&EFK
United States
Environmental Protection
Agency
Volume 5"
Great Lakes National
Program Office
536 South Clark Street
Chicago, Illinois 60605
EPA-905/4-79-029-1
(LA
The IJC Menomonee
River Watershed Study
Atmospheric Chemistry
Of PCBs and PAHs
Menomonee River
-------
FOREWORD
The Environmental Protection Agency was established to coordinate adminis-
tration of the major Federal programs designed to protect the quality of our
environment.
An important part of the Agency's effort involves the search for information
about environmental problems, management techniques, and new technologies
through which optimum use of the nation's land and water resources can be
assured and the threat pollution poses to the welfare of the American people
can be minimized.
The Great Lakes National Program Office (GLNPO) of the U.S. EPA, was
established in Region V, Chicago to provide a specific focus on the water
quality concerns of the Great Lakes. GLNPO also provides funding and
personnel support to the International Joint Commission activities under
the U.S.- Canada Great Lakes Water Quality Agreement.
Several land use water quality studies have been funded to support the
pollution from Land Use Activities Reference Group (PLUARG) under the
Agreement to address specific objectives related to land use pollution to
the Great Lakes. This report describes some of the work supported by this
Office to carry out PLUARG study objectives.
We hope that the information and data contained herein will help planners
and managers of pollution control agencies make better decisions for
carrying forward their pollution control responsibilities.
Madonna F. McGrath
Director
Great Lakes National Program Office
-------
EPA-905/4-79-029-I
March 1980
Atmospheric Chemistry of PCBs and PAHs
Volume 9
by
A.M. Andren
P.V. Doskey
and
J.W. Strand
Wisconsin Water Resources Center
and
Water Chemistry Program
University of Wisconsin - Madison
for
U.S. Environmental Protection Agency
Chicago, Illinois
Grant Number R005142
Grants Officer
Ralph G. Christensen
Great Lakes National Program Office
This study, funded by a Great Lakes Program grant from the U.S. EPA, was
conducted as part of the TASK C-Pilot Watershed Program for the International
Joint Commission's Reference Group on Pollution from Lake Use Activities.
GREAT LAKES NATIONAL PROGRAM OFFICE
ENVIRONMENTAL PROTECTION AGENCY, REGION V
536 SOUTH CLARK STREET, ROOM 932
CHICAGO, ILLINOIS 60605
U.S. Environmental Protection Agew
,
7? West Jackson Bpufevard, 12th FtoBT
Chicago. IL 606044590
-------
DISCLAIMER
This report has been reviewed by the Great Lakes National Program
Office of the U.S. Environmental Protection Agency, Region V Chicago,
and approved for publication. Mention of trade names of commercial
products does not constitute endorsement or recommendation for use.
ii
^H&Affcifca'tort fefnvn;:.
-------
PREFACE
The significance of atmospheric input of PCBs and PAHs to
Lake Michigan is assessed utilizing dry and wet deposition information
obtained over the lake's surface. Possible cycling of these compounds
over Lake Michigan is discussed.
iii
-------
CONTENTS
Title Page i
Disclaimer ii
Preface iii
Contents iv
*Part I Transport of Airborne PCBs to Lake Michigan I-i
*Part II PAHs in air over Lake Michigan Il-i
*Detailed contents are presented at the beginning of each part.
iv
-------
PART I
TRANSPORT OF AIRBORNE PCBs
TO LAKE MICHIGAN
P, V, DOSKEY
A, W, ANDREN
I-i
-------
ABSTRACT
The air over Lake Michigan was sampled during 1977 to develop a
collection method for PCBs and obtain data about their atmospheric transport
and dry deposition onto the lake. A resin, XAD-2, was the most efficient
collection medium for PCB vapor and was incorporated into standard high volume
air samples for the collection of particulate and vapor phase PCBs. PCB
concentrations in air samples taken over Lake Michigan were lower than those
taken from urban areas; i.e., Milwaukee. Aroclors 1242 and 1254 were the main
components of vapor phase PCBs while in some instances the particulate phase
PCBs contained Aroclor 1260. The particulate phase PCBs over Lake Michigan
contained a larger percentage of the more volatile mixtures than those
reported in urban areas such as Chicago and Milwaukee. PCBs tend to associate
with particulates 0.002 to 0.1 ym in diameter. The amount and organic carbon
content of the particulate phase appear to control vaporization and
revolatilization of PCBs.
The net atmospheric input to Lake Michigan based upon the liquid and gas
phase control models was 2848 and 8655 kg/yr, respectively. Bubble ejection
was not found to be a probable water to air transport process over the lake.
It is estimated that the atmopshere contributes >70% of the PCB load to Lake
Michigan,
I-ii
-------
CONTENTS-PART I
Title Page I-i
Abstract I-ii
Contents I-iii
Figures I-iv
Tables I-v
1-1. Introduction • 1-1
1-2. Conclusions 1-2
1-3. Background 1-4
General Information 1-4
Collection Methods 1-6
Air Concentrations and Composition 1-8
Nature of Airborne PCBs 1-10
1-4. Sample Cleanup and Quantification. 1-13
XAD-2 Cleanup 1-13
Sample Cleanup 1-15
Sample Quantification 1-15
1-5. Sampling 1-22
1-6. Sample Results and Discussion 1-26
Comparison and Interpretation of Air Concentrations
and Composition 1-26
Vaporization 1-35
Experimental 1-37
Results and discussion. 1-38
Volatilization 1-40
Experimental 1-41
Results and discussion 1-43
Conclusions ..*..«..«.< 1-51
1-7. Flux Calculations 1-53
Washout Coefficients and the Wet Flux 1-53
Particulate Flux 1-60
Vapor Flux 1-60
Bubble Ejection 1-70
Photolysis 1-71
References 1-73
Appendix I-A 1-79
XAD-2 Cleanup Procedure 1-79
Sample Cleanup Procedure 1-79
I-iii
-------
FIGURES
Number Page
1-1 Aroclor 1242 in XAD-2 extract before and after alumina
cleanup 1-14
1-2 Lake Michigan air sample and Aroclor 1242 (175°C) 1-17
1-3 Lake Michigan air sample and Aroclor 1254 (205°C).. 1-18
1-4 lake Michigan air sample and Aroclor 1248 (185°) 1-19
1-5 Lake Michigan air sample and Aroclor 1248 (205°) 1-20
1-6 Lake Michigan cruise tracks 1-23
1-7 Milwaukee sampling locations 1-25
1-8 Wind roses for Lake Michigan sampling stations 1-31
1-9 KQIj against time for Aroclors 1016 and 1221 1-44
1-10 KQL against time for Aroclors 1242 and 1254 for acetone
spike experiment 1-45
1-11 KQL against time for Aroclors 1242 and 1254 in fly ash
experiment 1-46
1-12 Loss of PCBs (%) with time for acetone spike experiment 1-48
1-13 Loss of PCBs (%) with time for fly ash experiment 1-49
1-14 Collision efficiency against particle size 1-57
1-15 Deposition velocity against particle size... 1-61
I-iv
------- OCR error (C:\Conversion\JobRoot\000002QG\tiff\200072P2.tif): Unspecified error
-------
1-21 PCB vaporization 1-39
1-22 Liquid and gas phase mass transfer coefficients for PCBs 1-42
1-23 Liquid phase mass transfer coefficients (m/hr) for
Aroclors 1221 and 1016 and Aroclors 1242 and 1254 1-42
1-24 Standard errors for revolatilized PCB vapor 1-50
1-25 Composition of Aroclor mixtures 1-50
1-26 Experimentally and theoretically derived Henry's Law
constants (H) and liquid phase mass transfer
coefficients (KQL) 1-52
1-27 Vapor phase washout coefficients and Henry's constants
for Aroclor 1242 and 1254 1-55
1-28 Atmospheric and laboratory determined washout ratios for
pesticides 1-55
1-29 Collision efficiencies and washout ratios for different
particle sizes 1-58
1-30 PCB washout coefficients and particle sizes 1-58
1-31 Predicted PCB concentration in precipitation 1-58
1-32 Vapor flux to Lake Michigan (F = KGCfe) 1-63
1-33 Henry's constant (H) and gas phase mass transfer
coefficients (KQG) for Aroclors 1-63
1-34 Vapor flux to Lake Michigan [N = KOG(CH-P)/RT] 1-64
1-35 Revolatilization rates from Lake Michigan 1-66
1-36 KQL against wind speed 1-68
1-37 Revolatilization rates from Lake Michigan against
wind speed 1-69
1-38 Microlayer residence times against revolatilization
rates 1-68
1-39 Estimated net atmospheric inputs to Lake Michigan based
on gas and liquid phase control model 1-72
1-40 Inputs of PCBs (kg/yr) to Lake Michigan 1-72
I-vi
-------
1-1. INTRO.DUCTION
Polychlorinated biphenyls (PC3s) have become an environmental
problem in Lake Michigan. Concentrations in lake trout, exceeding the
5 ppm standard set by the FDA,
trations in Lake Michigan fish
fishing industry on the lake.
Monsanto, the sole produce
voluntarily banned the sale of
Even though Monsanto no longer
United States will not come int
Substances Control Act of 1976,
continue to escape into the em
formerly used.
The detection of PCBs in
suggests that the atmosphere IE
this hypothesis, in the Great
mination of PCBs in precipitation
significant contribution to the
(Table 1-1).
Antarctic snow samples (1) strongly
a major transport route. Proof of
l),akes region, has come from the deter-
over Lake Michigan (2) and its
present PCS loading of the lake
Any study of the cycling c
ment must deal with its source,
initiated to obtain a better ur
the atmosphere and their dry de
The main objectives of thi
1. To develop a method oi
was both qualitative and quant
2. To emoloy the method t
trations of PCBs in air over L
3. To utilize the data tc
Michigan.
Table 1-1. Inputs of PCBs
Source
Industrial discharges
Precipitation
Streams and wastewater
Others*
TOTAL
are frequently observed. High PCB concen-
have caused a closing of the commercial
r of PCBs in the United States,
PCBs for use in "open" systems in 1971.
produces PCBs, the ban on PCBs in the
o effect until July 1, 1979 (Toxic
PL94-46). After the ban, PCBs will
ironment from products in which they were
f a trace contaminant in the environ-
transport, and sink. This study was
derstanding of the transport of PCBs in
position onto Lake Michigan.
s study were:
collection for atmospheric PCBs which
tative.
o obtain information about the concen-
.ke Michigan.
estimate the dry flux of PCBs to Lake
(kg/yr) to Lake Michigan (2)
Prior to 1975
1977
effluent
25,000
4,800
750
2,750
33,300
*Accounts for dry deposition and unknown industrial discharges.
1-1
-------
1-2. CONCLUSIONS
Laboratory experiments performed with XAD-2, polyurethane foam,
polyurethane foam coated with DC-200 and florisil, showed XAD-2 to be
the most efficient collection medium for PCB vapor and well-suited for
high volume air sampling. A standard high volume air sampler was
modified to incorporate XAD-2 resin for the collection of particulate
and vapor phase PCBs. The sampler is 90 to 100% efficient for
collection of even the most volatile vapor phase PCB mixtures. Retention
efficiencies depend upon the volatility of the isomers with Aroclor
1221 being retained at an average of 83%.
