.29
1.28
0.87
Total Volume
m3/hr
4.62
4.98
5.15
5.28
4.59
4.94
5.18
5.27
4.49
4.48
5.13
5.23
.The maximum pumping rate (no flow restriction) for this commercial system is 89 1/min (5.28 m^/hr)
Cartridge dimensions were 1.56 cm i.d. x 6.0 cm in length.
-------
SECTION IX
APPLICATION OF DEVELOPED INSTRUMENTATION AND
METHODOLOGY TO THE ANALYSIS OF AMBIENT AIR
In previous sections of this report detailed experimental design
and techniques were described for collecting and analysis of hazardous
vapors. Preliminary results on the application of this methodology and
instrumentation to field sampling and the analysis of air samples by
combined gas-liquid chromatography/mass spectrometry/computer are dis-
cussed in this section.
EXPERIMENTAL
Site of sampling
An urban site which presumably was concentrated with automobile ex-
haust in the presence of strong sunlight was chosen. An environment con-
ducive to photochemical smog is known to exhibit high levels of ozone,
NO. and NO . Two sites which had a history of high levels of ozone were
£ A
at the CHAMP stations in Santa Monica and West Covina, CA . Furthermore,
West Covina during the month of April, 1974 also had appreciable amounts
of N0_ and NO . On the basis of the presence of these pollutants and all-
£» X
phatic and aromatic hydrocarbons in air, the possibility of the formation of
epoxides and nitrosamines forming in ambient air was suspected to occur.
Sample Collection
Ambient air samples were collected with a Universal Sampler 5-1068
equipped with a multiport head as described in Section VIII. The sor-
bents selected for the collection of pollutants were Tenax GC and Chromo-
sorb 101 which was based upon their performance described in Section V.
The samplers (cartridge holders plus glass cartridges) were prepared in
109
-------
Inlet-
Hanilold
mem
GLC
-
1
Separator
Figure 42. Gas-liquid chromatograph-mass spectro-
meter computer (GLC-MS-CGti) outlay.
Ill
-------
pattern total maximum and minimum m/e peak intensity, and standard deviation
from calibrated m/e, and (2) an electrostatic plot of total ion current plots
and/or normalized mass spectra.
The operating parameters for the glc-ms-comp system for the analysis
of cartridges containing trapped pollutants are given in Table 22. These
conditions were used throughout this entire study.
RESULTS AND DISCUSSION
Figure 43 represents a total ion current plot from the mass spectro-
meter for an air sample from West Covina, CA. This profile was typical
for samples also collected in Santa Monica. The identity of several peaks
are given in Table 23 for this sample and Table 24 for a Santa Monica sam-
ple. The major constituents were aromatics and aliphatics. One oxygenated
compound, tenatatively,identified as styrene epoxide in this sample, was not
detected in samples from Santa Monica. Styrene, itself, was likewise not
detected; however, methylvinyl benzene was present. Styrene was reported
20
in Houston air by Bertsch et^ all. ; styrene epoxide was not. The discovery
of styrene epoxide was the most significant finding in this entire study.
Mass spectra were also obtain of the background from Tenax GC (35/60);
it was concluded that ethylene oxide was an artifact generated from thermal
desorption of this sorbent.
