EP A/600/A-94/239
FINAL REPORT
APPLICABILITY OF CANISTER SAMPLING FOR HAZARDOUS AIR POLLUTANTS
by
Thomas J. Kelly and Michael W. Holdren
BATTELLE
505 King Avenue
Columbus, Ohio 43201
March 1994
Contract Number 68-D0-0007
Work Assignment No. 45, Subtask 4
Project Officer
William McClenny
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
MD-44
Research Triangle Park, North Carolina 27711
-------
ABSTRACT
This paper evaluates the applicability of sampling with evacuated canisters for volatile
organic compounds listed among the 189 Hazardous Air Pollutants (HAPs) in the 1990 U.S.
Clean Air Act Amendments. Nearly 100 HAPs have sufficient vapor pressure to be
considered volatile compounds. Of those volatile organic HAPs, 52 have been tested
previously for stability during storage in canisters. The published HAP stability studies are
reviewed, illustrating that for nearly all of the 52 HAPs tested, canisters are an effective air
sampling approach. However, the published stability studies used a variety of canister types
and test procedures, and generally considered only a few compounds in a very small set of
canisters. A comparison of chemical and physical properties of the HAPs has also been
conducted, to evaluate the applicability of canister sampling for other HAPs, for which
canister stability testing has never been conducted. Of 45 volatile HAPs never tested in
canisters, this comparison identifies 9 for which canister sampling should be effective, and
17 for which canisters are not likely to be effective. For the other 19 HAPs, no clear
decision can be reached on the likely applicability of canister sampling.
ii
-------
INTRODUCTION
The collection of whole air samples in canisters for determination of volatile
organic compounds (VOCs) in the atmosphere is common practice. Passivated stainless steel
canisters are used in a variety of air monitoring programs, and form the basis for the U.S.
Environmental Protection Agency's Compendium Method TO-14 (Winberry et al., 1988;
McClenny et al., 1991a). The stability of volatile alkanes, alkenes, aromatic hydrocarbons,
and their halogenated derivatives in canisters is well established (McClenny et al. 1991a, and
references therein).
Recently attention has been focused toward volatile organic compounds containing
oxygen, nitrogen, and sulfur. Such compounds are often collectively called polar VOCs
(PVOCs). In part, this attention results from the recognition that atmospheric reactions of
VOCs often produce PVOCs as products, and that PVOCs in turn may be quite reactive.
Thus understanding of atmospheric chemical processes requires measurement of PVOCs. In
addition, the potential toxicity of PVOCs in air has led to regulatory efforts directed toward
them. For example, Title III of the 1990 Clean Air Act Amendments (CAAA) defines as
hazardous air pollutants (HAPs) a list of 189 diverse chemicals and chemical groups, and
mandates control and reduction of the human health risks from these compounds in ambient
air (Clean Air Act Amendments, 1990). Nearly 100 of the HAPs are sufficiently volatile
(i.e., vapor pressure >0.1 mm Hg at room temperature) to be considered VOCs (Kelly, et
al., 1994). Some of those HAPs are common aromatic and halogenated VOCs, but over half
of the volatile HAPS can be classified as PVOCs.
To meet the requirements of the CAAA, measurements of the diverse HAPs are
needed. This requirement has prompted surveys of the existing and potential measurement
methods for the HAPs (McClenny et al., 1991b; Keith and Walker, 1992; Kelly et al.,
1994). The application of canister sampling to volatile HAPs would seem a natural extension
of the existing methods for VOCs, and indeed the methods surveys noted above consider that
approach. However, reactivity and water solubility, more than polarity, are the defining
characteristics of most PVOCs. As a result, the stability of the PVOCs during storage of
sampled air in canisters is not assured. Several experimental studies have tested the stability
1
-------
of PVOCs and VOCs in canisters (e.g., Westberg et al., 1981; Gholson et al., 1990; Merrill
and Zapkin, 1991; Pate et al., 1992; Oliver, 1993; Kelly et al. 1993), and stability in
canisters has also been evaluated from a theoretical standpoint (Coutant and McClenny,
1991; Coutant, 1993). These studies particularly indicate the importance of humidity in
reducing the loss of volatile compounds to the canister walls.
The purpose of the present paper is to summarize published information on the
stability of volatile HAPs, both VOCs and PVOCs, during storage in canisters. A
comparison of chemical and physical property data then identifies additional HAPs for which
canisters should be an effective atmospheric sampling approach. That same comparison also
indicates volatile HAPS for which canisters are not likely to be useful.
REVIEW OF STABILITY STUDIES
In this section several experimental studies are summarized, with the intent of
identifying HAPs for which canister stability has been demonstrated, or alternatively shown
to be inadequate. This review concentrates on the stability of VOCs and PVOCs designated
as HAPs by the 1990 Clean Air Act Amendments. In summarizing the experimental studies,
guidance has been drawn from theoretical considerations (Coutant and McClenny, 1991;
Coutant, 1993) and from anecdotal evidence as to the parameters that may affect the stability
of VOCs and PVOCs stored in canisters. As a result, indication is provided (when available
from the original references) on the material, age, or history of the canisters used; the degree
of humidification of the stored samples; the number of compounds in the stored test
mixtures; the duration of the test; canister pressure during storage; and the concentrations of
the compounds tested. Canister stability testing has also been conducted by researchers other
than those cited here. However, the results of those tests have not been compiled or were
unavailable for this review (R. Rasmussen, Oregon Graduate Institute; L. Ogle, Radian
Corp.; personal communications, 1993).
Westberg et al., (1981): In this study 6 L Summa polished stainless steel canisters
were used to assess the stability of a variety of VOCs and PVOCs. In one test four canisters
2
-------
were spiked with low concentrations of four terpene compounds during pressurization with
ambient air. The canisters were analyzed for a total of 28 VOCs (including 8 HAPs), 24 of
which were native in the ambient air and 4 of which were the spiked terpenes. Analysis was
conducted at weekly intervals for comparison to an initial analysis at the time of filling the
canisters. The initial concentrations were 1 to 30 micrograms per cubic meter (/xg/nf).
Most of the tested compounds were stable within 10 percent over a two-week storage period,
with the terpenes exhibiting poorer stability. In a second test, 11 PVOCs (including one
HAP) were spiked into ambient air at concentrations of 20 to 40 pg/m3. Most of the PVOCs
were stable within about 10 percent over one week of storage, and within 30 percent over 3
weeks. Considerable variability in results was noted from one canister to another. Low
molecular weight (i.e., Cj to C6) alcohols were reported to be unstable in canisters, even
within a one day storage period.
Holdren et al. (1984): This study evaluated the stability of 16 VOCs, including 14
HAPs but no PVOCs, in seven new Summa polished stainless steel canisters of 6 liter
volume. The stored mixture contained all 16 compounds, at concentrations of 0.6 to 2.7
ppbv, in humidified air. Storage was conducted for 7 days, with determination of the
canister contents on the day the canisters were filled (day 0), and on days 2, 4 and 7. The
bias of measured concentrations on day 7 relative to those on day 0 was less than 5 percent
for 13 of the 16 VOCs, and the maximum bias was -18.8 percent, for benzyl chloride.