Air samples taken over Lake Michigan were less concentrated in
PCB than those taken in urban areas such as Milwaukee. Vapor phase
PCBs were mainly composed of Aroclors 1242 and 1254 while the particulate
phase PCBs also contained Aroclor 1260 in some instances. The par-
ticulate phase PCBs over Lake Michigan contained a larger percentage
of the more volatile mixtures than those reported in urban areas such
as Chicago or Milwaukee. Adsorption theory based upon the volatility
of PCBs and amount of particulate surface area fails to explain this
finding. The higher proportion of PCBs in the particulate phase over
Lake Michigan and Milwaukee as compared to the ocean is probably
a result of the higher concentration of particulate matter near urban
centers.
Vaporization of PCBs from PCB coated particulates placed on a
glass fiber filter did not occur- under conditions where the particulate
matter contained large amounts of organic carbon. While the experiment
did not give entirely conclusive evidence, the desorption kinetics
under environmental conditions should be slow since particulates
collected over Lake Michigan contained large percentages of the more
volatile PCBs.
Calculations based on the revolatilization experiment produced air/
water partition coefficients for Aroclors 1242 and 1254 which were
orders of magnitude less than the theoretically determined Henry's Law
constants. The coefficients were in the same range as those reported
for saturated water solutions of Aroclors 1221 and 1016. Theoretical
calculations of the vapor phase washout coefficients for 1242 and 1254
showed them to be the same order of magnitude as the air/water partition
coefficients. The magnitude of the coefficients indicated gas phase
control of PCBs. The revolatilization rate for PCBs was found to be
slower under conditions where particulate matter was present, showing
that if PCBs are liquid phase controlled, environmental revolatilization
rates cannot be estimated in filtered-organic free aquatic environments.
Theoretical calculations based upon washout coefficients indicated
that PCBs are associated with particulates between 0.002 and O.lp in
diameter. The calculations could not conclusively predict the existence
of vapor phase PCBs but the best estimates of wet fluxes from air
concentrations can be made by employing both a particulate and vapor
phase washout coefficient.
1-2
-------
Due to the scarcity of experimental evidence for gas phase control
of PCBs, atmospheric fluxes to Lake Michigan based upon both a liquid
and gas phase control model were calculated. Bubble ejection of PCBs
was not an important water to air transport process in Lake Michigan.
More experimental evidence is needed to-estimate the alteration of
airborne PCBs by photolysis. The net atmospheric input based upon the
liquid and gas phase control models was 2848 and 8655 kg/yr respectively.
At the present time the atmosphere is estimated to be contributing
greater than 70% of the PCB load to Lake Michigan. These estimates show
that futher investigations of the transport of PCBs must be made.
Futher studies should involve a determination of the vapor/particu-
late partitioning of airborne PCBs, determinations of Henry's Law
constants under simulated environmental conditions, and estimations of
the importance of photolytic processes on atmospheric PCBs. The effects
of the atmospheric transport of PCBs on composition, vapor/particulate
partitioning, and deposition rates should also be explored so accurate
estimates of fluxes can be made.
1-3
-------
1-3. BACKGROUND
General Information
.PCBs have had a widespread application_due to their, general
stability, inertness and dielectric properties. In the United States
they were produced by the Monsanto Company under the trade name Aroclor.
Aroclors are mixtures of PCS isomers which are designated by a four
digit number. The first two digits give the molecule type, while the
second two give the weight percentage of chlorine. For example, Aroclor
1242 is a mixture of biphenyl molecules with 42% chlorine by weight.
A listing of the Monsanto Aroclor products along with their molecular
weights and average chlorine numbers can be found in Table 1-2.
Some of the products in which PCBs were used include: transformer
fluids, hydraulic fluids, PCB-impregnated paper in capacitors, caulking
compounds, carbonless duplicating paper, plasticizers, adhesives
and wax extenders. Nisbet and Sarofim (3) and Hutzinger et al. (4)
give a complete breakdown of their former uses. Versar (5) compiled a
summary of the total amount of Aroclors used in different products from
1930-1974 and also, the environmental load by Aroclor type. The most
widely used Aroclors were 1242, 1248, 1254 and 1260. Aroclor 1016
was used after 1970 but to a much lesser extent than the other four.
Polychlorinated biphenyls have been found in virtually every
compartment of the environment. The atmosphere is one of the major
compartments and is probably the most important transport route, at
least on a global basis. This is exemplified by their discovery in
Antarctic snow samples (1).
There are several possible routes into the atmospheric environment.
One route is vaporization from PCB containing formulations. Also, due
to their widespread utie in many products, they can be volatilized
during incineration in poorly operated commercial and municipal incin-
erators. Incineration at 2000+°F for 2 sec will destroy PCBs. They
can also volatilize from products buried in sanitary landfills.
Murphy and Rzeszutko (2) have determined the concentration of PCBs in
a sanitary landfill's vent gases. The concentration was much higher
than those normally found in ambient air, making landfills a definite
PCB source. Other sources of polychlorinated biphenyls include small
domestic and apartment incinerators and open burning dumps. Table 1-3
contains Nisbet and Sarofim's (3) estimates of the rates of input of
PCBs to the North American continent. They estimated the atmospheric
emission rate of PCBs to be 1.5 to 2 x 10^ Tonnes/yr, amounting to a
vaporization of 30,000 Tonnes since 1930.
An examination of the physical properties of PCBs may give some
insight as to why their mode of transport is important. PCBs can
theoretically exist as 209 different isomers (4). Two important
physical properties, from an environmental standpoint, are their vapor
pressures and solubilities. The solubility of the isomers decreases
with increasing chlorine content (6,7). Vapor pressures of some
Aroclor mixtures can be found in Table 1-4. Hutzinger et al. (4)
give a complete listing of the physical and chemical properties of PCBs.
1-4
-------
Table 1-2. Average chlorine numbers and molecular weight of
Aroclor mixtures (4)
Aroclor
Average Cl No/molecule
Molecular wt(
1221
1232
1016
1242
1248
1254
1260
1262
1268
1.15
2.04
2.96
3.10
3.90
4.96
6.30
6.80
8.70
192
221
256
261
288
327
372
389
453
Table 1-3. Gross estimates of rates of input and accumulation of PCBs in
1970 (3)
Input
Rates, tonnes/yr
PCB
Vaporization of plasticizers
Vaporization during open burning
Leaks and disposal of industrial fluids
Destroyed by incineration
Disposal in dumps and landfills
Accumulation in service
1 to 2 x 103
4 x 102
4 to 5 x 10;
3 x 103
18 x 103
7 x 103
1248 to 1260
Mainly 1242
1242 to 1260
Mainly 1242
1242 to 1260
1242 to 1254
Table 1-4. Vapor pressures and saturation vapor densities of Aroclor mixtures
ry
Aroclor Temp, °C Vapor pressure,* mm Hg Sat. vapor density,** mg/mm
1242
1248
1254
1260
25
25
25
25
4 x 10~4
5 x 10~4
8 x 10~4
4 x 10~5
5.5
7.9
1.4
0.73
*Mackay and Wolkoff (Ref. 8)
**Calculated using the ideal gas law (PV = nRT).
1-5
-------
A calculation of the saturation vapor densities of PCBs from their
vapor pressures and the ideal gas law indicates that PCBs could be
present in the atmosphere at extremely high concentrations (Table 1-4).
This is only a theoretical calculation and does not necessarily indicate
the volatility of the PCS mixtures in nature. It would be expected
that the more volatile mixtures would be present in the atmosphere.
Mieure et al, (9) have shown that the vapors above a standard Aroclor
1016 mixture were enriched in the more volatile components. This points
out that the atmospheric component of the mixture may bear no resem-
blance to the original Aroclor product. This fractionation makes
interpretation of the data extremely difficult. The affinity of PCBs
in aqueous solutions for surfaces (4) would seem to indicate that they
would be adsorbed to particulates but the extent to which this occurs
in the atmosphere is presently uncertain.
Collection Methods
Many methods have been employed to measure the concentrations of
PCBs in air. Three important facts must be kept in mind when designing
a collection method:
1. PCBs are a mixture of isomers having variable volatilities.
2. Atmospheric samples most likely contain fractionated PCBs
which do not reflect the composition of the original Aroclor mixtures.
3. Air concentrations are usually in the sub-part per trillion
range.
Keeping these factors in mind, certain criteria must be met by the
collection method so both a qualitative and quantitative sample can
be obtained:
1. The sampler must have a high collection efficiency for all
the PCB isomers, even the most volatile species.
2. Since the concentrations in air are usually in the nanograms
per cubic meter range, a substantial flow rate through the collection
medium is essential to avoid extremely long sampling periods.
3. It must be relatively easy to recover the PCBs from the
collection medium since only a small quantity is collected and lengthy
procedures would allow more opportunity for loss of sample.
4. Interfering substances from the blank medium must be few to
minimize difficulties in quantification and interpretation of gas
chromatograms.
Many methods have been employed to collect PCBs from air but
at the present time none is uniformly accepted. The methods can be
grouped into two categories. One category is designed to measure air
1-6
-------
concentrations. These methods employ pumps to force air through a
collection medium. The other category contains methods to measure
fluxes of PCBs. An adsorbent medium is coated on a surface, such as
glass, onto which PCBs deposit.
One type of forced air method employs impingers containing
ethylene glycol. These samplers have been found to be adequate in
areas containing high levels of pesticides (10). Both their low
collection efficiency and flow rates are inadequate for high volume air
sampling (11). Two types of high volume air samplers have been tested.
These employ a filter for collecting particulates and a chamber behind
the filter holder assembly containing an organic vapor trap. Harvey
and Steinhauer (12) used ceramic saddles coated with silicone oil to
collect PCB vapor. They reported a collection efficiency of 70% for
either tri- and hexachlorobiphenyl or Aroclor 1254. Bidelman and
Olney (11) used polyurethane foam plugs as a vapor trap. Collection
efficiencies of 96 to 99% were obtained in laboratory experiments
using tri-, tetra-, and pentachlorobiphenyl isomers. In actual air
sampling experiments collection efficiencies of 79 to 99% were
reported. Environmental Science and Engineering, Inc. (13) also
evaluated the polyurethane foam method and reported collection effi-
ciencies of 85% or greater. Lewis et al. (14) reported polyurethane
foam collection efficiencies of 70 to 85% using Aroclor 1242. These
researchers found that the collection efficiency was not uniform for
all the components of a 1242 mixture. The range was 41 to 95%, the
more volatile components being retained less efficiently.
An examination of the polyurethane foam collection efficiencies
shows that they are both inconsistent and not extremely quantitative.
The polyurethane foam method has been used quite extensively but the
reported collection efficiencies are rather low in some cases.
Collection methods have also been employed to measure PCB fluxes.
Nylon screens coated with silicone oil have been used to collect PCBs
and pesticides, but their collection efficiency is unknown (15,16,17,18),
Glass plates coated with a 5:1 mixture of hexane:mineral oil were used
in California to collect PCBs (19). A collection efficiency of 50%
was obtained for Aroclor 1254 over a week long period. Aroclor 1242
was volatilized from the plates during this period. The ability of
these types of collection methods to measure a "natural" PCB flux
must come under question. If PCBs are present mainly in the vapor
phase, as has been conjectured by Bidelman and Olney (20) and Harvey
and Steinhauer (12) , samplers using coatings of silicone oil cannot
be used to measure deposition rates. Due to the organic coating's
high partition coefficient, samplers of this design efficiently
scavenge PCB vapor as well as particulates. These sampling surfaces
cannot reflect a natural surface such as turf or water. This sampling
method is still adequate for measuring and comparing source strengths.