Because it was evident that the majority of the constituents were ali-
phatics and aromatics, single ion plots of 71 and 85 (Fig. 44) and 91, 105,
120, and 134 (Fig. 45) were obtained to demonstrate their distribution pat-
tern throughout a chromatogram. The low intensities of the aliphatic ions
(Fig. 44) indicated that some selectivity was exhibited by the Tenax GC
cartridge, i.e. it was less effective in trapping aliphatic compounds than
aromatics or polar pollutants. This selectivity was desirable because the
112
-------
Table 21. PROTOCOL FOR SAMPLING AMBIENT AIR IN LOS ANGELES, CA
Experiment Cartridge Type
1
2
3
Tenax GC (35/60)
1.5 cm i.d. x 6 cm
A
B
Tenax GC (35/60)
1.0 cm i.d. x 3 cm
A
B
C
D
Chromosorb 101 (60/80)
A
B
C
D
Sampling parameters
time (hr)
5
5
3
3
3
3
3
3
3
3
rate (1/min)
24
24
12
12
12
12
12
12
12
12
volume (m^)
6.6
6.6
2.07
2.07
2.07
2.07
2.07
2.07
2.07
2.07
Location Remarks
c1—.— *.-* it/ A_ .j .« .« /"* A
West Covina + internal
standards-
MEK & NB
West Covina + internal
standards-
MEK & NB
-------
UJ
UJ
>
UJ
UJ
tr
tr
o
16 17
O
i
mi |i|i|i M|i|i|rp I'M I MM
8700 8750 8800
885O 89OO 8950
MASS SPECTRUM NO.
'MI'I'I'IMM1!1!1!1!1!1!1!
90OO
it i i
50°
| i i i i i i i i iii i i i i i i iii i i i i i
80 110 140 170 200 23O
COLUMN TEMPERATURE ( °C )
Figure 43. Total ion current plot during gas- liquid chromatography mass
spectrometry of air sample from West Covina, CA. See text
for conditions.
-------
Table 22. OPERATING PARAMETERS FOR GLC-MS-COMP SYSTEM
Parameter Setting
Inlet-manifold
desorption chamber 225°C
valve 125°C
Capillary trap - minimum -195°C
maximum +175°C
thermal desorption time 4 min
GLC
OV-17 Ni capillary, 450 ft
column 50-230°C, 4°C/min
carrier (He) flow ~3 ml/min
transfer line to ms 210°C
MS
scan range m/e 20 -»• 300
scan rate, automatic-cyclic 1 sec/decade
filament current 300 yA
multiplier 6.0 ,
ion source vacuum ~4 x 10 torr
115
-------
Table 23 (continued). . POLLUTANTS IN AMBIENT AIR FROM WEST COVINA, CA
Peak No. RRT Name
32 2.030 dimethylethylbenzene
33 • 2.035 dimethylethylbenzene
34 2.059 methyldiethylbenzene
35 2.070 n-tridecane (tent.)
36 2.150 unknown
37 2.162 CiAH30
38 2.178 tetramethylbenzene
39 2.233 p_-tolualdehyde
40 2.261 acetophenone
41 2.310 unknown
42 2.330 nitrobenzene
43 3.380 unknown
44 3.530 naphthalene
45 3.561 unknown
117
-------
Table 24 (continued). POLLUTANTS IN AMBIENT AIR FROM SANTA MONICA, CA
Peak No. RRT Name
33 2.161 acetophenone
34 2.330 nitrobenzene
119
-------
to
o
m/e 71
m/e 85
r i ' r r i ' r r r i1 rr r r PT r rTTTrirm
50 100 150 200
MASS SPECTRUM NO.
250
• r j1 1 '
300
I i r I I i i i i i I i i I i I i r i i I r i I
SO 80 110 140 170
COLUMN TEMPERATURE (°C)
i | i i
2OO
r i 1 i
230
Figure A4. Single ion plots of ions common to aliphatic cracking series.
-------
m/e 134
m/e 120
m/e 105
m/e 91
I '
50
I
80
I
I
110
' ' ' I ' ' '
140 170 200
COLUMN TEMPERATURE (*C )
I
230
[rri 'I 'I'l'I'l'l' I 'I'M I1 IT1 [' I MM1 I1!1 I1 M PI 'I1 riTPIT(M MM
0 50 100 150 200 25O 300 350
MASS SPECTRUM NO.
Figure 45. Single ion plots for ions representative of aromatic cracking series,
-------
hazardous vapors which are to be collected and identified under this pro-
gram are considered to be semi-polar to polar.