Oliver et al. (1986): This study employed a combination of new and previously
used 6 L and 3 L Summa polished canisters in evaluating the stability of 18 VOCs, including
16 HAPs but no PVOCs. Concentrations of all the VOCs were less than 2 ppbv in all tests,
and ambient air was the matrix for filling the canisters, up to 30 psig. Fifteen of the VOCs
were spiked into ambient air in filling the canisters; three others (benzene, toluene, and o-
xylene) were used at the levels native in the ambient air. Four new 6 L canisters, four new
3 L canisters, and five used 6 L canisters were used in a 7-day storage test, with
measurements made on days 0, 2, 4, and 7. A 30 day storage test was also conducted with
three other used 6 L canisters; measurements were made on days 0, 1, 15, and 30. All
3
-------
compounds were found to be reasonably stable in the canisters, with most exhibiting mean
rates of change of less than 1 percent per day of storage in both the 7 day and 30 day tests.
Jayanty (1989): In this study a single 6 L Summa polished stainless steel canister
was used to test the stability of 18 VOCs, including 17 HAPs but no PVOCS. The 18 VOCs
were prepared together at concentrations of 4 to 12 ppbv in air humidified to an unspecified
level. The canister was filled to an initial pressure of 30 psig. Stability was tested during an
11 day storage, with the VOC content of the canister determined on days 0, 1,7, and 11.
For all of the 18 VOCs, concentrations measured on day 11 agreed well (i.e., typically
within 5 percent) with the day 0 concentrations and with the expected concentrations put into
the canister.
Gholson et al. (1989): In this study the stability of VOCs in canisters was
evaluated under conditions simulating the sampling of dispersing emissions from hazardous
waste incinerators. New 6 L stainless steel Summa polished canisters were used in several
tests with up to 18 VOCs, including 17 HAPs but no PVOCs. In an initial test, 5 VOCs (all
HAPs) were stored at concentrations of 10 to 15 ppbv in three canisters pressurized to 15
psig with dry air. Analyses were conducted on all three canisters after 4 days, and on one
canister after 11 days of storage. No loss of any of the 5 VOCs was found. The same VOC
mixture in air was also tested in one canister containing 2,800 ppm of water and 120 ppm of
HC1, over a 14 day period with measurements on days 0, 1,5, 8, and 14. All 5 VOCs were
found to be stable within about 8 percent over the 14 day period. In a separate test, a
mixture of 18 VOCs (17 HAPs) were tested in a single canister at concentrations of 5 to 12
ppbv in nitrogen, in the presence of 5,600 ppm of water and 120 ppm of HC1. The water
content was stated to exceed the saturation vapor pressure in the canister. Storage over 14
days, with testing on days 0, 1,3, 7, and 14, showed stability generally within 10 percent
for the tested compounds. However, large positive artifacts were observed for a few
compounds, possibly due to gas chromatograph column degradation caused by the water/HCl
in the canister. The test most pertinent to ambient sampling was an 11-day study of the same
18-component mixture, in nitrogen with 150 ppm H20 but with no HC1, which showed
4
-------
stability generally within IS percent for all compounds, with no positive artifacts.
Gholson et al., 1990: This study was conducted using four new 6 L aluminum
canisters and one new 6 L stainless steel canister. The investigation of aluminum canisters
was motivated by previous work (Allen et al., 1987) that indicated improved
chromatographic peak shapes for PVOCs when an aluminum sample loop was used, relative
to those obtained with a stainless steel sample loop. Two of the aluminum canisters and the
one stainless steel canister were Summa polished, the others were not. Fn one test, a mixture
of 18 VOCs (17 HAPs, no PVOCs) each at about 4 ppbv in air was stored in the four
aluminum canisters. Two of the canisters (one Summa polished, one not) were humidified
with 2,000 ppm of water, whereas the other two contained dry samples. Storage was
conducted over 10 days, with analysis of canister contents on days 0, 1,4, 7, and 10. In a
second test, seven compounds (4 HAPs, 2 PVOCs) were stored at 10 to 20 ppbv levels in
one polished aluminum canister, one unpolished aluminum canister, and the one stainless
steel canister, with 170 ppm of water added to all three canisters. This mixture was stored
for 14 days, with analysis on days 0, 1,3, 7, and 14. Storage stabilities for the VOCs were
poor in dry canisters, but good in humidified canisters. Summa polishing of the aluminum
canisters was found to have only a small effect on VOC stability. Aluminum canisters were
found deficient relative to stainless steel for storage of PVOCs, and showed no advantage for
storage of VOCs.
Holdren et al. (1991): This study was one of the most comprehensive, addressing
the stability of a variety of compounds in nine new Summa-polished canisters obtained from
three different vendors. Three separate tests were conducted, using purified air of 70 percent
relative humidity, over storage periods of up to 33 days. The first test involved 42 VOCs
(the 41 compounds on the TO-14 target list, plus 1,3-butadiene), including 31 HAPs. A
mixture containing nominally 10 ppbv of each compound was analyzed on days 0, 20, and
32, and a separate 2 ppbv mixture was analyzed on days 0, IS, and 33. The second test
involved 18 VOCs, including 15 PVOCs and 12 HAPs. A 10 ppbv mixture was analyzed on
days 0, 10, and 33, and a 2 ppbv mixture on days 0, 11, and 30. The third test was of 13
5
-------
toxic compounds, including 8 PVOCs and 9 HAPs. A 10 ppbv mixture was analyzed on
days 0, 10, and 30, and a 2 ppbv mixture on days 0, 20, and 32. Ignoring duplication in the
three tests, a total of 67 compounds, including 47 HAPs, were tested in this study. Most
compounds tested showed good stability over 30 days storage, though some polar compounds
were more stable at the nominal 10 ppbv level than at 2 ppbv.
Parmar (1991): This study tested the stability of four volatile sulfur compounds,
including the HAPs carbonyl sulfide (COS) and carbon disulfide (CSj), in one Summa
polished stainless steel canister. Initial concentrations of 1,200 ppbv COS and 1,400 ppbv
CS2 were stored in dry nitrogen for 3 days. The loss of COS in the canister was reported to
be 8.3 percent in one day, 12.3 percent in 2 days, and 17.5 percent in 3 days. The
corresponding losses for CS2 were reported to be 1.1 percent, 1.5 percent, and 4.2 percent.
Merrill and Zapkiii (1991): This study used three 6 L Summa polished stainless
steel canisters to test the stability of four PVOCs (including the HAP methanol) over a 35
day period in air of 60 percent relative humidity. The initial concentrations were 20 to 30
ppbv, except for methanol which was at 60 ppbv. Analysis of the canister samples on days
0, 7, 15, 20, and 35 showed no significant losses of any of the four PVOCs.
Pate et al. (1992): In this study five 6 L stainless steel canisters, three Summa
polished and two not polished, were used to test the stability of 12 VOCs, including 8 HAPs
and 10 PVOCs. The compounds were present at levels of 5 to 13 ppbv in both dry and
humidified mixtures. The humidified mixtures exceeded saturation for water vapor.
Canister pressure was indicated as 3 atmospheres. Storage was conducted over 31 days, with
analyses on days 0, 1,3, 7, 14, and 31. Stability in humidified canisters was very good,
whereas that in dry canisters was relatively poor. Summa polishing had little apparent effect
on the stability of the tested compounds.