From the preceding discussion it is evident that many methods
have been employed but PCB sample collection methodologies can still
be improved upon.
1-7
-------
Air Concentrations and Composition
Atmospheric concentrations and airborne fallout have been reported
for a number of areas. Concentrations are highest in urban areas and
particularly near industrial users.
Concentrations in remote areas are usually in the sub-nanogram
per cubic meter range. Sampling air in Bermuda, Bidelman and Olney (11)
found concentrations of 0.30 to 0.65 ng/m (as 1248) while Harvey and
Steinhauer (12) found a concentration of 0.5 ng/m3 (as 1254). Concen-
tration ranges over the North Atlantic have also been determined.
Bidelman and Olney (20) found concentrations ranging from 0.2 to 1.6
ng/m3 (as 1242 or 1248). Harvey and Steinhauer (12) found an exponential
increase in PCB concentrations in air as they proceeded toward the New
York Metropolitan area. At 2000 km from the coast (Grand Banks) they
found ranges of 0.05 to 0.16 ng/m3 (as 1254). Off Bermuda, concentra-
tion ranges of 0.15 to 0.4 ng/m3 (as 1254) were observed while PCB
concentrations in air 500 km from New York City (Georges Banks) ranged
from 0.58 to 1.60 ng/m3 (as 1254).
Bidelman (21) collected a number of samples over the North
Atlantic during the period 1973 to 1977 (Table 1-5). Examination of
the results indicates that the concentrations have decreased. The
restriction by Monsanto in 1970 on the sale of PCBs for use only in
closed systems may have contributed to the decrease. A certain lag
period may have taken place, the length of which depended upon the
atmospheric residence time of the mixtures. The change in composition
of the air samples from Aroclor 1248 to 1254 is more difficult to explain.
The environmental loads of Aroclor 1248 and 1254 during the period
1930 to 1974 are very nearly equal (5). The change may be due to a
different release rate of the two mixtures, dependent upon the type
of products in which they were used. Another explanation may be that
the deposition rate of Aroclor 1248 is much greater than that of 1254.
A seasonal trend for Bidelman's samples was not apparent.
Concentrations of PCBs in urban air are usually an order of
magnitude higher than those found in remote areas. On the University
of Rhode Island campus and urban Providence, Rhode Island, Bidelman
and Olney (20) found concentrations ranging from 2.1 to 9.4 ng/m .
Harvey and Steinhauer (12) found concentrations of 3.9 to 5.3 ng/m3
(as 1254) in Vineyard Sound located near the New York metropolitan
area. Murphy and Rzeszutko (2) measured concentrations of 3.6 to 11.0
ng/m3 in urban Chicago air. At least 85% of each sample was found to
most closely resemble an Aroclor 1242 mixture. Air sampled in
Jacksonville, Florida was found to contain 3 to 36 ng/m3 (13). The
samples most closely resembled Aroclors 1242 and 1254.
Atmospheric concentrations near industrial users of PCBs1 are
higher than those normally found in urban areas. Concentrations near
a General Electric facility in New York were found to be in the range
of 20 to 4000 ng/m3 (22). After ceasing to use PCBs at their facilities,
average levels dropped below 300 ng/m3. Average air concentrations
around dump and contaminated dredge spoil areas were 3240, 2160
and 107 ng/m3 (as 1016). Environmental Science and Engineering, Inc. (13)
collected samples around a major industrial user and found concentra-
tions to be in the range of 300 to 498 ng/m3 (as 1016). Tables 1-6,
1-8
-------
Table 1-5. Trends in concentration of PCBs (as Aroclor
1248) in N. Atlantic air (Bermuda)—seasonal
and long term (21)
o
Time In air, ng/m
April 73-Sept. 73 0.76 ± 0.55(5), 4 were <0.33
Oct. 73-March 74 0.51 ± 0.53(7), 5 were <0.24
April 74-Sept. 74 0.20 ± 0.09(19)
Oct. 74-March 75 No samples taken
April 75-Sept. 75 0.076 ± 0.063(6)
Oct. 74-March 76 No samples taken
April 76-Sept. 76 0.077 ± 0.035(11)*
Oct. 76-March 77 0.074 ± 0.099(11)*
*These samples matched Aroclor 1254 more closely.
Table 1-6. PCB concentrations in marine air (23)
o
Location and date In air, mg/m Ref.
Bermuda, 1973
Bermuda, 1973
Bermuda, 1974
Cruises, Bermuda-U.S. , 1973-74
Chesapeake Bay, 1973
Grand Banks, 1973
Georges Banks, 1973
0.
0.
0.
<0.
1.
0.
0.
15
19
08
05
0
05
58
to
to
to
to
to
to
to
0.
0.
0.
1.
2.
0.
1.
5
66
48
6
0
16
6
11
11
11
11
12
,20
,20
,20
,20
12
12
1-9
-------
1-7, and 1-8 contain a summary of the air concentrations cited in
this section and also additional data from Bidelman et al. (23).
Attempts to measure PCB fallout have also been made. Measure-
ments in Sweden showed a range of 620 to 10,510 ng/m^-month (17).
Fluxes determined by Bengston and Sodergren (18) in Iceland showed
a range of 40 to 1050 ng/rn2-month. McClure and Lagrange (24) measure^
an average Aroclor 1254 flux at La Jolla, California of 3 x 10~'
g/m2-day 9 yg/m . For reasons stated in the collection
methods section, these numbers should not be used to represent
natural PCB fluxes. A summary of the data can be found in Table 1-9.
The reported data definitely shows that concentrations are lowest
in remote areas and highest in urban areas, particularly near industrial
users. Since there are many ways of collecting and quantifying air
samples, extreme caution must be taken when comparing results.
Differences in quantification methods will give variable results. The
air composition data show that samples were reported predominantly as
Aroclors 1242, 1248 and 1254. The determination of the composition of
an air sample is very subjective unless a systematic method of
chromatogram division such as the Webb and McCall (25) method is
used. Samples can still be used to determine trends and comparisons
can be made if methods of collection and quantification are similar.
Nature of Airborne PCBs
Determinations of PCB concentrations in air by different high
volume collection methods have shown that greater than 90% of the air-
borne PCBs pass through a glass fiber filter and are collected on a
PCB adsorbent material (2,11,12,20). Whether these PCBs are associated
with particles too small to be collected by a glass fiber filter or
are actually present in ambient air as vapor is uncertain. The problem
is compounded by the hypothesis that PCBs initially associated with
particulates can be vaporized during collection, become entrained in
the flow and, subsequently, be collected as vapor (2,12,20). Operation-
ally defined particulate and vapor concentrations have been determined,
but the true nature of atmospheric PCBs has not been investigated
extensively.
Theoretical calculations have been made to estimate the amount
of PCBs present in the particulate phase. Murphy and Rzeszutko (2)
used the composition of their rain and air samples to estimate the
vapor/particulate partitioning. Two assumptions were made. The first
was that all the Aroclor 1260 in air is associated with particulates.
This is a logical assumption since the vapor pressure of Aroclor 1260
is approximately an order of magnitude lower than the vapor pressures
of the lighter mixtures (1242, 1254). The second was that all PCBs in
rain are due to particulate washout. This assumption is based upon
the Henry's Law constant of Aroclor 1260. The previously reported
Henry's Law constant (9) predicts that the concentration of scavenged
vapor is negligible. Following from these two assumptions and using
the average concentration of 1260 in the air samples (3%) and that in
the rain samples (13%), the amount of PCBs on particulates is 3/13 or
27%. Whether or not the second assumption is correct is uncertain
since the Henry's Law constants for Aroclor mixtures may be in error
and PCBs could possibly be gas phase controlled (26). If PCBs are
1-10
-------
Table 1-7. PCB concentrations in continental air (23)
o
Location and date In air, ng/m Ref.
Kingston, R.I. , 1973-75
Organ Pipe National Park,
Ariz., 1974
Hays, Kans., 1974
Northwest Territories,
Canada, 1974
La Jolla, Cal., 1973
Vineyard Sound, Mass., 1973
University of R.I., 1973
Providence, R.I. , 1973
Chicago, 111., 1975-76
Jacksonville, Fla. , 1976
1 to 15
<0.02 to 0.41
<0.03
<0.002 to <0.07
0.5 to 14
4 to 5
2.1 to 5.8
9.4
3.6 to 11.0
3 to 36
23
23
23
23
24
12
20
20
2
13
Table 1-8. PCB concentrations in continental air near sources
o
Location and date In air, ng/m Reft
Washington County, N.Y., 1978 20 to 4000(300)* 22
Washington County, N.Y.
dredge spoil sites, 1978 107 to 3240 22
Jacksonville, Fla., 1977 , 498 13
Concentration after PCB use was stopped.
I-11
-------
gas phase controlled, rain may be able to scavenge PCS vapor and an
estimate of this type could not be made.
Junge (27) estimated the amount of vapor/particulate partitioning
of chlorinated hydrocarbons (CHC) over the ocean using a different
approach. He applied adsorption theory along with the number of
particles per cubic centimeter, the surface area represented by the
particles and the saturation vapor pressures of the chlorinated
hydrocarbons. Ten percent of the CHC over the ocean was predicted
to be in the particulate phase. Using Junge's approach less than 5%
of the PCBs over the ocean would be associated with particulates.
The situation in urban and rural air is more complex due to the
increased number of particles. For Aroclor 1254 and 1260 approximately
10 to 45% would be associated with particulates while less than 5% of
the 1242 would be found in the particulate phase. This can only be
considered an estimate since adsorption not only depends upon the
amount of surface area and the vapor pressure of the adsorbate but
also the nature and subdivision of the adsorbent.
Table 1-9. PCB fallout data
2
Location and date PCB, ng/m -month Ref.
Sweden, 1972 620 to 10,510 17
Iceland, 1974 40 to 1,050 18
La Jolla, Cal., 1977 9,000 24
Southern Cal., Bight, 1974 1,560 to 34,890 19
Bratislava, Czech., 1972 to 1973 120 to 3,440 28
Tsavo East, Kenya, 1973 20 28
Nakuru, Kenya, 1973 15 to 25 28
Gambia, 1973 184 28
1-12
-------
1-4. SAMPLE CLEANUP AND QUANTIFICATION
XAD-2 Cleanup
Considerable difficulty was experienced in obtaining a low
blank extract from XAD-2 resin. Air sampling required use of the resin
on a dry basis; upon drying however, contaminants are released from
inside the beads (29). These contaminants are formed during the
manufacturing process. It was necessary to remove as many of the
contaminants as possible so they would not affect the collection
efficiency and complicate interpretation of chromatograms.
Several cleanup procedures were attempted in order to reduce the
contamination level. A 24-hr soxhlet extraction with hexane as well
as three successive 24-hr extractions, failed to produce a low blank.