In contrast the intensities of aromatic ions (Fig. 45) were relatively
large and constituted a high background between 140° to 230° which impeded
the detection and identification of possible hazardous vapors occurring at
trace levels. It was concluded that the chromatographic conditions should
be improved whereby the aliphatic and aromatic compounds are resolved as a
group from hazardous vapors of interest. The use of more polar stationary
phases and longer high resolution columns are currently under investigation.
122
-------
SECTION X
REFERENCES
1. Matz, J. The Importance of Nitrosamines for Communal Hygiene.
Z. Ges. Hyg. Ihre Grenzgebiete (Berlin). 18(12):903-8, 1972.
2. Norpoth, K., Manegold, G., Brlicker, R. and H. P. Amann. Investi-
gations on the Problem of the Release of Alkylating Compounds in the
Course of the Smouldering Process. Zbl. Bakt. Hyg. 156(B):341-52,
1972.
3. Jones, P. W. Analysis of Nonparticulate Organic Compounds in Ambient
Atmospheres. 67th Air Poll. Cont. Assoc. Mtg., Denver, Colorado.
Paper No. 74-265. June 1973.
4. Calvert, J. G. and J. N. Pitts, Jr. Photochemistry. John Wiley & Sons,
Inc., New York, 1966. pp. 366-557.
5. Dunham, C. L. (ed.). Biological Effects of Atmospheric Pollutants-
Particulate Polycyclic Organic Matter. Nat. Acad. Sci., Washington,
D. C., 1972. pp. 95-117.
6. Ibid. pp. 248-51.
7. Van Duuren, B. L. Carcinogenicity of Halo-ethers II Structure-Activity
Relations of Analogs of Bis-(chloromethyl)ether. J. Nat. Cancer Institute,
48(5):1431-39, 1972.
8. Van Duuren, B. L. Dimethylcarbamyl Chloride, A Multipotential Carcinogen.
J. Nat. Cancer Institute. 48(5):1539-40, 1972.
9. Van Duuren, B. L. Epoxides, Hydroperoxides, and Peroxides in Air
Pollution. Int. J. Environ. Anal. Chem. 1(3):233-41, 1972.
10. Van Duuren, B. L. Interaction of Some Mutagens and Carcinogenic Agents
with Nucleic Acids. Proc. Int. Symp., 1968. p. 149.
123
-------
11. Van Duuren, B. L. and G. Witz. "Phosphorescence Spectroscopy". In
Methods Pharmacol. C. F. Chignell, Ed. Appelton-Century-Crofts.,
Kew York, 1972. pp. 63-110.
12. Aue, W. A. Collection and Analysis of Organic Air Pollutant Trace
Substances. Environ. Health. 5:1-7, 1971.
13. Altschuller, A. P. Continuous Monitoring of Methane and Other
Hydrocarbons in Urban Atmospheres. J. Air Pollution Cont. 16:87-91,
1966.
14. Brooman, D. L. and E. Edgeley. Concentration and Recovery of Atmos-
pheric Odor Pollutants Using Activated Carbon. J. Air Pollution
Cont. Assoc. 16:25-9, 1966.
15. Jones, W. M. The Absorption of Benzene Vapor from an Air Stream by
..Beds of Charcoal. J. Appl. Chem. 16:345-9, 1966.
16. Krumperman, P. H. Erroneous Peaks from Porapak-Q Traps. J. Agr.
Food Chem. 20:909, 1972.
17. Williams, F. W. and M. E. Umstead. The Determination of Trace Con-
taminants in Air by Concentrating on Porous Polymer Beads. Anal. Chem.
40:2234-4, 1968.
18. Raymond, A. and G. Guiochon. Gas Chromatographic Analysis of C0-C-0
0 lo
Hydrocarbons in Paris Air. Environ. Sci. Techn. 8:143-8, 1974.
19. Hollis, D. L. Separation of Gaseous Mixtures Using Porous Polyaromatic
Polymer Beads, Anal. Chem. 38:309-16, 1966.