Oliver (1993): This study tested the stability of 9 PVOCs, including 5 HAPS,
using four new 6 L Summa polished canisters. The PVOCs were spiked into ambient air to
6
-------
fill the canisters. Testing was conducted over 7 days, with analysis on days 0, 2, 4, and 7,
in two separate tests. The first test used PVOC levels of 1 to 4 ppbv, the second 3.5 to 8
ppbv. All the tested PVOCs were stable within 25 percent over 7 days of storage, with most
exhibiting stability within 15 percent.
Kelly et al(1993): As part of a method development and ambient sampling
effort for PVOCs, canister stability was tested for 15 VOCs, including 9 HAPs and 13
PVOCs, with used 6 L Summa polished stainless steel canisters. A mixture of these 15
compounds at about 4 ppbv each, in air of 80 percent relative humidity and 5 psig pressure,
was analyzed after 4 days and after 12 days of storage. Linear regression of the day 4
results relative to the day 0 results for all 15 compounds indicated 13 percent loss over the 4
days of storage. Individual compounds showed losses of 2 to 38 percent over that time
period. The PVOC compounds showed relatively poor stability after 12 days of storage.
Holdren et al. (1994): This study was intended primarily to evaluate differences
in storage stability among nine new Summa polished stainless steel canisters, treated in three
different ways. Three of the canisters underwent no further treatment, three underwent the
Silcosteel® surface deactivation process, and three were treated with the Silcosteel treatment
plus an added deactivating reagent. A target group of 9 VOCs, including 6 PVOCs and 7
HAPs, was prepared at nominal levels of 20 ppbv each, and of 2 ppbv each, in humidified
high purity air. Actual concentrations of individual compounds in the 20 ppbv mix were
10.7 to 30.4 ppbv, and in the 2 ppbv mix were 1.1 to 3.0 ppbv. The 20 ppbv mixture was
evaluated with both 52 percent and 21 percent relative humidity air (RH measured at 1 atm
pressure); the 2 ppbv mixture was evaluated with 21 percent relative humidity only.
Analyses were conducted on days 0, 4, and 7 of storage. No significant differences were
observed among the three canister treatment processes. Stability of the target compounds
was within the 25-30 percent uncertainty of the measurement method.
7
-------
HAPs STABILITY RESULTS
The reported degree of stability of a VOC in a canister is the end product of
several factors, not all of which have been properly documented in the published studies. To
begin with, the preparation and introduction of low levels of VOCs into canisters may
introduce variability and uncertainty in stability testing, particularly for PVOCs. For
example, the precision of replicate standards has been shown to be poorer for PVOCs than
for VOCs (Kelly et al., 1993), and initial concentrations reported in canisters do not always
agree with those expected (Pate et al., 1992). Furthermore, the stability of a compound in
air within a canister is affected by a complex interaction of factors, including the vapor
pressure, polarizability, water solubility, and aqueous reactivity of the compound; the
characteristics of the canister surface; the past use of the canister itself; the reactivity of the
compound with other species present in the air sample; the humidity of the air; the canister
pressure; the temperature of the canister during sampling and storage; and the competitive
adsorption of the compound relative to that of water vapor and trace chemicals in the sample
(Coutant and McClenny, 1991; Coutant, 1993). Finally, the reported stability in a canister is
based upon an analytical measurement, and thus is subject to uncertainties from that source.
Methanol, for example, has been found to be stable in canisters in some studies (Merrill and
Zapkin, 1991; Pate et al., 1992), but in others analytical difficulties have prevented its
determination (Westberg et al., 1981; Oliver, 1993). In actuality, the reported studies have
determined the overall recoverability of VOCs in the respective test and analysis systems, not
merely the stability of VOCs within a canister. From an analytical standpoint, overall
recoverability is the more important factor. However, the impact of analytical capabilities on
reported stability results has rarely been discussed (Oliver et al., 1986; Holdren et al.,
1994).
The summaries presented above show that canister stability testing has generally
been limited in nature. Most of the reported stability tests have addressed relatively few
compounds, and most have used very few canisters. The comparability of results among the
various studies is uncertain, because of differences in the performance of the tests. For
example, storage tests have been conducted for as little as 7 days, and for as long as 33
8
-------
days. Most studies have used purified air or nitrogen as the diluent gas, but in a few studies
ambient air was used. Thus the potential range of other chemicals present in the canister
with the target VOCs has varied widely. The degree of humidification of the test mixtures
and the pressurization of the canisters differ among the studies, and in some studies these
parameters are not well documented. Stability studies have usually been conducted with
VOC concentrations ranging from about 1 ppbv to 20 ppbv, but one study (Parmar, 1991)
used levels of over 1 ppmv. The size, age, surface treatments, and even materials of
construction of the canisters used have also varied. As a result, the published results do not
constitute a single cohesive data set. '
Despite the ambiguities in reported tests of VOC canister stability, the results of
those tests are of interest as indications of the applicability of canister sampling for HAPs.
Table 1 summarizes the canister stability results for the 52 HAPs for which such testing has
been reported. Table 1 lists the HAPs in the same order as in the CAAA list, and indicates
the stability reported, the concentration level tested, the pertinent literature, and any
comments on the results. Because of the diversity of the stability tests, as noted above,
Table 1 shows results for each study cited, rather than a summary stability value for each
HAP.
The diversity of the reported stability tests requires explanation of the entries in
Table 1. Unless otherwise noted, the entries in Table 1 refer to tests using Summa polished
stainless steel canisters, with humidified, purified air as the diluent gas. When a published
study included more than one stability test, the entries refer to the longest period of storage
within each study. Results shown in Table 1 were also selected to be those most pertinent to
ambient sampling of HAPs. For example, results in Table 1 from reference 5 (Gholson et
al., 1989) are from a test conducted without high levels of HC1 in the canister. A few
studies reported stability results using parallel flame ionization and electron capture gas
chromatographic detectors, and for some compounds reported contradictory trends in canister
stability from the two detectors (Oliver et al., 1986; Jayanty, 1989, Gholson et al., 1989).
In such cases, the worse-case result, i.e., that indicating poorer stability, is shown in
Table 1. The means of calculating the stability of VOCs in canisters also differed among the
-------
TABLE 1. SUMMARY OF CANISTER STABILITY TESTS ON CAAA HAZARDOUS AIR POLLUTANTS
jcompound
Stability*
(ft Change/Days)
Teited
Cone.