An extensive extraction scheme using warm water, methanol, acetone,
ethyl acetate, dichloromethane and petroleum ether was tried. A
24-hr soxhlet extraction with each of these solvents of decreasing
polarity failed to reduce the contaminant level. While the inter-
fering organic constituents could not be identified as PCBs, they
still obscured the early portion of the chromatograms where many of
the more volatile PCB components elute.
A 24-hr soxhlet extraction with petroleum ether gave better results
than either of the two previous methods. It was possible that the
contaminants had a vapor pressure similar to that of the other sol-
vents and instead of the compounds being removed during the extraction
they were continually condensing on the XAD-2. A 72-hr extraction
with petroleum ether, changing the solvent every 24 hr, was found to
be satisfactory. Eluting the XAD-2 extract through an alumina column
(6% H20 w:w) showed no PCB interfering contaminants present (Fig. 1-1).
The XAD-2 cleanup procedure can be found in Appendix I-A.
Some difficulty was also experienced with reuse of the resin.
Samples taken in the Milwaukee area over a 3 day period turned the
usually white XAD-2 resin yellow. A 24-hr soxhlet extraction with
petroleum ether failed to produce a clean blank. The resin was found
to be permanently contaminated and could no longer be used. These
problems were not experienced when collecting air samples over Lake
Michigan. During the duration of the Lake Michigan sampling program
the resin was continually reused for both sampling and lab experiments.
Discoloration of the XAD-2 or unsatisfactory blanks were not observed.
The discoloration of the resin may be due to the organic pollution
load or the type of volatile organics present in the sampling area.
A more extensive cleanup procedure may be necessary. This was not
investigated.
In applications (uf XAD-2 for air sampling, the best cleanup
method is soxhlet extraction with a low boiling solvent such as
petroleum ether. Clean blanks are obtained which show no interfering
components when eluted through an alumina column. Also reuse of
XAD-2 may not be possible in heavily polluted areas unless a more
extensive cleanup procedure is employed.
I- 13
-------
*AROCLOR 1242 PEAKS
BEFORE ALUMINA CLEANUP
AFTER ALUMINA CLEANUP
Fig. 1-1. Aroclor 1242 in XAD-2 extract before and after alumina cleanup,
1-14
-------
Sample Cleanup
In the process of deciding upon a quantitative, as well as qualita-
tive, cleanup procedure, many methods were tried with limited success.
Air samples were collected in Madison, Wisconsin and submitted to
various techniques. The method of Ketseridis et al. (30) was modified
somewhat by coupling it with column chromatography. The sample was
extracted with concentrated sulfuric acid and KOH in methanol. It
was then submitted to either florisil, alumina, or silicic acid
chromatography. A very clean sample was obtained but the recovery
was low. Using TCB, the acidic and basic extractions coupled with
alumina chromatography produced an analytical recovery of 68%.
Junk et al. (29) used an acidic (1 N HCl) and basic (0.05 N NaOH)
extraction to elute organic acids and bases from XAD-2 resin. The
procedure was modified somewhat in this study. The resin was eluted
with acid and base and then soxhlet extracted with a mixture of
petroleum ether and acetone. The extract was then eluted through an
alumina column. The PCBs were not clearly visible.
Sample extracts submitted to column chromatography using alumina
(6% H20 w:w), florisil (2% H20 w:w) and silicic acid (3.3% H20 w:w)
gave similar results. Chromatograms of extracts seemed extremely
complex, even more so than those obtained by researchers using similar
cleanup techniques. From these results it was conjectured that
XAD-2 was a more efficient collection medium for organic vapors than
other media such as polyurethane.
It was decided that alumina chromatography would be used while
trying to optimize GC conditions to resolve as many PCB components as
possible. A listing of the analytical methodology of air sample
analysis can be found in Appendix I-A. Analytical recoveries for
direct spikes on the resin, with subsequent elution of the concentrated
extract through alumina were: 85.5, 80.1 and 64.9% for TCB, Aroclor
1242 and Aroclor 1221 respectively.
The best cleanup procedure for air samples collected in this
study was found to be alumina column chromatography. Satisfactory
recoveries were obtained using both labelled and unlabelled PCBs.
Recoveries were found to depend upon the volatility of the PCB mixture
with Aroclor 1221 being the most difficult to recover.
Sample Quantification
The difficulties experienced with the quantification of air samples
are the result of two factors: interfering volatile organics present
in ambient air and nonconformance of samples to standards. Many of
the volatile organics elute very early when chromatographed and totally
obscure the components of an Aroclor 1221 mixture. Cleanup procedures,
such as alumina, remove some of the interfering organics.
Air samples very rarely contain PCB components in the same
proportion as those found in the original Aroclor mixture. Also,
1-15
-------
mixtures of Aroclors sometimes occur in air, as was the case in this
study (Figs. 1-2 to 1-5). It became apparent that the samples most
closely resembled a mixture of Aroclors 1242 and 1254 rather than
Aroclor 1248. The standard error can be used as a measure of the
sample's conformance to the standard. The errors were rather sizeable
in most cases (Table 1-10). One reason for the large error is
enrichment of the lighter components of the original mixture in the
vapor phase.
The samples were quantified by electron capture gas chromatography
on a Varian Aeorograph Series 1700 gas chromatograph equipped with
two Sc (3n) foils. A 3.35 m column (2 mm ID) packed with 1.5%
OV-17:1.95% OF-1 on 80/100 mesh Gas Chrom W(AW) and a 3.05 m column
(2 mm ID) packed with 4% SE-30:6% OV-210 on 80/100 mesh Gas Chrom W(AW)
were used for identification purposes. Since the cleanup procedure
did not remove all PCB interfering substances, it was necessary to
use 2 different liquid phases for positive identification of PCB
components. Coinjection of samples with Aroclor standards was also
used for identification purposes.
The samples could not be temperature programmed due to severe
baseline drift. They were alternatively chromatographed at three
different temperatures to obtain the best resolution and sensitivity
possible. The isothermal temperatures were 175, 205 and 230°C
for Aroclors 1242, 1254 and 1260 respectively. Samples and standards
were always in the linear range of the detector so peak heights were
used for quantification. Three standards were run each day and the
precision of the injections was always 5% or better. Duplicate
injections of the sample always produced a precision within 5 to 10%.
The samples were quantified by dividing the peak heights of the
sample peak by the sensitivity (deflection per amount injected) of
the matching peak in the Aroclor standard. The amount of PCB repre-
sented by the peaks in the sample were then averaged to obtain the
amount of Aroclor in the sample. The standard error can be obtained
by finding the deviation of the peaks from the average.
To determine the Aroclor composition of the sample, chromatograms
were divided according to the method of Webb and McCall (26), which is
presently one of the best GC quantification methods. The total amount
of PCBs is found by determining the actual amount of PCB under each
peak. This is possible if the weight percentage of each component in
an Aroclor mixture is known. To apply the weighting factors they
obtained, the resolution of the gas chromatographic columns must be
identical to theirs. The resolution of the gas chromatographic
columns in this study was much better. Chromatographing the air
samples under their conditions, in an attempt to reproduce their
resolution, would have totally obscured peaks in the sample matching
Aroclor 1242. This is a result of the many early eluting volatile
organics. Since more than one Aroclor standard was used to quantitate
a sample, duplication of results was avoided by using nonoverlapping
peaks in each of the PCB mixtures. Five peaks from Aroclor 1242,
seven from Aroclor 1254 and five from Aroclor 1260 were used to
quantitate the samples.
There are other alternative methods which are better than the one
1-16
-------
Column: 4% SE-30/
6% OV-210
175" C
LAKE MICHIGAN AIR SAMPLE
AROCLOR 1242
Peaks matching standard
Column: 1.5% OV-17/1.95% QF-1 175°C
Fig. 1-2. Lake Michigan air sample and Aroclor 1242 (175 C).
1-17
-------
LAKE MICHIGAN AIR SAMPLE
AROCLOR 125t
Peaks matching standard
Column: 4% SE-30/6% OV-210 205*C
Fig. 1-3. Lake Michigan air sample and
Aroclor 1254 (205°C).
1-18
-------
AROCLOR 1248
LAKE MICHIGAN AIR SAMPLE
* Peaks matching standard
Column: 1.5% OV-17/1.95% QF-1 185*C
Fig. 1-4. Lake Michigan air sample and Aroclor 1248 (185 C)
1-19
-------
LAKE MICHIGAN AIR SAMPLE
Peaks matching standard
Column: M%SE-30/6% OV-210 20S"C
Fig. 1-5. Lake Michigan air sample and Aroclor 1248
(205°C).
1-20
-------
used in this study. The Webb and McCall (26) method is excellent
for samples which do not exactly conform to the original Aroclor
mixture. The only drawback is that weighting factors must be determined
for the resolution of the instrument on which the analysis is being
performed. To obtain an accurate number all the components that are
resolved in the standard must also be resolved in the sample. Gas
chromatography-mass spectrometry is an excellent tool to use for
identification. The instrument is practically unisensitive to a
family of compounds such as PCBs, so a much more concentrated sample
than the ones collected in this study is necessary. Perchlorination
of the PCBs to decachlorobiphenyl (31) would be an excellent way to
quantitate samples but this method has not been perfected enough to
give consistent and quantitative recoveries (32). Perchlorination
also would destroy any qualitative information about the sample but
GC analysis of the raw sample could still be used to obtain this
information. In the future, PCB analysis must move towards quantifica-
tion of separate PCB components to gain a better understanding of the
partitioning of PCBs between air, land and water.
Table 1-10. Standard errors of Lake Michigan and Milwaukee samples
Standard error for Aroclors**
Standard error for Aroclors**
Sample*
1242
1254
Sample* 1242
1254
1260
XLM2
XLM4
XLM5
XLM6
XLM7
XLM8A
XLM9A
XMIL1
21.4
26.4
19.5
26.2
17.2
18.1
20.4
15.2
10.1
20.3
36.5
27.1
24.6
6.4
34.9
20.2
LM2
LM3
LM4
LM5
LM6
LM8A
LM9A
MIL1
MIL 3
MIL5
MIL7
MIL8
MIL9
MIL10
3.7
13.1
10.6
14.0
19.1
13.5
13.1
5.5
11.4
10.2
10.6
6.9
13.4
9.5
20.7
20.3
37.0
20.0
14.7
17.3
10.9
21.6
15.9
17.2
22.4
19.1
13.7
22.0
18.7
9.8
17.0
25.9
17.0
26.0
19.8
43.8
35.5
*Samples with X are XAD-2 extracts, others are filter extracts.
**Standard error is (Amount under each peak - average amount)
Average amount
1-21
-------
1-5. SAMPLING
Air was sampled over Lake Michigan and in Milwaukee, Wisconsin.
The samples collected over Lake Michigan were taken aboard the EPA
research vessel, Roger R. Simons, during the spring, summer and fall
of 1977. Air sampling was performed in conjunction with the water
sampling program being conducted by the EPA and air monitoring by
Dr. Herman Sievering.
During the EPA cruises, transects of the southern lake basin
were made (Fig. 1-6 and Table 1-11). The ship's running time averaged
approximately 8 hr a day. In order to collect a detectable quantity
of PCBs, a single sampling period usually extended over a period of
3 days. Samples were collected over large areas (Table 1-12) and
under variable weather conditions.