20. Bertsch, W., Chang, R. C. and Z. Zlatkis. The Determination of Organic
Volatiles in Air Pollution Studies: Characterization of Profiles.
J. Chromatog. Sci. 12:175-182, 1974.
124
-------
21. Zlatkis, A., Lichenstein, H. A. and A. Tishbee. Concentration and
Analysis of Trace Volatile Organics in Gases and Biological Fluids
with a New Solid Adsorbent. Chromatographia. 6:67-70, 1973.
22. Rasmussen, R. A. A Quantitative Cryogenic Sampler-Design and Operation.
American Lab. 4:19-27, 1972.
23. Rohrschneider, L., Jaeschke, A. and W. Kubik. Air Pollution at
Heights of up to 500 m above Industrial Sites. Chem. Ing. Tech.
43:1010-17, 1971.
24. Kaiser, R. E. Enriching Volatile Compounds by a Temperature Gradient
Tube. Anal. Chem. 45:965-7, 1973.
25. Bellar, T. A., Brown, M. F. and J. E. Sigsby, Jr. Determination of
Atmospheric Pollutants in the Parts-per-Billion Range by Gas Chroma-
tography. Anal. Chem. 35:1924-27, 1963.
26. Lonneman, W. A., Bellar, T. A. and A. P. Altschuller. Aromatic
Hydrocarbons in the Atmosphere of the Los Angeles Basin. Environ. Sci.
Technol. 2:1017-20, 1968.
27. Altschuller, A. P., Lonneman, W. A., Sutlerfield, F. D. and S. L.
Kopczynski. Hydrocarbon Composition of the Atmosphere of the Los
Angeles Basin-1967. Environ. Sci. Technol. 5:1009-16, 1971.
28. Aue, W. A. and .P. M. Teli. Sampling of Air Pollutants with Support
Bonded Chromatographic Phases. J. Chromatog. 62:15-27, 1971.
29. McEwan, D. J. Automobile Exhaust Hydrocarbon Analysis by Gas Chroma-
tography. Anal. Chem. 38:1047-53, 1966.
30. Farrington, P. S., Pecsok, R. L., Meeker, R. L. and T. J. Olson.
Detection of Trace Constituents by Gas Chromatography-Analysis of
Polluted Air. Anal. Chem. 31:1512-16, 1959.
125
-------
41. Scheutzl, D., Crittenden, A. L. and R. L. Charleston. Application
of Computer Controlled High Resolution Mass Spectrometry to the
Analysis of Air Pollutants. J. Air Poll. Cont. Assoc. 23:704-9,
1973.
42. Jeltes, R. Sampling of Nonpolar Air Contaminants on Porapak Porous
Polymer Beads. Atm. Envion. 3:587-8, 1969.
43. Mann, J. R. and S. T. Preston. Selection of Preferred Liquid Phases.
J. Chromatog. Sci. 11(4):216-220, 1973.
44. Tiggelbeck, D. Increasing Selective Efficiency in Cigarette Filter
Charcoals. Proc. 4th Int. Tobacco Sci. Cong., Athen, 1966. pp. 923-
44.
45. Saunders, R. A., Umstead, M. E., Smith, W. D., and R. H. Gammon.
The Atmospheric Trace Contaminant Pattern of Sea Lab II. Proc.
3rd Ann. Conf. Atmos. Contamination Confined Spaces, AMRL-TR-67-200,
1967.
46. Duel, C. L. Collection and Measurement of Atmospheric Trace Con-
taminants. Aerojet Electrosystems Co., Azusa, CA. Final Report Cont.
NAS 1-8714, NASA Doc. No. 71-19636.
47. Saunders, R. A. Analysis of Spacecraft Atmospheres. NRL Rept. 5316,
1962.
48. Chiantella, A. J., Smith, W. D., Umstead, M. E., and J. E. Johnson.
Aromatic Hydrocarbons in Nuclear Submarine Atmospheres. Am. Ind.