(PPbv)
Referenced Comment*®
Compound
Tested
Subility* Cone,
(ft Change/Days) (ppbv)
References* Comments' [
Acctilddiyde
30/4
30.4
13
Caibon
1/30
17,9
7
30/7
3.0
disulfide
21/32
3.7
4/3
1400
8
[<51
lAcetonitrile
4/33
20.5
7
15/7
28.2
13
14/30
4.3
15/7
2.8
25/7
8.6
11
[1.21
.38/4
4
12
Carbon
1/7
2.0
2
tetrachloride
15/30
0.9
3
(2]
Acrylooilrile
5/33
16.3
7
5/11
4.5
4
6/30
3.4
8/11
4.7
5
W
5/31
13.9
10
48/10
4.0
6
n
12n
4.3
11
[2.31
3/10
4.0
[81
15 n
7.3
10/32
12.8
7
14/4
4
12
13/33
2.7
20/7
25.8
13
|
20/7
2.6
Caitoonyl
18/3
1200
8
[6]
1
sulfide
•
lAllyl chloride
2/32
13.1
7
|
15/33
2.7
Chloro-
3/7
1.4
2
I
benzene
7/30
1.0
3
[21 |
Benzene
19/21
2.1
1
12]
10/11
10.0
4
5/30
0.6
3
[2]
14/11
10.0
5
[4] 1
2/11
4.7
4
14/10
4.2
6
[9] I
9/11
4.7
5
[4]
24/14
10.2
[10] 1
1/10
4.3
6
[51
5/14
10.2
[1U |
6/32
12.0
7
9/32
U.O
7
16/33
2.5
14/33
2.3
5/33
12.1
1
7/30
2.5
Chloroform
in
1.9
2
1
14/4
4
12
34/30
1.2
3
PI |
f
10/7
19
13
3/11
4.8
4
\
20/7
1.9
1/11
4.8
5
[4] I
1
5/10
4.4
6
[51 1
gBenzyl chloride
19/7
1.0
2
10/32
14.1
7
11/30
0.75
3
[21
7/33
2.9
9/32
10.9
7
36/33
2.3
Cumene
10/32
7.7
7
13/33
1.6
Bromoform
5/32
12.3
7
13/33
2.6
1,3-Butediene
4/11
4.7
4
2/11
4.7
5
[4]
9/10
9.3
6
[51
14/32
11.8
7
34/33
2.5
19/33
11.0
10
-------
TABLE i. (Continued)
Compound
Stability"
(% Change/Days)
Tcited
Cone.
(PPbv)
Referenced Comment**
Compound
Toted
Stability" Cone.
(% Change/Days) (ppbv)
Reference** Comment!*
1.4-
11/32
9.8
7
Ethylene
in
1.9
2
Dichlorobenzeae
14/33
2.0
dichloride
12/30
0.9
3
[2]
2/11
5.1
4
14-
7J/7
0.82 (ei>)
2
10/11
5.1
5
[4]
Dichloropropene
1.6/7
0.64 (Irani)
2/10
4.0
6
[5]
4/30
0.49 (eii)
3
(21
9/32
; 13.8
7
15/7
0.39 (tnni)
9/33
2.9
12/32
12.9 (eii)
7
13/33
2.7 (eis)
Ethylene
26/33
11
7
8/32
13.9 (trans)
oxide
15/33
2.9 (tnns)
Ethylidene
9/32
12.8
7
1,4-Dioxane
100/4
19.2
6
[10]
dichloride
11/33
2.7
77/14
19.2
(111
25/30
12.6
7
Hexachloro-
11/7
1.3
2
72132
2.6
butadiene
35/30
1.0
3
[2]
16/32
6.2
7
Elhyl acrylate
3/33
9.9
7
10/33
1.3
19/30
2.1
3/31
4.9
10
Hexane
16/21
1.6
1
[2]
in
1.1
11
[2]
1
3/7
3.6
Methanol
<5'/35
60
9
1
15/4
4
12
46/31
131
10
¦Ethyl benzene
16/21
0.6
1
[2]
Methyl
4/11
8.5
4
I
12/11
9.4
4
bromide
15/11
8.5
5
[4]
1
17/11
9.4
5
[41
4/10
4.1
6
[5]
R
50/10
3.9
6
[51
6/32
12.7
7
10/32
9.5
7
34/33
2.6
17/33
2.0
Methyl
19/32
10.1
7
Elhyl chloride
12/32
11.4
7
chloride
17/33
2.1
29/33
2.4
Methyl
1/7
1.9
2
Ethylene
2/7
1.4
2
chloroform
11/30
1.4
3
[2]
dibromide
3/30
1.2
3
[2]
4/11
10.1
4
11/11
9.7
4
1/11
10.1
5
[4]
6/11
9.7
5
[4.12]
1/10
4.1
6
[5]
20/10
4.3
6
[51
17/32
6.1
7
14/32
13.5
7
6/33
13
12/33
2.8
3/31
53
10
11
-------
TABLE 1. (Continued)
Compound
Stability"
(% Change/Days)
Tested
Cone,
(ppbv)
Referenced Comment**
Compound
Tested
Stability* Cone.
(% Change/Days) (ppbv)
References* Comments* |
Methyl ethyl
14/21
13
1
121
1,1,2,2-
9/32
10.1
7
¦ketone
4/33
12.0
7
Tetrachlofo-
15/33
2.1
94/30
2.5
ethtnc
7/31
26
10
58/7
1.2
11
[21
Tettachloro-
3.2n
1.6
2
1/7
4.1
ethyksne
10/30
1.0
3
PI
14/14
4
12
10/11
5.2
*
30/7
19.0
13
3/11
5.2
5
14] j
20/7
1.9
6/10
4.6
6
[51 1
13/32
10.8
7
1
Methyl iso-
21/30
8.6
7
13/33
2.2
|
butyl ketone
24/32
1.8
>
j
Methyl
7/30
10.1
7
Toluene
20/21
3.9
1
[21 1
gmethaciylate
15/32
2.1
20/30
1.0
3
[21
1
1/7
1.2
11
121
4/11
9.5
4
1
1
4/7
3.9
6/11
9.5
5
HI |
5/10
4.3
6
[91
Methyl
9/33
9.1
7
33/14
10.1
[101
E-butyl ether
12/30
1.9
16/14
10.1
[111
3/31
7.1
10
8/32
10.4
7
18/4
4
12
11/33
2.2
14/7
14.2
13
5/33
10.1
30/7
1.4
8/30
2.1
6/31
5.0
10
Methylene
3/11
5.2
4
15/4
4
12
.
chloride
4/11
5.2
5
HI
25/7
16.0
13
55/10
4.3
6
13]
30/7
1.6
26/32
17.1
7
27/33
3.6
1,2,4-Tri-
5/32
7.1
7
chloro-
7/33
1.5
Nitrobenzene
26/30
10.4
7
benzene
Propylene
3/11
10.5
4
l.U-Tri-
12/32
12.1
7
dichloride
6/11
10.5
5
M
chloroethme
18/33
2.5
<1/10
4.1
6
[S]
12/32
11.4
7
Trichloro-
2/7
1.8
2
17/33
2.4
ethylene
11/30
1.0
3
PI
5/11
5.4
4
Propylene oxide
17/33
15.4
7
9/11
5.4
5
M
9/4
4
12
[131
1/10
4.3
6
[51
14/32
11.6
7
Styrene
8/32
9.6
7
18/33
2.4
34/33
2.0
12
-------
TABLE 1. (Continued)
Stability*
Compound (% Change/Days)
Tested
Cone.
(Ppbv)
Referenced Comments*
Compound
H
Tested 1
Stability* Cone. 1
(% Change/Days) (ppbv) References' Comments'|
2,2,4-Trunelhyl-
2201
0.6
1
12)
m-Xykne
17/21
2.?