The samplers were located on the focsle deck of the ship, as far
away from the stack as possible, to eliminate contamination from ship
fuel. Contamination was easily avoided when the ship was underway
but problems occurred when the wind was blowing over the stern of the
ship. The samplers were turned off under these conditions. Under
heavy sea conditions the focsle deck takes on a large quantity of sea
spray. The samplers were turned off and covered under these conditions.
Precipitation events and fog also required that the samplers be covered
in order to eliminate contamination from wet fallout. To avoid
sample collection of on shore point sources, the samplers were operated
only when the ship was at least 3.2 km off shore.
Meteorological data were taken at hourly intervals during the
sampling periods. The data included relative humidity, air temperature,
wind speed and wind direction.
Samples were also taken during the winter of 1977 to 1978 in the
Milwaukee area within the Menomonee River Watershed. A stationary
sampling network had been previously established for the IJC Menomonee
Watershed study (33). Two of the seven locations were chosen for PCB
sampling: the Falk Corporation and River Lane (Fig. 1-7). The Falk
Corporation is located in the industrial section of the city while the
River Lane station is located in a rural area. These locations were
chosen so a comparison could be made between the environmentally
different areas. Daily access to the samplers was not possible, which
resulted in continuous operation of the samplers for extended lengths
of time. Weather data were not available for these sampling locations.
1-22
-------
SHEBOYGAN
MILWAUKEE
- CRUISE I
CRUISE II
— CRUISE III
•• CRUISE IV
KENOSHA flL-•_•._._ ^'*<.•. /
V. ~~ "'-^""^Vii
CHICAGO
GRAND
HAVEN
HOLLAND
SOUTH
HAVEN
MICHIGAN
CITY
0 10 20 40
l__j > «
Scale
in
Miles
Fig. 1-6. Lake Michigan cruise
tracks.
1-23
-------
Table 1-11. Lake Michigan cruise tracks
Cruise
Date
Decription
II
III
IV
11 to 17 June, 1977
14 to 19 Aug., 1977
20 to 26 Aug., 1977
17 to 24 Aug., 1977
Transects from Chicago to Michigan City,
Michigan City to Chicago, Chicago to
Holland, Holland to Milwaukee, Milwaukee to
Grand Haven, Grand Haven to Sheboygan,
Sheboygan to Milwaukee
Stationary sampling northwest to Michigan
City
Transects from Chicago to Michigan City,
Michigan City to Kenosha, Kenosha to South
Haven, South Haven to Milwaukee, Milwaukee
to Grand Haven, Grand Haven to Sheboygan,
Sheboygan to Milwaukee
Sheboygan to Grand Haven, Grand Haven to
Holland to Grand Haven, Grand Haven to
Milwaukee, Milwaukee to South Haven, South
Haven to Kenosha, Kenosha to Michigan City,
Michigan City to Chicago
Table 1-12. Lake Michigan sample locations
Sample
Date
Location
2 15 to 17 June, 1977
3 14 to 17 Aug., 1977
4 18 to 19 Aug., 1977
5 20 to 21 Aug., 1977
6 22 to 24 Aug., 1977
25 Aug., 1977
8A 17 to 21 Sept., 1977
9A 22 to 24 Sept., 1977
Milwaukee to Grand Haven, Grand Haven to
Sheboygan, Sheboygan to Milwaukee
Stationary sampling northwest to Michigan
City
Stationary sampling northwest to Michigan
City
Chicago to Michigan City, Michigan City ot
Kenosha
Kenosha to South Haven, South Haven to
Milwaukee, Milwaukee to Grand Haven
Grand Haven to Sheboygan
Sheboygan to Grand Haven, Grand Haven to
Holland to Grand Haven, Grand Haven to
Milwaukee, Milwaukee to South Haven
South Haven to Kenosha, Kenosha to Michigan
City, Michigan City to Chicago
1-24
-------
LITTLE
MENOMONEE
RIVER
Scale
in
Miles
Fig. 1-7. Milwaukee sampling locations,
1-25
-------
1-6. SAMPLE RESULTS AND DISCUSSION
Comparison and Interpretation of Air Concentrations and Composition
A total of 15 air samples were collected for PCB analysis, eight
from Lake Michigan and seven from the Milwaukee area. To simplify the
discussion of results, PCB passing through a glass fiber filter is
assumed to be vapor. What is operationally defined as vapor may
actually be particles too small to be collected by a glass fiber filter.
PCB vapor over Lake Michigan was composed mainly of Aroclor 1242
(Table 1-13). The XAD-2 extracts had an average composition of 75% of
Aroclor 1242 and 25% of Aroclor 1254. The average vapor concentration
was 0.87 ng/m^ (0.44 to 1.33 ng/rn-^) . The particulate phase composition
was similar to that of the vapor in almost all cases (Table 1-14). Four
of the seven filter samples contained a small percentage of Aroclor 1260.
The average composition of the vapor phase was 86% of Aroclor 1242, 13%
of 1254 and 1% of 1260; whereas the particulate phase was 33% of Aroclor
1242, 39% of 1254 and 30% of 1260. The average composition of the
particulate PCBs was 69% of Aroclor 1242, 23% of 1254 and 8% of 1260.
Table 1-13. PCB content of Lake Michigan XAD-2 extracts
Sample
XLM2
XLM4
XLM5
XLM6
XLM7
XLM8A
XLM9A
Average
Amount of
air, m3
813
900
929
1102
422
897
716
PCBs,
ng
905
597
550
487
303
1192
891
Percent
1242
69.1
83.6
80.0
80.1
815
66.4
65.2
75.1
1254
30.9
16.4
20.0
19.9
18.5
33.6
34.8
24.9
PCBs,
ng/m3
1.11
0.66
0.59
0.44
0.72
1.33
1.24
0.87
I- 26
-------
Table 1-14. PCB content of Lake Michigan filter extracts
Sample
LM2
LM3
LM4
LM5
LM6
LM8A
LM9A
Average
Amount of
air, m3
813
2428
900
929
1102
897
716
PCBs,
ng
98
231
119
172
114
109
106
1242
74.5
72.7
79.0
59.3
70.2
58.7
67.9
68.9
Percent
1254
25.5
27.3
21.0
22.6
18.4
24.8
23.6
23.3
1260
0
0
0
18.1
11.4
16.5
8.5
7.8
PCBs,
ng/ra3
0.12
0.10
0.13
0.19
0.10
0.12
0.15
0.13
Particulate
PCB, ppm
1.7
3.7
7.5
2.4
5.6
4.2
2.8
4.0
Table 1-15. Lake Michigan samples, PCBs in total sample
Sample
2
4
5
6
8A
9A
Average
Amount of
air, m3
813
900
929
1102
897
716
PCBs,
ng
1003
716
722
601
1301
997
1242
69.6
82.8
75.1
78.2
65.8
65.5
72.8
Percent
1254
30.4
17.2
20.6
19.6
32.8
33.6
25.7
1260
0
0
4.3
2.2
1.4
0.9
1.5
PCBs,
ng/m3
1.23
0.80
0.78
0.55
1.45
1.39
1.03
TSP,*
pg/m3
69.9
17.3
17.9
31.0
28.4
54.5
36.5
*Total suspended particulates.
1-27
-------
The results from the Lake Michigan samples are similar in
concentration to samples taken by Bidelman and Olney (20) in the
North Atlantic in 1973. These researchers found a concentration range
of 0.21 to 1.6 ng/m^ in marine areas. Additional results from
their study (Table 1-5) showed that from 1973 to 1975, samples
matched Aroclors 1248 and 1242 but the composition changed to Aroclor
1254 in 1976. The•Lake Michigan samples were different in that they
were composed of a mixture of Aroclors 1242 arid 1254 rather than single
Aroclors. Bidelman and Olney (20) found that 2% of the total PCBs
was associated with particulate matter. A larger percentage (13%)
of the PCBs over Lake Michigan are associated with particulate
matter. The Lake Michigan samples are also similar in concentration,
but not composition, to those taken by Harvey and Steinhauer (12).
PCB concentrations in the North Atlantic, 500 km off the New York
Metropolitan area, ranged from 0.58 to 1.60 ng Aroclor 1254/m-^ and
<1% of the PCBs were associated with particulate matter.
The concentrations of the Lake Michigan samples are lower than
those reported for urban areas within the continental U.S. Data for
urban studies were discussed in Section 1-3 (Table 1-7). Concentrations
in urban areas are usually greater than 1.0 ng/nH. An average
concentration for urban Chicago air, reported by Murphy and Rzeszutko (2),
was 7.7 ng/m^ (Table 1-16). The average composition of the vapor
phase was 86% of Aroclor 1242, 13% of 1254 and 1% of 1260; whereas
the particulate phase was 33% of Aroclor 1242, 39% of 1254 and 30% of
1260. The average concentration of PCBs in the particulate matter
was 3 ppm, similar in concentration to the Lake Michigan samples.
Four percent of the PCBs were associated with particulate matter which
was a lower percentage than the partitioning over Lake Michigan.
The Milwaukee filter extracts (Table 1-17) were quite different
from those of Lake Michigan. Unlike the Lake Michigan samples and
similar to urban Chicago air, the Milwaukee particulate matter com-
position was weighted more towards heavily chlorinated PCB mixtures.
At the Falk Corporation the average composition was 14% of Aroclor
1242, 59% of 1254 and 27% of 1260; whereas the River Lane samples
had an average composition of 56% of 1242 and 44% of 1254. The average
PCB concentration in particulates was higher at the Falk Corporation
(7.0 ppm) than either Lake Michigan (4.0 ppm) or Chicago (3.0 ppm).
The River Lane station concentration (2.5 ppm) was similar to average
values over Lake Michigan and Chicago. The two sample pairs, namely
MIL7 and MILS and MIL9 and MIL10, were taken concurrently at the two
Watershed stations. The Falk Corporation showed higher concentrations
in both cases. The vapor concentration and composition reported for
one Milwaukee sample taken at Falk was 73% of Aroclor 1242 and 27%
of 1254. The composition was comparable to that of the Lake Michigan
samples. Of the PCBs, 16% were found to be associated with particulate
matter, a number comparable to the 13% obtained from the Lake Michigan
samples. The vapor phase concentration (2.25 ng/irr) was three times
higher than the average found over Lake Michigan but lower than that
reported for Chicago.