Hyg. Assoc. J. 27:186-92, 1966.
49. Saunders, R. A. Atmospheric Contamination in SEA-LAB I. Proc. Conf.
Atmos. Contamination Confined Sapces. AMRL-TR-65-230, 1965.
127
-------
50. Grob, K. and G. Grob. Gas-liquid Chromatographic-Mass Spectrometric
Investigations of C^-CL.. Organic Compounds in an Urban Atmosphere
An Application of Ultra Trace Analysis on Capillary Columns. J.
Chroiaatog. 62:1-13, 1971.
51. Jennings, G. and H. E. Nursten. Gas Chromatographic Analysis of
Dilute Aqueous Systems. Anal. Chem. 39:521-3, 1967.
52. Herbolsheimer, R., Funk, L. and H. Drasch. Viability of Activated
Carbon as Adsorbant for the Determination of Trichloroethylene in
the Air. Staub-Reinhalt. Luft. 32:31-3, 1972.
53. Perry, R. H. and C. H. Chilton. Chemical Engineers Handbook. McGraw
Hill, New York, N. Y., 1973. pp. 5-53.
54. Daniels, F. and R. A. Albertz. Physical Chemistry. John Wiley and
Sons, Inc., New York, 1962. pp. 607rll.
55. Langmuir, I. The Constitution and Fundamental Properties of Solids
and Liquids. Part I. Solids. J. Amer. Chem. Soc. 38:2221-95, 1916.
56. Langmuir, I. The Absorption of Gases on Plane Surfaces of Glass, Mica
and Platinum. J. Amer. Chem. Soc. 40:1361-1403, 1918.
57. Sune, J. Private Communication. Environmental Protection Agency,
RTP, NC.
128
-------
SECTION XI
LIST OF PAPERS ACCEPTED OR SUBMITTED FOR PUBLICATION
1. COLLECTION AND ANALYSIS OF TRACE ORGANIC VAPOR POLLUTANTS IN
AMBIENT ATMOSPHERES
I. A Technique for Evaluating the Concentration of Vapors by
Sorbent Media.
E. D. Pellizzari, J. Bunch, B. Carpenter and E. Sawicki
J. Environmental Science and Technology. Accepted for Publication.
2. COLLECTION AND ANALYSIS OF TRACE ORGANIC VAPOR POLLUTANTS IN
AMBIENT ATMOSPHERES.
II. Studies on Thermal Desorption of Organic Vapors from Sorbent
Media
E. D. Pellizzari, B. Carpenter, J. Bunch and E. Sawicki
J. Environmental Science and Technology. Accepted for Publication.
3. COLLECTION AND ANALYSIS OF TRACE ORGANIC VAPOR POLLUTANTS IN AMBIENT
ATMOSPHERES
III. The Design of a Sampler System for Trace Quantities of Organic
Vapors
B. Carpenter, E. D. Pellizzari, J. Bunch and E. Sawicki
J. Environmental Science and Technology. Submitted for Publication.
129
-------
APPENDIX
130
-------
M
fc"
1 "
s
UJ •
p
5-
>
I
11 29 38
i
M,i
i
i » i • i • i
1
i in 1 in tl
1
11 i • i • i • i • i • i • i • t • i • i • i • i • i - i
•
•
•
ll,. 1.1
« ' ^o ' «!• ' Vo ' a'o ' 4i ' ilo ' lio ' lio ' ilo ' Uo ' lio ' iio ' do ' lio ' lio ' zlio ' 210 zio ' wo ' 1^0 ' zio
•100
• 10
10
• 40
• 20
• 0
M/e
Figure 46. GLC-MS of bis-(chloromethyl)ether
-------
* •
s.
s
ssr
p
5,
• i • i • i • i •
L
11 28 38 «
i • i
,1 il
• i • i i
i
l|l|l|l|l|l|l|t|>|!I1|llV|»|lll1
a
•
*
1 II
h t'i Vi «'B 4i ib iio 120 lio ilo tio i«o 1^0 lio do 210 2*10 220 2&o 2^0 2^0
19V
10
(0
40
• 20
• 0
M/e
u>
KJ
Figure 47. GLC-MS of bis-(2-chloroethyl)ether.