1
(2,14)
pentane
7/32
9.3
7
17/33
1.9
Vinyl acetate
39/33
11.6
7
55/11
2.4
o-Xykne
12/21
0.8
1
PI
am
7J
10
120/30
0.14
3
(2]
31/4
4
12
15/7
0.15
16/11
11.7
4
Vinyl chloride
3 n
1.1
2
10/11
11.7
5
M
11/7
0.8
3
12]
59/10
4.1
6
(5]
<1/11
5.9
4
9/32
9.8
7
S/ll
5.9
5
14]
10/33
2.0
83/10
3.9
6
(51
10/32
28.3
7
p-Xylene
17/21
<2.3
1
P.15]
25/33
5.9
Vinylidene
4/7
2.1
2
*
chloride
4/30
1.1
3
PI
6/14
9.4
6
110]
7/14
9.4
(11)
7/32
12.9
7
iifa'"» > >>¦„«
11/33
ip-nwwrwMi
2.7
a Stability indicate*! as absolute value of pereeKRnge from wSf^n^^^^nn^ueFluntion ofitQrap"uMlay«T
b References: 1, Westberg et al., 1981; 2, Holdren el al., 1984; 3, Oliver et al., 1986; 4, Jayanty, 1989; 5, Gholson, ct al.r 1989; 6,
Gholson et al., 1990; 7, Holdren et al., 1991; 8, Ptrmar, 1991; 9, Menill and Zapkin, 1991; 10, Pale et a]., 1992; 11, Oliver,
1993; 12, Kelly et a!., 1993; 13, Holdren et at., 1994.
c Comments:
(11 Coeluted with cthanol.
[2] Ambient air used as diluent.
{3) Coeluted with 2-propanol.
(4] Humidified nitrogen used as diluent
(5] Aluminum canisters.
(6] Test conducted in dry N2.
(7) A1 pusivatcd cans.
(8) Al unpassivat/ed cans.
(9) Al canisters, 2000 ppm H20.
{10] Al canister, 170 ppm Hp.
(11) Stainless steel canister, 170 ppm HjO.
(121 Corrected for instability of control canister.
(13) Coeluted with acetone.
(14J Coeluted with p-*ylene.
(151 Coeluted with m-xylene.
d Estimated; based on lack of a significant concentration trend with time.
13
-------
studies. Most studies reported canister results for each individual day on which
measurements were made. The final day's results are shown in Table 1. However, Oliver
el al. (1986) instead performed a linear regression with time over all canister analysis days.
Canister stability was then reported in terms of a mean percentage change per day of storage
(Oliver et al., 1986). Results from Table 1 were calculated as the product of the mean
percent change per day and the total number of days of storage. Several studies conducted
tests at more than one initial concentration level; separate entries are shown in such cases, in
Table 1. Separate results are also shown for the cis- and trans- isomers of 1,3-
dichloropropene, which have been tested separately (the CAAA HAPs list makes no
distinction between these isomers).
From an analytical standpoint, the important factor in canister stability is the
absolute change in concentration of a volatile compound, relative to its initial concentration
in the canister. The direction of any change in concentration, i.e., positive or negative, is
less important. As a result, the stability results in Table 1 are shown as absolute values of
the percentage change during storage. However, the results tabulated are based on both
positive and negative changes reported in the cited studies. A negative change in
concentration during storage, i.e., a loss of compound within the canister, is the more
commonly expected result. Such loss can result from adsorption on the canister surface, or
from chemical reactions with water or other chemicals in the canister (Coutant, 1993). A
positive change, i.e., an increase in measured concentration, can also occur, as a result of
the partitioning of volatile compounds between the gas phase and condensed water on the
canister surface (Coutant, 1993). The extent of this effect depends on the Henry's law
equilibrium of the volatile compound between the aqueous and gas phases. In this scenario,
a volatile compound may be partitioned largely into condensed water on the canister surface,
under the initial canister conditions of pressure, temperature, and relative humidity.
However, as sample (and water vapor) is removed during successive analyses of the canister,
changes in pressure result which may reduce the amount of condensed water, and
consequently shift the partitioning of the volatile compound toward the gas phase. As a
result, later analyses appear enriched in the volatile compound, and a positive trend in
concentration with time is observed. In addition, normal variability in preparation of test
14
-------
mixtures and in canister analysis may cause either positive or negative variations in canister
stability results. All of these factors may play a role in any given stability study, but in
general cannot be resolved based on the published information.
The quantitative stability results in Table 1 may be normalized in a sense by
comparison to a quantitative criterion of the stability required to conduct ambient sampling.
Although rapid transfer of canisters from the field to the lab has been employed to minimize
storage times (e.g., Kelly et al.t 1993), a longer storage schedule is more typical. For the
present discussion, we assume that stability within 25 percent over a two-week storage period
is sufficient for the great majority of ambient measurements. For many HAPs, the reported
stability results show considerable variation. However, relative to this criterion, Table 1
indicates that nearly all of the HAPs tested to date are sufficiently stable that canister
sampling is an effective approach to ambient measurements. Possible exceptions are
carbonyl sulfide and acetaldehyde, for which storage times of only a few days cause large
changes in concentration. However, very limited testing has been done for these compounds,
and further tests would be valuable, especially for carbonyl sulfide, which has only been
tested during storage in dry nitrogen. In addition, Table 1 shows variable stability results for
methyl ethyl ketone, propylene oxide, and vinyl acetate; further testing is also needed for
these HAPs. Some of the variability in reported canister stability for these compounds
undoubtedly arises from difficulties in the analysis step.
DISCUSSION
The canister stability reported for the HAPs (Table 1) is largely the product of the
chemical and physical properties of the HAPs. Consequently, the usefulness of canister
sampling for other HAPs, not yet tested for stability, should correlate with their respective
properties. In this section, comparisons of HAPs properties are used to identify HAPs for
which canisters are, and are not, likely to be useful for sample storage.
Chemical and physical properties of volatile HAPs were compiled in a recent
survey of proven and potential ambient measurement methods for the 189 HAPs (Kelly et
al., 1994), and in a modelling study of VOC canister stability (Coutant, 1993). The
-------
properties compiled were vapor pressure, boiling point, polarizability, water solubility,
Henry's law equilibrium constant, aqueous reactivity, and reactivity in the atmosphere. The
latter two properties were quantified in terms of the typical half-life for reaction in each
matrix (Spicer et ai, 1993; Kelly et al., 1994). Data on chemical and physical properties of
HAPs were compiled from a variety of reference materials. Key references for vapor
pressure were Jones and Bursey (1992) and Weber et al. (1991); for polarizability Sansone,
et al. (1979), CRC (1979), and Keith and Walker (1993); for water solubility Mackay et al.
(1993) and Verschueren (1983); for Henry's law equilibrium Hine and Mookerjee (1975),
Mackay and Shiu (1981), Yaws et al. (1991), Eklund et al. (1991), and Betterton (1992);
and for reactivity data Howard et al. (1991) and Spicer et al. (1993). The overall ranges of
properties found for the volatile HAPs were: vapor pressure, 0.15 to 3,800 mm Hg;
polarizability, 8 to 50 cm3/mole; water solubility, <0.1 to > 100 g/L at 25°C; and
dimensionless Henry's law partition coefficient (Cmta/C^), 2 x 10"2 to 2.5 x 10s. Reactive
lifetimes in water and in the atmosphere ranged from a few minutes to many months.
Reactivity with water was considered the more important factor, since atmospheric reactivity
depends upon factors (e.g., radical reactions, photolysis) which are of no consequence in a
canister. Henry's law partitioning was ^Iso given much greater weight than simple bulk
solubility, because of its greater relevance to the behavior of volatile compounds in a
canister.