The wind roses and other meteorological data for the Lake Michigan
samples are given in Fig. 1-8 and Table 1-18. The samples, with the
exception of number 7, were usually collected over a 3 day period
during which time the meteorological conditions were subject to
1-28
-------
Table 1-16. Chicago air samples (2)
Date
Amount of
air, in3
PCBs,
ng
Percent
1242 1254 1260
PCBs,
ng/m3
27 Nov. 1975 1,788
5 Dec. 1975 1,106
20 Feb. 1976 1,072
29 Mar. 1976 1,138
19 Apr. 1976 954
Average
27 Nov. 1975 x 1,788
5 Dec. 1975 1,106
20 Feb. 1976 1,072
29 Mar. 1976 1,138
19 Apr. 1976 954
Average
TOTAL SAMPLE
12,017
8,288
3,912
11,337
10,528
20.7
6.8
12.0
24.5
5.3
13.9
POLYURETHANE FOAM EXTRACTS
11,469
7,517
3,646
11,275
10,385
80.3
93.2
86.4
75.3
94.7
86.0
19.1
6.2
10.2
24.4
4.7
12.9
0.6
8.5
6.3
0.2
0.7
3.3
0.5
0.6
3.4
0.2
0.5
1.0
FILTER EXTRACTS
6.7
7.4
3.6
9.9
11.0
7.7
6.41
6.80
3.40
9.91
10.88
7.5
27 Nov. 1975
5 Dec. 1975
20 Feb. 1976
29 Mar. 1976
19 Apr. 1976
Average
1,788
1,106
1,072
1,138
954
547
771
266
62
143
45.2
0.7
14.3
75.3
28.4
32.8
54.2
12.6
36.6
42.6
48.8
39.0
2.6
85.5
46.0
0.2
15.2
29.9
0.31
0.70
0.25
0.05
0.15
0.29
Table 1-18. Average meteorological data for Lake Michigan air samples
Sample
Temp., °C
Rel. humidity, %
Bar. pressure, cm Hg
2
3
4
5
6
7
8A
9A
17.0
24.2
21.2
18.6
15.7
15.7
14.2
14.0
74
61
39
66
63
62
—
—
76.50
76.48
76.61
76.27
76.53
76.71
1 76.33
76.33
1-29
-------
Table 1-17. PCBs in Milwaukee samples
H
1
00
o
Sample*
MIL1
MIL3
MILS
MIL7
MILS
MIL9
MIL10
Averages
XMILI**
Date
Nov. 30 to Dec.
Dec. 5 to Dec.
Dec. 12 to Dec.
Jan. 31 to Feb.
Jan. 31 to Feb.
Feb. 7 to Feb.
Feb. 7 to Feb.
Fa Ik Corp.
River Lane
2,
9,
16
3,
3,
10
10
1977
1977
, 1977
1978
1978
, 1978
, 1978
Amount of
air, m3
1821
3667
3632
2817
3262
2822
2662
1821
PCBs,
ng
650
1120
1184
1176
345
1271
193
4102
1242
9.1
11.6
15.1
26.7
69.3
7.5
42.0
14
56
73
Percent
1254
54.9
57.7
61.7
50.9
30.7
69.2
5.0
59
44
27
1260
36.0
30.7
23.2
22.4
0
22.3
0
27
PCBs,
ng/m3
0.36
0.31
0.33
0.42
0.11
0.44
0.07
0.37
0.09
2.25
TSP
ug/m3
51.0
50.1
54.0
63.9
37.6
49.3
35.3
53.7 '
36.5
Particulate
PCBs , ppm
7.1
6.2
6.1
6.6
2.9
8.9
2.0
7.0
2.5
*Even numbered samples taken at River Lane, odd numbered samples taken at Falk Corp,
**XAD-2 extract corresponding to MILL
-------
No. 3
No.
No. 2
270
No. 5
180
No. 6
No. 8
No. 7
'«
No. 9
Numbers denote average wind speed in knots
Fig. 1-8. Wind roses for Lake Michigan sampling stations.
1-31
-------
considerable change. Due to these factors, it was not possible to
correlate PCB concentrations with meteorological conditions. Since
PCBs are more concentrated in urban air, it would be expected that
southwest winds would bring elevated concentrations of PCBs over
the lake from the Chicago and Milwaukee areas or metropolitan-
industrial areas in general. The complex wind roses do not permit
interpretation of the data to indicate this. PCB concentrations were
highest on the June and September cruises but a seasonal trend can
presently not be discerned due to the limited amount of data.
Meteorological data were not available for the Milwaukee samples.
One result that was clearly evident from the preceding discussion
was that urban particulates contained a higher percentage of the
heavily chlorinated mixtures (Aroclors 1254, 1260) while the particulates
over Lake Michigan contained higher percentages of the more volatile
mixtures (Aroclor 1242). Adsorption theory, as applied by Junge (28),
would predict that the less volatile PCB constituents would be
associated with particulate matter. The Lake Michigan samples showed
just the opposite trend indicating that other factors may be
influencing the PCB composition of air particulates over the lake.
One possible approach to interpret the results in this study
might be to consider the extent of particulate/vapor partitioning
and composition as being a function of five factors:
1. The source of PCB emission
2. The amount of total suspended particulates (TSP)
3. The type of particulate matter present
4. The atmospheric residence time
and possibly
5. Photolysis.
The number of conclusions drawn is limited by the scarcity of data.
The following interpretation is largely speculation but it can be used
to design better sampling schemes for future investigations of aerial
PCB transport.
The source of PCB emission will influence the composition of
airborne PCBs in close proximity to that source. If the source is
a major industrial user of a particular Aroclor, the same mixture
would be reflected in the atmosphere with possible enrichment of the
more volatile components. Sources such as poorly operated incinerators
or sanitary landfill gas vents might emit mixtures of Aroclors in
variable proportions due to the many former uses. The composition
of Aroclors emitted from landfills would depend upon the volatility
of the mixtures. Murphy and Rzeszutko (2) found that PCB in the vent
gases of a sanitary landfill consisted of 98% of Aroclor 1242, the
most volatile mixture of the more widely used Aroclors. The temper-
atures in a poorly operated incinerator, while not high enough to
destroy PCBs (200Q+°C), would be high enough to vaporize all the
Aroclor mixtures. The amounts and composition present in the atmosphere
1-32
-------
would thus depend upon the amounts and composition of the incinerated
PCB containing products.
The amount of total suspended particulates present in the
atmosphere may also determine the extent of vapor/particulate partition-
ing. Thomas et al. (34) found that the vapor concentration of polyaromatic
hydrocarbons (PAHs), at a generation source having temperatures
ranging from 80 to 90°C, equalled the saturation vapor density of
the pure state and exceeded that of the adsorbed state. Even in the
presence of excess soot the vapor phase PAHs were in excess. Attainment
of adsorption equilibrium was slower than condensation (homogeneous
nucleation) of the vapor. Since PCBs are more volatile than PAHs it
seems logical that vapor phase PCBs close to a source, such as an
incinerator stack, will be in greater concentrations than the adsorbed
species. Distant from the source of generation, the PCBs will become
diluted, the total amount of particulates will become greater than the
amount of volatile PCBs, and adsorption will become a factor. If
this theory applies to PCB adsorption, higher TSP loads will result
in higher concentrations of PCBs in the particulate phase.
This correlation was not evident in the sets of samples taken
on Lake Michigan or at the Falk sampling site. 'The Falk samples
did show high particulate PCB and TSP concentrations while the River
Lane samples, which were taken concurrently, showed an opposite trend.
Table 1-19' contains a comparison of the average particulate PCB
concentration, PCB vapor concentration, and TSP for four locations.
A coincident correlation exists between particulate PCB concentration
and the TSP, with high TSP values producing high particulate PCB
concentrations. A correlation did not exist between the percentage
of the total PCBs in the particulate phase and TSP. Chicago, which
had a TSP concentration intermediate between that of Lake Michigan
and the Falk Corporation, had the lowest percentage (2). Over the
ocean however, where TSP values would probably be lower than those
reported for Lake Michigan and urban areas, particulate PCBs comprised
only 2% of the total amount (11).
The type of particulate matter present in the atmosphere may
also influence the extent of adsorption. Murphy and Rzeszutko (2)
reported that the concentration of PCBs in non-polar organic extractable
solids (18 ppm) was greater than that in the total particulate matter
(3 ppm). This may indicate that PCBs have a tendency to associate
predominantly with aerosols containing non-polar organic matter. If
the organic composition of aerosols has a spatial dependence the PCB
composition may also show the same trend. Strand (35) showed that
TOC in aerosols varies considerably. If PCBs are closely associated
with TOC in aerosols a better correlation might exist between PCBs
and TOC than PCBs and TSP.
The residence time of PCBs and particulates may also influence
the vapor/particulate partitioning as well as the composition. If
it is assumed that the particulate matter from Milwaukee and Chicago
is moving out over the lake, the particles will have aged somewhat
when they reach the lake. In urban areas, where PCB concentrations
are usually high, the more highly chlorinated mixtures may associate
with particulates more readily due to their low volatility, while
the lower chlorinated mixtures remain predominantly in the vapor
phase due to slow adsorption kinetics. As the particulates move out
1-33
-------
Table 19. Average TSP and particulate PCB and vapor phase PCB concentrations
Location
TSP,
yg/m3
Particulate
PCBs, ppm
Vapor phase
PCB, mg/m3
Falk Corp., Milwaukee
Chicago
Lake Michigan
River Lane, Milwaukee
53.7
48
40.6
36.5
7.0
3
4.0
2.5
2.25
7.5
0.87
*(Ref. 2).
1-34
-------
over the lake, the more volatile mixtures associate with particulate
matter making vapor and particulate more homogeneous in composition.
The size of the particulate associated PCBs, which will determine
their residence time, may also determine their composition. Sampling
air over the North Atlantic from 1973 to 1976, Bidelman (21) has found
a change in PCB composition from Aroclor 1248 to 1254. The sale of
PCBs in the U.S. was discontinued after 1974 so the change in composition
cannot be due to a new source. Aroclor 1248 may be associated with
particles which are depositing more rapidly.
The effect of photolysis on the PCB composition of air has not
been investigated. The main reaction of the photolysis of PCBs in
water or solvents is stepwise dechlorination (36,37,38). Stepwise
dechlorination of PCBs in air would result in a predominance of the
lower chlorinated mixtures. The process must be investigated more
thoroughly before any conclusions can be made.
The vapor/particulate partitioning and composition of airborne
PCBs is probably a complex combination of the five factors presented.
This makes interpretation of results extremely difficult. In the
discussion of results, seasonal trends were not investigated. The
Lake Michigan samples were taken during the spring, summer and fall
while those reported for Milwaukee and Chicago were taken predominantly
during the winter. A correlation between particulate composition
and season is not presently discernable. The limited number of samples
used in the interpretation and the absence of meteorological
correlations also present difficulties. The collection and quanti-
fication method may also influence the particulate/vapor partitioning
and the reported composition of an air sample. Nevertheless, the
above speculations can be used as a tool in designing future sampling
schemes and laboratory experiments.
Vaporization
The degree of alteration of the vapor/particulate partitioning
of airborne PCBs by a high volume air sampler is unknown. Vaporization
of PCBs from collected particulate matter has been hypothesized (2,12,20)
The vaporization of certain PAHs from collected particulates under
ambient conditions has already been reported (39). Vaporization was
greatest for the more volatile PAHs, such as fluoranthene and pyrene
but losses were also observed with benzo(a)pyrene. PAHs are much
less volatile than PCBs. This can be exemplified by comparing their
calculated saturation vapor densities (Table 1-20).
In the sampling system employed in this study, volatilized PCBs
would be collected as vapor. This process would lead to an under-
estimation of the particulate PCB concentration. The degree of
particulate/vapor partitioning is very important when estimating wet
and dry fluxes.