-------
M
£
CO
5
UJ
p
5
£
IB ' 20 ' 30 ' « ' da ' «!o
•100
•• 10
•• «o
•• 40
•• 20
do ' IQO ' iio' iio' i&o' i4o' lio' ito'
M/e
lilO ' 190 ' 2)10 ' 210 220 ' 2^0 ' MO ' 250
10
U)
Figure A8. GLC-MS of 3-propiolactone.
-------
2i
5
P
54-
• i • i ' i ' i • i • i ' i ' i 'i i • i • I ' i • i • i • i • i • i ' i • i ' i • i • i • i • i • i 1100
10
•• to
•• 20
IB 28 39
+9 59 *9 70
OB
03 110 110 120 190 HO 190 1(0 110 190 ISO 200 210 220 290 210 250
M/e
Figure 49. GLC-MS of vinylene carbonate.
-------
K
M
5
UJ
p
oc
» « i""| . dll_
—ii—j i | i p—i ( P—f
II 20 30 40 50 61
7B ' 8*0 Jo"
100
• 10
• 60
. . 4Q
•• 20
nO 110 120 190
M/e
lid 1^0
200 210 220 230 240 250
CO
Ul
Figure 50. GLC-MS of N-diethylnitrosamine.
-------
JH
5-.
_L
10 20 90 40 50 60 70 80 00
110 120 190 HO 150 160 170 180 ISO 200 210 220 290 210 250
M/e
u>
Figure 51. GLC-MS of nitroir.ethane
-------
M
1 '
5
UJ •
p
(T
_J •
1 1 ' 1 ' 1 ' 1
11 2B 3D +0
i i i i i
1 i ii
do ' e!o
1
a
,1 ,1
Vo ' 90 ' Jo iko ' lio ' tio ' lio ' ito ' lio ' ito ' 1^0 ' lio ' lio ' zlio ' 210 ' 220 ' wo ' zio ' 2^0
M/e
'109
• 10
• (0
• 40
• 20
• o
Figure 52. GLC-MS of ethylmethanesulfonate.
-------
r*
. .
M
a; .
5
UJ •
p
5-
• i * i • i • i
i in 1
18 28 ' ft ' 48
. i . i .,. i . i . i .,.,.,.,.,.,.,.,.,,,.,,,.,.,,,
.
| ,
•
ft ' ft ' Vo ' B'Q ' ft ' ibo ' no ' 120 ' lio ' ito ' lio ' ito ' 1^0 ' lio ' lio ' 250 ' 210 ' 220 ' 2^0 ' 2^0 ' sb
JVfl
10
• (0
• 10
• 20
• o
M/e
10
00
Figure 53. GLC-MS of glycidaldehyde,
-------
I*
. .
£
i.
5
UJ •
p
(E
• 1 • l ' I ' 1
i
I III
11 21 3D 4
|j
> 1 > 1 1 1 1 1 1 1 1 1 1 1 • 1 I 1 > 1 > 1 > 1 • 1 • 1 • 1 • 1 1 1 1 1 1 1 1 1 • 1
L „,!
1 1 i ll 1 Ilil 1 1 1 i
D 50 60 JO 80 4l lb 110 180 lIlO Uo l&O lU l!lO lIlO 190 290 210 220 2&0 2^0 2&0
•100
10
io
• 40
• 20
• o
M/e
VO
Figure 54. GLC-MS of propylene oxide.
-------
*
fc"
i.
z
1-4
UJ •
p
5 ;
.,.,.,.,.,.,.,.,.,
1
11 ' 2B 30 ' 4
,1
0 ' S
1 Ji
g ' 60
' >!' 1
• i < i i i i i • i • i • i • i • i • i • i > i • i > i • i • i
.