In the present study, the properties of HAPs never tested for stability in canisters
were compared to those of HAPs shown to be reasonably stable in canisters (Table 1), to
infer the likely stability of the untested HAPs. An example of such a comparison is shown
in Table 2. Table 2 compares the properties of 1,2-dibromo-3-chloropropane, which has not
been tested for canister stability, to those of 1,4-dichlorobenzene (see Table 1). As Table 2
shows, the vapor pressure of the dibromochloropropane is low, but essentially the same as
that of the dichlorobenzene. The two compounds also have the same polarizability, very
similar water solubility, and very similar Henry's law partitioning. Both chemicals are non-
polar compounds, and are relatively unreactive both in air and water. As a result, it is
reasonable to conclude that the stability of 1,2-dibromo-3-chloropropane in a canister should
16
-------
TABLE 2. EXAMPLE COMPARISON OF HAPs PROPERTIES
HAP
Vapor Pressure Polarizability* Water Solubility Henry's Law Partitioning*
(mm Hg at 25°C) (cm3/mole) (g/L at •Q (cjcj
Reactive Half-Life ]
(Days)
Aqueous Air
1,2-Dibromo-3-chloro-
propane*
1 ,4-Dichlorobenzeneb
0.8
0.6
36.3
36.3
<0.1 at 18
< 1 at 23
20
15
14-180° 6-61
28-180* 8-84
J
a: Not tested for canister stability,
b: Tested for canister stability (see Table 1).
c: Polarizability « (MW/p) (n2-l)/(na+2), where MW « molecular weight, p * density, and n = refractive index.
d: Entries are dimensionless Henry's law constants, concentration in water/concentration in air (CJCJ. Value for l,2-dibromo-3-
chloropropane is based on data for 1,2-dibromopropane (Hine and Mookuijee, 1975; Yaws etal., 1991).
e: Hydrolysis rate is negligible; indicated lifetime based on aqueous aerobic biodegradation.
-------
M
be comparable to that of 1,4-dichlorobenzene, i.e., canister sampling should be effective for
1,2-dibromo-3-chloropropane.
Similar comparisons have been made for other HAPs, both by close similarity of
properties, as in Table 2, or by ranking properties of chemically analogous HAPs. Examples
of the latter approach include comparison of vinyl bromide with vinyl chloride, or of
hexachloroethane with carbon tetrachloride, hexachlorobutadiene, and other highly
chlorinated HAPs. A clear conclusion is not possible for all HAPs, due to the ambiguity or
absence of some chemical and physical property data. In addition, some HAPs are relatively
unusual compounds for which it is difficult to fmd suitable comparisons among the HAPs
previously tested in canisters. However, for many of the untested HAPs some assignment
can be made as to the likely effectiveness of canister sampling. The results of this
comparison of properties are shown in Table 3. This table lists the volatile HAPs which
have not been tested for stability, in the same order as in the CAAA list. Shown in Table 3
are an indication of the likely canister stability for each HAP, and comments or supporting
information. The likelihood of canister stability is indicated as "Y", i.e. stability is likely;
"N", i.e., stability is not likely; or "?", stability is unknown or doubtful based on the
available information. In most cases, the supporting comments call attention to specific HAP
properties that strongly determine the conclusion reached concerning canister stability. In
those cases where canister stability is likely, the comments in Table 3 indicate the HAPs used
as the basis for a comparison of properties.
Table 3 lists 45 volatile HAPs not yet tested for canister stability. Of those HAPs, 9
are indicated as likely to be stable in canister sampling, and 17 as unsuitable for canister
sampling. For 19 of the HAPs in Table 3, no clear decision can be made as to their likely
stability in canisters. The 9 HAPs for which canister stability is likely are generally analogs
of other HAPs previously shown to be stable in canisters (Table 1), or are known to be
unreactive. The other 36 compounds are polar, water-soluble, and potentially reactive
compounds. For these HAPs, little or no canister testing has been done on analogous
compounds. Canister sampling for the 17 HAPs designated by "N" in Table 3 is strongly
discouraged. Reliance on canister sampling for the 19 HAPs designated by a "?" is not
recommended, even for short storage periods, without prior canister stability testing.
18
-------
TABLE 3. VOLATILE HAPs NOT TESTED FOR CANISTER STABILITY, AND
INDICATIONS OF LIKELY STABILITY
HAPs
Likely Stability
in Canister*
Comments'" . |j
Acetophenone
?
Low vapor pressure, substantial 1
solubility |
Acrolein
?
High water solubility |
Acrylamide
N
Low vapor pressure, high water j
solubility 1
gAcrylic acid
N
High water solubility |
HAniline
?
High water solubility, low vapor
pressure
|bis(ChIoromethy!)etber
N
Reacts with liquid water
flCatechol
N
Substantial water solubility, low vapor
pressure
Hchloroacetic acid
N
High water solubility
Chloromethyl methyl ether
N
Reacts with liquid water
Chloroprene
Y
By comparison to 1,3-butadiene
DCresols/Cresyllic acid
N
Only o-cresol has vapor pressure >0.1
mm Hg
jo-Cresol
N
Minimally volatile, highly soluble
[Diazomethane
N
Reactive
11,2-Dibromo-3-chIoropropane
Y
By comparison to p-dichlorobenzene
iDichloro ethyl ether
Y
By comparison to 1,4-dichlorobenzene,
benzyl chloride, and methyl iso-butyl
ketone
iDiethyl sulfate
N
Reacts with liquid water
19
-------
TABLE 3. (Continued)
HAPs
Likely Stability
in Canistei*
Comments*
|N,N-Diinethylaniline
?
Low vapor pressure
jDimethyl carbamoyl chloride
N
Reacts with liquid water
jN.N-Dimethyl formamide
?
High water solubility 1
11,1-Dimethyl hydrazine
N
High water solubility, reactive I
pimethyl sulfate
N
High solubility and reacts with liquid
water
Epichlorohydrin
?
Substantial water solubility, minimal
previous stability testing of epoxides
1,2-Epoxybutane
?
High water solubility, minimal
previous stability testing of epoxides
lEthyl carbamate
?
Minimal vapor pressure, high water
solubility
REthyleneimine
?
Potentially reactive, high water
solubility
Formaldehyde
?
Soluble, potentially reactive 1
Hexachloroethane
Y
By comparison to hexachlorobutadiene, j
carbon tetrachloride, and other highly |
chlorinated compounds 1
Isophorone
Y
By comparison to benzyl chloride and
nitrobenzene
Methyl hydrazine
?
Reactive, high water solubility
Methyl Iodide
Y
By comparison to methyl chloride and
methyl bromide
¦Methyl isocyanate
N
Reactive
||2-Nitropropane
Y
By comparison to nitrobenzene
20
-------
TABLE 3. (Continued)
HAPs
Likely Stability
in Canister*
Comments'*
N-Nitroso-N-methyl urea
N
Reacts with liquid water
N-N itroso-dimethy 1 amine
7
Soluble and possibly reactive
N-Nitrosomorpholine
7
Minimal vapor pressure, substantial
solubility
Phenol
7
Minima] vapor pressure, substantial
solubility
Phosgene
N
Reactive
1,3-Propane sultone
7
Reacts with liquid water
Ibeta-Propiolactone
?
Reacts with liquid water
IPropionaldehyde
?
By comparison to acetaldehyde
1,2-Propyleneimine
N
Reactive, high water solubility
Styrene oxide
N
Reacts with water
Triethylamine
7
Potentially reactive, substantial
solubility
Xylenes (mixed)
Y
Individual isomers previously tested
Vinyl bromide
Y
By comparison to vinyl chloride
a: Y = Yes, likely to be stable in a canister.