Another problem associated with high volume air sampling is
the performance characteristics of the filter media. The particulate/
vapor partitioning of PCBs is operationally defined by the sampling
system. The possibility exists that all airborne PCBs are associated
1-35
-------
Table 1-20. Saturation vapor densities of PAHS (ng/m3) and
PCBs (mg/m3) (41)
Compound Vapor density, ng/m3
PAHs
Pyrene 74,000
Benz-(a)-anthracene 1,050
Benz-(a)-pyrene 85
Benz-(e)-pyrene 85
Benz-(k)-fluorene 16
Benz-(ghi)-perylene 1.6
PCBs
Aroclor 1242 5.5
Aroclor 1248 7.9
Aroclor 1254 1.4
Aroclor 1260 0.73
1-36
-------
with particulate matter, but some of the particles are too small to
be captured by a glass fiber filter. At the normal flow rate of
the PCB sampler (0.566 to 0.708 m3/min) the collection efficiency is
100% for particles greater than 0.3y in diameter (40). Glass fiber
filters may collect particles less than 0.3y in diameter but the
collection efficiency is unknown.
An attempt was made to determine the extent to which the sampler
employed in this study altered the vapor/particulate partitioning
of PCBs.
Experimental
An experiment was designed in which PCB coated particulate matter
was dispersed on a glass fiber filter and subjected to normal sampling
conditions. Fly ash and soil were used instead of atmospheric
particulate matter. Both substances were dried before use. The soil
sample was pulverized using a mortar and pestle. The fly ash and
soil were then sieved in order to remove particles greater than 44y
(sand fraction).
The Lake Michigan air samples were used as the basis for
estimating the amount of PCB that was coated on the particulate matter.
A mixture of 75% Aroclor 1242 and 25% Aroclor 1254 was used. Assuming
that all the airborne PCBs were originally associated with particulate
matter, the average particulate PCB concentration would be 20 ppm.
This number is half an order of magnitude less (4 ppm) if it is
assumed that all the particulate PCBs are captured by the glass
fiber filter.
The PCB coated particulate matter was prepared by immersing soil
and fly ash in petroleum ether containing Aroclors 1242 and 1254
and allowing the slurry to equilibrate for a period of two days. The
petroleum ether was evaporated slowly using a gentle vacuum and taken
to dryness with a slow stream of nitrogen. The dried substances
were transferred to bottles and put on a wrist action shaker for
30 min to insure homogeneity of subsamples. A portion of the coated
particulate matter was submitted for analysis to determine its PCB
content. The composition was never the same as the sought composition
due to adsorption of PCB on container walls and losses during evap-
oration. The total organic carbon content of the fly ash and soil
was also determined.
Two different experiments were performed. The first consisted
of using prefiltered laboratory air. Two samplers were interconnected,
one being used as a prefilter. A glass fiber filter was cut to fit
on the top of one of the XAD-2 capsules. Coated fly ash was dispersed
on the filter as evenly as possible. The sampler was operated for
48 hr during which time 1152 m^ of air flowed over the particulate
matter at 0.4 nrVmin. The total amount of air was approximately
the same as the amount sampled during a normal collection period.
Analysis of the filter and XAD-2 showed that all the PCB was
vaporized from the particulate matter. There were two unforeseen
1-37
-------
problems associated with this type of experiment. A partial vacuum
greater than the pressure drop experienced during a normal sampling
situation, may have been created between the two XAD-2 capsules
which increased the tendency for vaporization. The use of a prefilter
insured that organic free air was continuously being drawn over the
particles. The PCBs may have vaporized to attain equilibrium with
the flowing air, which was completely devoid of vapor phase PCBs.
A second set of experiments was designed to avoid these problems.
Two samplers were run concurrently and outdoors to model a normal
sampling situation as closely as possible. One sampler was used as
a blank. Using this experimental design insured that the particulates
would be in quasi-equilibrium with the air being drawn over them.
Also, the pressure drop across the filter would be the same as it is
when a normal sample is taken (5 mm Hg).
Fifty to one hundred milligrams of coated particulate matter was
dispersed on the filter surface as evenly as possible. The sampler
was run for 12 hr during which time 504 tip of air flowed over the
particles at 0.7 nrVmin. Difficulties with interpretation of results
were avoided by sampling for a shortened length of time. This time
reduction should have no serious effects on the results. During
longer sampling periods the collection of a large quantity of organic
vapors interferes with quantification of the most volatile PCB isomers.
Since these would be the most important components in an experiment of
this type, a 12 hr sampling period was chosen.
Results and discussion
The results show that more PCB was vaporized from fly ash than
soil (Table 1-21). The sampler used to collect the air blank collected
a larger amount of particulate matter in both experiments. Correcting
the loss of total PCB for the lack of reproducibility between the
samplers showed that 19.8 and 90.1% of the PCB was lost from the soil
and fly ash, respectively. It is also apparent that the more volatile
Aroclor 1242 is lost to a greater extent. The chromatograms of the
filter extracts showed that the heavily chlorinated components of
Aroclors 1242 and 1254 were vaporized to a lesser extent.
Desorption from the particulate matter could be controlled by
two factors. Kinetic effects determine the rate of vaporization. If
the desorption kinetics are rapid all the PCBs may be lost from the
particulate matter during a normal sampling period. The type of
particulate matter onto which PCBs are adsorbed may also be a factor.
The experiments may indicate that PCBs adsorbed to particles containing
a high organic carbon content are unavailable for vaporization. The
fly ash contained 4.9 x 10~^% organic carbon while the soil contained
57%.
There is one major difference between the particulate matter
used in this experiment and the particulate matter present in air.
Airborne particulates should be in equilibrium with the surrounding
air. It seems unlikely that equilibrated particulates would be
stripped of their PCB content by a flow of air in x-^hich they had
1-38
-------
Table 1-21. PCB vaporization
Sample
Soil
Fly Ash
PCBs,
ng
692
1310
Percent
1242 1254
BEFORE
64.3
66.0
Particulate
PCBs, ppm
Total
organic C, %
SAMPLING
35.7
34.0
10.4
18.1
57
4.9 x 10~4
Soil 510 56.5
Air blank
Particulate 142 51.4
Vapor 3542 85.7
Percent loss 26.3 35.3
Fly ash 123 39.0
Air blank
Particulate 61 0
Vapor 4008 86.5
Percent loss 90.6 94.4
AFTER SAMPLING
43.5 7.7
48.6 6.4
14.6
10.1
62.0 1.7
100 5.5
13.5
83.2
1-39
-------
resided for a long period of time. Vaporization could take place if
PCBs were weakly adsorbed onto inorganic particulates such as fly ash.
Aerosols collected over Lake Michigan have been found to contain
20 to 70% organic carbon (35). As was the case with the soil, PCBs may
not vaporize from particles containing organic carbon. The high
percentage of_ Aroclor 1242 reported on Lake Michigan particulate matter
may suggest that if vaporization does occur, the kinetics of desorption
are rather slow.
Caution must be taken when extrapolating these results to the
collection of an air sample. Soil and fly ash are not the same as
airborne particulate matter. Also, the PCS was coated on the particulate
matter by immersing it in an organic solvent, a process which may
have altered the particle matrix.
The experiments indicate that vaporization may be a function of
the organic carbon content of particulates. PCBs adsorbed onto
particulates containing a high organic carbon content may not vaporize.
PCBs weakly adsorbed to inorganic particulates will be vaporized
during the collection process. Since Lake Michigan aerosols contain
20 to 70% organic carbon, vaporization of PCBs from these aerosols
may not occur.
Volatilization
Volatilization of certain low solubility organic compounds from
water bodies to the atmosphere has been proposed as an important
environmental pathway (9). PCBs are sparingly soluble in water and
have been included in this category (9,42). Versar (5), prepared a
first order PCS mass balance model for Lake Michigan and estimated
a PCS evaporation rate of 1160 kg/yr. A combination of an estimate
for dry fallout of 2908 kg/yr (5) with a wet deposition rate of
4800 kg/yr (2) suggests that approximately 14% of the atmospheric
input may be evaporating from the lake's surface. These estimates
show that the revolatilization process is a potentially important
environmental pathway and thus worthy of further investigation.
Recently the revolatilization process has been questioned. The
phase (gas or liquid) which controls the transfer of PCBs across the
air/water interface is now uncertain (27). The Henry's Law constant
(H), which is a measure of the partitioning of a gas between air and
water, has been calculated for Aroclor mixtures using solubility and
vapor pressure data (42). In order to calculate H, the vapor pressure
and aqueous solubility must refer to the same state of matter.
Individual PCB isomers are solids at room temperature while Aroclors,
which are complex mixtures of PCB isomers, are viscous liquids. This
fact was not taken into consideration when the Henry's Law constants
were calculated. Since Aroclors are non-ideal mixtures, the individual
isomers, when present in air or water, will be partitioned individually
and therefore must be treated accordingly. The published values of
H may be in error by a factor of a thousand (27). PCBs may in fact
be gas phase controlled, resulting in a flux of vapor phase PCBs to
the water surface rather than from it.
1-40
-------
The Whitman (43) two layer model, as applied by Liss and Slater (44),
may be used as a first approximation to estimate the flux of gases
across the air/sea interface. The liquid (kj,) and gas (kg) phase mass
transfer coefficients for PCBs can be estimated using the mass transfer
coefficients for C02 and B/^O. The values for PCBs can be found in
Table 1-22. The overall liquid phase mass transfer coefficient (Kg^)
can be obtained experimentally (42). Kg-^ can be calculated using the
following equations
C = C0 exp(-KOL t/L) Eq. (1)
where
KgL is the overall liquid phase mass transfer coefficient
t is time
L is the depth of water
CQ is the original soluble concentration
C is the soluble concentration at time t
Examining the equation which is used to calculate an overall mass
transfer coefficient (44) :
1/KOL - l/kL + l/O&G/RT) Eq. (2)
where
RT is 2.45 x 10~2 atm-m3/mol at 25°C
it can be observed that with a knowledge of KQL an experimental H
can be obtained.
PCBs are rather insoluble in water and have a large affinity for
surfaces (4). PCBs in Lake Michigan may be associated with particulate
matter and therefore unavailable for revolatilization unless desorption
is rapid. Analysis of Lake Michigan water indicates that 67% of the
PCBs are present in the particulate phase (2). This number is
operationally defined since PCBs may be adsorbed to particles too small
to be collected by a glass fiber filter.
Hetling et al. (22) performed a revolatilizatibn experiment using
PCB saturated water solutions. They found almost total revolatilization
of Aroclors 1221 and 1016 within a matter of days. They also stated
that the introduction of organic particulate matter slowed the rate
of evaporation. Data was presented only for the results of the
saturated water solution experiments.
Experimental
Assuming that the partitioning of PCB between the particulate
and soluble phase was important, an experiment was designed to examine
the revolatilization rate under conditions similar to those found in
the environment. A water sample was taken from Lake Emrick, a lake
1-41
-------
Table 1-22. Liquid and gas phase mass transfer coefficients
for PCBs
Aroclor Mol. Wt. kr, m/hr kT . m/hr
vJ J_i
1221
1016
1242
1254
192
256
261
327
9.19
7.96
7.88
7.04
0.10
0.08
0.08
0.07
Table 1-23. Liquid phase mass transfer coefficients (m/hr) for
Aroclors 1221 and 1016 (22) and Aroclor 1242 and 1254
(this study)
Sample Time,* hr 10., (1242) Km (1254)
FLY ASH EXPERIMENT (This Study)
IRV1 112 2.10 x 10~4 7.88 x 10~5
IRV2 306 1.19 x 10~4 5.26 x 10~5
IRV3 493 8.96 x 10~5 3.56 x 10~5
IRV4 877 6.23 x 10~5 2.33 x 10~5
ACETONE SPIKE EXPERIMENT (This Study)
2RV1
2RV2
2RV3
2RV4
1
2
3
4
112
306
493
877
24
48
72
96
1.14 x 10 4
5.82 x 10";?