•
1
1/0 »'o ds do no i
k ' do ' do ' do ' do ' rio ' do ' do ' 2^0 ' 210 ' 220 ' »o ' 2^0 ' 250
19V
•0
(0
• «0
. 20
• 0
M/e
Figure 55. GLC-MS of styrene oxide.
-------
5
P
uj
ill . 4 I
rrw*
-• 10
• 10
• «0
•• 20
18 21 31
8 60 ' Vo ' 8*8 ' ^8 ' 180 ' lio 120 1^0
M/e
lio lio 1^0 lio lio 200 210 220 zio MO »o
Figure 56. GLC-MS of butadiene diepoxide.
-------
t
t-t
p
tc
UJ
I ' I ' I • I « I ' I
_J,I,I,_J,J_
i I I100
80
•• «0
ID 20 30 1-0 5D
7*8 ' 88 ' h ' ibo' iio' iio ' lio' iio' iio '
M/e
210 2J>0 ' 230 M0250
Figure 57. GLC-MS of acrolein.
-------
LU
P
cn
LU
oe
ill
100
•• BO
•• so
• 40
•• 20
ILL
10 20 3D +D 50
&'o
70*8 flO ibO 110 120 190 HO 150 160 110 150 ISO 200 210 220 290 240 250
M/e
Figure 58. CLC-MS of methyl ethyl ketorve.
-------
*
M
Z .
5
UJ •
P
tr
^
• i • i < i • i • i • i • i
J
i 1
1 i ' i "
1 I
l'o ' 20 ' 38 ' « ' 50 ' & ^0 ' 90 ' Jfl ' 1
i i i i i i i • 1 • i . i • i < i • i . i . i • i • , . , . ,
.
•
100
BO
• to
• 10
• 20
ho ' lio ' 120 ' iio ' ilo ' iio ' ilo ' do ' iio ' iio ' 290 ' 210 ' 220 ' afo ' 2^0 ' 2^0
M/e
Figure 59. Mass spectrum of malelc anhydride.
-------
to
P
5
i
rrioo
.. 10
•• «o
•• 10
•• 20
do' do' do' do' do' do' do' do' do' do' 210' 210' 220' s*io' 210' 2^0
M/e
10 20 3D to 50 60 Vg 9*0
Figure 60. Mass spectrum of succinic anhydride.
-------
M
2.
5
UJ •
P
5-
»
•
•
11 21 ' 3*8
1
i § i i i i • i i • i • i • i • i • i • i
ll Ll ,i ll
•
1 , ii
+'g h u Vi «'o SB ibo 110 i
o ' iio ' ito ' do ' lio ' do ' do ' lio ' 210 ' 210 ' 2^0 ' 2^0 ' 2^0 ' 2$o
M/e
109
• 10
10
• 40
20
• 0
Figure 62. GLC-MS of tetramethylene sulfolane.
-------
» *
g.
s
y '
p
5-
^
,,,,.,,,.,. i . i . i • i
J
111 ill 1
11 28 38 ' *0 ' !?iJ 68
1 ,.,!
• i > i • i < i • i < i • i • i • i • i • i • i i i • i • i > i
100
10
(0
40
• 20
jo ' so ' dg ' ilio ' 110 ' lio ' lio ' lU ' ib ' lU ' i^ro ' lio ' lio ' ako ' 210 ' zio ' zio ' 2^0 ' zio
M/e
co
Figure 63. GLC-MS of aniline.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/2-74-121
2.
3. RECIPIENT'S >CCESSIOf*NO.
4. TITLE AND SUBTITLE
Development of Method for Carcinogenic Vapor Analysis
in Ambient Atmospheres
5. REPORT DATE
July 1974
6. PERFORMING ORGANIZATION CODE
7. AUTHOR
------- |