N = No, not likely to be stable in a canister.
7 = Canister stability uncertain, due to absence of some chemical and physical property
data, or due to lack of appropriate compounds with which to make comparisons.
b: Commens summarize quantitative comparisons of several chemical and physical
properties; see text.
21
-------
The present study has foeussed on volatile organic HAPs, and has not considered the
several volatile inorganic compounds also on die HAPs list. The volatile inorganic HAPs
include chlorine, hydrogen fluoride, hydrogen chloride, hydrazine, and phosphine. Although
exhibiting high vapor pressures, these HAPs are generally too reactive, water soluble, and/or
polar to be amenable to canister sampling.
CONCLUSIONS
A review of canister stability tests shows such testing has addressed 52 compounds
designated as HAPs in Title III of the CAAA. For nearly all of the tested compounds,
stability during storage and analysis is sufficient to make canisters an effective sampling
approach. However, additional stability testing under carefully documented conditions is
needed to establish and understand the stability for some of the 52 HAPs. Further
improvements in analytical methods for some HAPs are also needed, because determination
of canister stability is dependent upon analytical measurements.
Comparisons of chemical and physical properties of the HAPs have identified 9
HAPs, never tested for canister stability, which are likely to be stable in canisters. Such
comparisons also identified 17 HAPs for which canister stability will certainly be poor. For
19 other HAPs, available property information and stability testing of analogous compounds
are insufficient to determine the likely effectiveness of canister sampling.
The canister stability tests reported in the literature have employed a great variety of
procedures, and have generally addressed small numbers of compounds with a small set of
canisters. Greater standardization of test procedures, and improved documentation of test
conditions, are needed to integrate test results and allow application of models of chemical
behavior in canisters.
22
-------
ACKNOWLEDGMENT
This work was conducted under the support of the U.S. Environmental Protection
Agency, Contract No. 68-D0-0007. The involvement of William McClenny of the U.S.
EPA in this study is gratefully acknowledged, as are helpful discussions with Robert Coutant.
Although the work described here was funded by the U.S. Environmental Protection Agency,
it has not been subjected to Agency review, and therefore does not necessarily represent the
views of the Agency, and no official endorsement should be inferred.
23
-------
REFERENCES
Allen, J.M., Jayanty, R.K.M., and von Lehmden, D.J. (1987) Comparison of sample loops
constructed of several different materials for gas chromatographic analysis of parts-per-
billion-Ievel organic gas mixtures. Anal. Cham 59, 1982-1984,
Betterton, E.A. (1992) Henry's law constants of soluble and moderately soluble organic
gases: Effects on aqueous phase chemistry, in Gaseous Pollutants: Characterization and
Cvcline. 1.0. Nriagu Ed., John Wiley & Sons, Inc., New York.
Clean Air Act Amendments of 1990, Conference Report to Accompany S. 1630, Report No.
101-952, U.S. Government Printing Office, Washington, DC, 1990, pp 139-162.
Coutant, R.W. and McClenny, W.A. (1991) Competitive adsorption effects and the stability
of VOC and PVOC in canisters. In Proceedings of the 1991 EPA/AWMA Symposium on
Measurement of Toxic and Related Air Pollutants. EPA-600/9-91/018, Publication No. VIP-
21, Air and Waste Management Association, Pittsburgh, PA, pp 382-388.
Coutant, R.W. (1993) Theoretical evaluation of stability of volatile organic chemicals and
polar volatile organic chemicals in canisters. Final Report to U.S. EPA, Contract No. 68-
D0-0007, Work Assignment No. 45, Subtask 2, Battelle, Columbus, OH.
CRC Handbook of Chemistry and Physics (1979) R.C. Weast, ed., 59th Edition, CRC press,
Boca Raton, FL.
Eklund, B., Smith, S., and Hunt, M. (1991) Air/Superfund National Technical Guidance
Study Series: Estimation of air impacts for air stripping of contaminated water, EPA-450/1-
91-002, U.S. Environmental Protection Agency, Research Triangle Park, NC.
Gholson, A.R., Storm, J.F., Jayanty, R.K.M., Fuerst, R.G., Logan, T.J., and Midgett,
M.R. (1989) Evaluation of canisters for measuring emissions of volatile organic air pollutants
from hazardous waste incineration. J. Air Waste Manage. Assoc. 39, 1210-1217.
Gholson, A.R., Jayanty, R.K.M., and Storm, J.F. (1990) Evaluation of aluminum canisters
for the collection and storage of air toxics. Anal. Chem.. 62, 1899-1902.
Hine, J. and Mookerjee, P.K. (1975) The intrinsic hydrophilic character of organic
compounds, correlations in terms of structural contributions, J. Ore. Chem.. 40, 292-298.
Holdren, M., Rust, S., Smith, R., and Koetz, J. (1984) Evaluation of cryogenic trapping as
a means for collecting organic compounds in ambient air. Final report to U.S.
Environmental Protection Agency, Contract No. 68-02-3487, Work Assignment 22, prepared
by Battelle, Columbus, OH.
24
-------
Holdren, M.W., Matthews, M.C., Orban, J.E., and Smith, D.L. (1991) Stability of volatile
organic compounds in Summa polished canisters. Final report to U.S. Environmental
Protection Agency, Contract No. 68-D0-0O07, Work Assignment 2, prepared by Battelle,
Columbus, OH.
Holdren, M.W., Skarpness, B.O., Strauss, W.J., and Keigley, G.W. (1994) Effectiveness of
Silcosteel treatment of canisters. Final report to U.S. Environmental Protection Agency,
Contract No. 68-DO-0007, Work Assignment 45, Task 3, prepared by Battelle, Columbus,
OH.
Howard, P.H., Boethling, R.S., Jarvis, W.F., Meylan, W.M., and Michalenko, E.M. 1991
Handbook of Environmental Degradation Rates. ISBN-0-8371-358-3, Lewis Publishers,
Chelsea, MI.
Jayanty, R.K.M. (1989) Evaluation of sampling and analytical methods for monitoring toxic
organics in air. Atmos. Environ.. 23, 777-782.
Jones, D.L. and Bursey, J. (1992) Simultaneous control of PM-10 and Hazardous Air
Pollutants, II: Rationale for selection of Hazardous Air Pollutants, II: Rationale for selection
of Hazardous Air Pollutants as potential particulate matter, EPA-452/R-93/012, U.S.
Environmental Protection Agency, Research Triangle Park, NC.
Keith, L.H. and Walker, M.M. (1992VEPA's Clean Air Act Air Toxics Database. Volume
I: Air Toxics Chemical and Physical Properties. ISBN-0-87371-819-4, Lewis Publishers,
Boca Raton, FL.
Keith, L.H. and Walker, M.M. (1993) EPA's Clean Air Act Air Toxics Database. Volume
II: Air Toxics Chemical and Physical Properties. ISBN-0-87371-820-8, Lewis Publishers,
Boca Raton, FL.
Kelly, T.J., Callahan, P.J., Pleil, J.D., and Evans, G.E. (1993) Method development and
field measurements for polar volatile organic compounds in ambient air. Environ. Sci.
Technol.. 27, 1146-1153.
Kelly T.J., R. Mukund, Gordon, S.M., and Hays, M.J. (1994) Ambient measurement
methods and properties of the 189 Title III Hazardous Air Pollutants. Final report to the
U.S. Environmental Protection Agency, Contract No. 68-D0-0007, Work Assignment 44,
prepared by Battelle, Columbus, OH, March 1994.
MacKay, D. and Shiu, W-Y. (1981) A critical review of Henry's law constants for chemicals
of environmental interest, J. Phvs. Chem. Ref. Data. 10, 1175-1199.
Mackay, D., Shiu, W.Y., and Ma, K.C. (1993) Illustrated Handbook of Phvsical-Chemical
Properties and Environmental Fate for Organic Chemicals. Volume III: Volatile Organic
Chemicals. ISBN-0-83731-973-5, Lewis Publishers, Chelsea, MI.
25
-------
McCIenny, W.A., Pleil, J.D., Evans, G.F., Oliver, K.D., Holdren, M.W., and Winberry,
W.T. (1991a) Canister-based method for monitoring toxic VOCs in ambient air. J. Air Waste
Manage. Assoc.. 41, 1308-1318.
McCIenny, W.A., Evans, G.F., Oliver, K.D., Daughtrey, E.H., Jr., Winberry, W.T., Jr.,
and Decker, D.L. (1991b) Status of VOC methods development to meet monitoring
requirements for the Cleaii Air Act Amendments of 1990. In Measurement of Toxic and
Related Air Pollutants, proceedings of the 1991 EPA/AWMA International Symposium,
Report No. EPA/600/9-91/018, publication number VIP-21, Air and Waste Management
Association, Pittsburgh, PA, pp. 367-374.
Merrill, R.G., Jr., and Zapkin, M.A. (1991) Application of canister sampling and analysis
methodology to volatile polar organic compounds. In Measurement of Toxic and Related Air
Pollutants, proceedings of the 1991 EPA/AWMA International Symposium, Report No.
EPA/600/9-91/018, publication number VIP-21, Air and Waste Management Association,
Pittsburgh, PA, pp. 602-606.
Oliver, K.D., Pleil, J.D., and McCIenny, W.A. (1986) Sample integrity of trace level
volatile organic compounds in ambient air stored in Summa polished canisters. Atmos.
Environ.. 20, 1403-1411.
Oliver, K.D. (1993) Sample integrity of trace level polar VOCs in ambient air stored in
summa-polished canisters. Technical Note TN-4420-93-03, submitted to U.S. EPA under
Contract No. 68-D0-0106, by ManTech Environmental Technology, Inc., Research Triangle
Park, NC.
Parmar, S.S. (1991) A study of holding times for H2S, COS, and CH3SH samples (gas
phase) in different containers. In Measurement of Toxic and Related Air Pollutants,
proceedings of the 1991 EPA/AWMA International Symposium, Report No. EPA/600/9-
91/018, publication number VIP-21, Air and Waste Management Association, Pittsburgh,
PA, pp 544-551.
Pate, B., Jayanty, R.K.M., Peterson, M.R., and Evans, G.F. (1992) Temporal stability of
polar organic compounds in stainless steel canisters, J. Air Waste Manage. Assoc.. 42, 460-
462.
Sansone, E.B., Tewari, Y.B., and Jonas, L.A. (1979) Prediction of removal of vapors from
air by adsorption on activated carbon. Environ. Sci. Technol. 13, 1511-1513.
Spicer, C.W., Pollack, A.J., Kelly, T.J., and Mukund, R. (1993) A literature review of
atmospheric transformation products of Clean Air Act Title III Hazardous Air Pollutants,
Final Report to U.S. Environmental Protection Agency, Contract No. 68-D80082, Battelle,
Columbus, OH.
Verschueren, K. (1983) Handbook of Environmental Data on Organic Chemicals. Van
26
-------
Nostrand Reinhold Company, New York.
Weber, R.C., Parker, P.A., and Bowser, M. (1981) Vapor pressure distribution of selected
organic chemicals, EPA-600/2-81/021, U.S. Environmental Protection Agency, Cincinnati,
OH.
Westberg, H.H., Holdren, M.W., and Hill, H.H., Jr. (1981) Analytical methodology for the
identification and quantification of vapor phase organic pollutants. Final report to the
Coordinating Research Council, Project No. CAPA-11-71, prepared by Washington State
University, Pullman, WA.
Winberry, W.T., Jr., Murphy, N.T., and Riggin, R.A. (1988) Method TO-14. In
Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient
Air. Report No. EPA-600/4-89-017, U.S. Environmental Protection Agency, Research
Triangle Park, NC.
Yaws, C., Yang, H-C., and Pan, X. (1991) Henry's law constants for 362 organic
compounds in water, Chem. Eng.. pp 179-185, November.
27
-------
TECHNICAL ftiPOft? DATA
PUmt rmi huimtnom aw tht it«mt mfon tvmpittln
1.WMORT NO. »•
EPA/600/A-94/239
l.» —~
i.TiTit A*P»U»TiTL«
Applicability of Canister Sampling for Hazardous
Air Pollutants
I.MRPORMIMO eRfiANlZATlBN COB!
T.AUTMOftlSi
T.J. Kelly and M.W. Holdren, Battelle, Columbus, OH
1. PBMPOMMINO OMBAMiZATieti nt»b*f HQ.
|, P|R'0»Miha OAfiANlZAIiOM MAUI AND AO0RKU
Battelle Laboratories
505 King Avenue
Columbus, OH 432,01.
#R06RAM titutw! kb.
1X.S»OKSORIMO A0INCV MAUI AND ABBRitt
US Environmental Protection Agency
Research Triangle Park, NC 27711
ts.vvri m*ort anb uriodcovirib"
M. fOMSoAiNO aoinct coot
It. 10^4.1 WlKTART NOT It
Ift. AUTRACI
==> The report evaluates the applicability of sampling with evacuated canisters for
volatile organic compounds listed among the 189 Hazardous Air Pollutants (HAPs)
in the 1990 US Clean Air Act Amendments. Nearly 100 HAPs have sufficient vapor
pressure to be considered volatile compounds. Of those volatile organic HAPs,
52 have been tested previously for stability during storage in canisters. The
published HAP stability studies are reviewed, illustrating that for nearly all
of the 52 HAPs tested, canisters are an effective air sampling approach.* However,
the published stability studies used a variety of canis.ter._tv.pes-.and—t-e"st pro-
cedures, and-genera1ly considered only a few compounds in a very small set of
canisters?-*'A comparison of chemical and physical properties of- the HAPs has been
conducted, to evaluate the applicability of"canister sampling for other HAPs, for
which canister stability testing has hever been conducted.<^jOf 45 volatile HAPs
never tested in canisters, this comparison identifies 9 for\?hich canister
sampling should be effective, and 17 for which canisters are "not likely to be
effective. For the other 19 HAPs,-no clear decision ban be reached on the
likely applicability of canister sampling.
It, KIV WORM AND OOCUMtWT ANAWWB
i DticRtrroRI
k.lBINTIf ltR»/onN IXDID TtRMX
c. CO(ati F«Id 'Croup
H OUTRllJTlOS ITATIMtN?
RELEASE TO PUBLIC
11 SICURIT T CkASi (Tlhj Htmorlf
UNCLASSIFIED
Jt. NO. O* '*61*
»e ii ewRiT* ci.au {Thu putt/
UNCLASSIFIED
UM'SI
|?A Pant 1U8-1 (••«. •• (UlbKI
------- |