4.06 x 10 5
2.57 x 10~5
(REF. 22)
KQL (1016)
1.04 x 10"3.
8.40 x 10 ~4
7.31 x 10~4
5.59 x 10";?
3.86 x 10~5
2.60 x 10~5
1.62 x 10~5
KQL (1221)
9.37 x 10~4
9.01 x 10~4
9.15 x 10~4
8.04 x 10~4
*Time sample taken after start.
1-42
-------
in southern Wisconsin. Unfiltered lake water (1 L) was used in
the experiment. Compressed air was fed into a 4 L flask through a
glass frit at 50 ml/min. The air was purified by two XAD-2 columns.
The outlet consisted of two glass columns filled with XAD-2 resin to
capture any revolatilized PCB. The water was stirred with a magnetic
stirrer to keep the particulate matter in suspension.
Two concurrent experiments were performed. It was not practical
to use a concentration as low as the 13 ng/L found in Lake Michigan (45)
Any amount of revolatilized PCB would have been less than the detection
limit of an electron capture detector. If the experiment was done on
an extremely large scale this may have been possible. The experiment
was designed so the solubilities of Aroclors 1242 and 1254 which are
0.240 mg/L and 0.012 mg/L, respectively (9), were not exceeded. In
one experiment 10 mg of PCB coated fly ash (1050 ng of Aroclor 1242,
1530 ng of 1254) was added to the lake water. In the other flask,
a PCB spike (2087 ng of Aroclor 1242, 2339 ng of 1254) was introduced
in lOyl of acetone. Acetone is water soluble. It was used with the
assumption that PCB introduced into the lake water in this manner
would initially be in solution. Ten milligrams of fly ash were also
added to keep the suspended solids load in each flask identical.
Analysis of the XAD-2 vapor traps was performed over regular intervals.
The duration of the experiment was one month after which time the
water was filtered with a glass fiber filter and batch extracted with
hexane. The non-filterable material was soxhlet extracted with
petroleum ether.
Results and discussion
The results from the experiments performed in this study as well
as those of Hetling et al. (22) will be used to discuss the revolatil-
ization of PCBs. By transforming Eq. (1) into the following form,
ln(C/C0) = -(KOL)t/L Eq. (3)
it can be observed that KQL should be a constant. In the experiments
performed in this study the actual amount of PCBs in solution (C)
was found by subtracting the vapor concentration at" time t from the
total amount of PCB spiked into the lakewater (Cg); therefore, (C)
may include both adsorbed and soluble PCBs. This calculation was
also necessary to interpret the experiment of Hetling et al. (22)
since only the amount revolatilized was determined. The values of
KQL for each of the evaporation studies can be found in Table 1-23.
Plotting the values of KQL versus time, it became apparent that the
overall mass transfer coefficient changed with time (Figs. 9-11).
The change in the overall mass transfer coefficient may be the
result of two factors, adsorption of PCBs to particulate matter and
different revolatilization rates for different PCB components. In the
fly ash and acetone spike experiments, particulate matter was present
in the system. Adsorption of PCBs to particulate surfaces will retard
the evaporation process (22). The change in concentration of the
PCBs in solution is due to both evaporative loss and desorption from
particulate matter. The experimentally derived KQL is actually a
1-43
-------
100
90
80
70
60
50
1*0
30
20
10
3
W
EC
O
*
w
II I I I
AROCLOR -100.6
AROCLOR 1221
K_. (m/hr)
UJ-i
J I I
10""* 23U56789 10~3 2 3 U 5 6
OL
Fig. 1-9. Km against time for Aroclors 1016 and 1221.
1-44
-------
900 -
800 -
700 -
600 -
500 -
i»00 -
300
200
100 -
r-i AROCLOR 1254
AROCLOR 1242
I I I I I I
10~5 2
3456789 10~H 23456
Fig. 1-10. K against time for Aroclors 1242 and 1254 for acetone
UJj
spike experiment.
1-45
-------
900
800
700
600
500
400
300
200
100
QAROCLOR 1254
^AROCLOR 1242
j I
10~5 23456789 lO'4 2345
Fig. I-11. K against time for Aroclors 1242 and 1254
\}Lt
in fly ash experiment.
1-46
-------
combined revolatilization and desorption rate. The nonlinearity of
the KOL versus time curve was the result of applying Eq. (1) to a
system it could not describe. A desorption term would have to be
added to linearize the results.
The processes of adsorption, desorption and revolatilization are
related in a complex fashion. Figures 1-12 and 1-13 show that the
loss curves for the fly ash and acetone spike experiments are identical
in shape, although the loss of Aroclor 1242 in the fly ash experiment
is much greater. Desorption from the fly ash had to occur before
the PCBs were revolatilized. The process must have been rapid since
the results are similar to those of the acetone spike experiment.
The solubilized PCBs must have subsequently adsorbed to the particulate
matter originally present in the lakewater. The KQL versus time curves
(Figs. 1-10 and 1-11) for Aroclor 1242 in both experiments are not
identical which suggests that adsorption, desorption and revolatili-
zation are related in a complex fashion. The Aroclors 1242 and 1254
in the acetone spike experiment might have associated with the fly ash in
that system, however this is unlikely since the fly ash is inorganic in
nature. A more logical explanation might be that the PCBs were adsorbed
to the particulates originally present in the lakewater.
The change in total mass transfer coefficient may also be the
result of different revolatilization rates for the individual components
of the two Aroclor mixtures. The explanation can be found by examining
Henry's Law:
P. = K xj Eq. (4)
where
p is the partial pressure
K is the mass transfer coefficient
x is the mole fraction
j is any solute in the system (in this case the individual
PCB isomers)
Henry's Law states that all real solutions should approach the behavior
described by this equation provided the solution is sufficiently
dilute. If several solutes are present, which is the case described
by these experiments, the solution must be dilute in all solutes.
Each of the different solutes has a different value of K (46).
The deviations of Aroclor mixtures from the behavior described
by Henry's Law are readily apparent. In the experiments performed in
this study, the PCB concentrations, while not exceeding the published
solubilities, definitely exceeded those in Lake Michigan. Uncertainties
may arise when extrapolating Henry's Law data from high to low
concentrations. Different revolatilization rates for individual
PCB components were observed. The standard error can be used as a
measure of the deviation from a standard Aroclor mixture. The standard
errors for each of the vapor samples can be found in Table 1-24. The
deviations were rather sizable indicating that the components did in
fact have different revolatilization rates. Figure 1-9 is a graph
of the data from Hetling et al. (22) for the revolatilization of
1-47
-------
50
30
20
10
QAROCLOR 1254
1242
I 1 I I I i
1°° 200 300 400 500 600 700 800 900
Fig. 1-12. Loss of PCBs (%) with time for acetone spike experiment.
1-48
-------
O O
J J
D O
O O
& K
o
o
o
o
00
o
o
o
o
ID
O
o
to
o
o
.3-
O
o
CO
o
o
CM
O
O
C
a)
j-i
OJ
cn
rt
1
•H
•u
•H
IS
03
e
Pu
CO
CO
M
O1
-ct
I
o
o
O
cn
o
oo
o
to
o
LO
o
CO
o
CM
-------
Table 1-24. Standard errors for revolatized PCB vapor
Sample*
Time,** hr
Standard error, *** %
Aroclor 1242
Arochlor 1254
IRV1
2RV1
1RV2
2RV2
1RV3
2RV3
1RV4
2RV4
112
112
306
306
493
493
877
877
11.2
7.7
25.0
19.3
27.2
28.1
40.4
27.5
19.9
19.8
5.2
6.9
13.7
15.5
37.4
32.2
*If first number is 1 it is fly ash if 2 it is acetone spile experiments.
**Time sample taken.
***(Amount under each peak - Average amount)
Average amount
Table 1-25. Composition of Aroclor mixtures (9)
Chlorines per
biphenyl
Composition, %
1221
1016
1242
ND is not detectable.
1254
0
1
2
3
4
5
6
7
8
11
51
32
4
2
0.5
ND
ND
ND
0.1
1
20
57
21
1
0.1
ND
ND
0.1
1
16
49
25
8
1
0.1
ND
0.1
0.1
0.5
1
21
48
23
6
ND
1-50
-------
Aroclor 1221 from a saturated water solution. It is apparent that the
change in KQL with time is very slight, in fact, KQ^ is almost a
constant. Eighty- three percent of Aroclor 1221 is either mono- or
di-chlorobiphenyl (Table 1-25) . The mass transfer coefficients of
these components are probably similar in magnitude. For Aroclors 1016,
1242 and 1254 the percentage compositions are more widely scattered
among the isomeric components. The mass transfer coefficients would
therefore be different.
A problem associated with the experiments described above is
PCS adsorption to container walls. A final extraction of the water
and particulate matter failed to account for all the PCBs. Either
insufficient extraction or adsorption of PCBs to the glass occurred.
A rinse of the glassware with hexane-acetone would have failed to
show if the adsorption of PCBs was above or below the water's surface.
A steady inflow and outflow of compressed air was used to minimize
adsorption above the water's surface. Irreversible adsorption was
not a problem in the experiment of Hetling et al. (22) since greater
than 90% of the revolatilized PCBs were recovered by the vapor trap.
values were determined for each of the experiments. Since
the KQL versus time curves were not linear, both minimum and maximum
values of KQL were reported. The maximum value of KQL was determined
by extrapolating the curve to time zero while the minimum transfer
coefficient was .reported as the lowest observed value. Henry's Law
constants were then calculated by employing Eq. (2). Table 1-26
contains the experimentally determined values of KQ-T and H for the
Aroclor mixtures as well as the theoretical values. It is impossible
to use the theoretically determined values as a comparison since the
vapor pressures used to calculate H are for liquids. The experimental
H values are similar to those of DDT (3.89 x 10~5 atm-m3/mol) and
lindane (4.93 x 10~' atm-m^/mol) (42) which are solids at room
temperature. The vapor pressures of DDT and lindane would not be
expected to be the same as those of PCBs. The experiments reported
in this study show that a discrepancy exists between values of H
determined experimentally and previously reported theoretical values.
Conclusions
The Mackay and Leinonen (42) model of revolatilization, which can
be used to calculate H, cannot be purely applied to the experiments
performed in this study for reasons which have already been elaborated
upon. An estimate of a value of H for this experimental system can
still be made.
The method of quantification prevents an accurate application
of Henry's Law to the system. Each PCB isomer would have to be
quantified separately. Isolating each individual isomer in an Aroclor
mixture by present gas chromatographic methods is impossible. If
Aroclors were quantified by chlorine numbers, more accurate information
could be obtained. One assumption would still have to be made, i.e.,
positional isomers of the same chlorine number have the same mass
transfer coefficient. This remains to be shown but it may simplify
measurements and make the approach more environmentally realistic.
In the environment PCBs are a complex mixture, not single isomers.
1-51
-------
Experiments should be designed which model the environment as closely
as possible. J
The Henry's Law constants determined in this study should
actually be termed air/water partition coefficients. Aroclors
are mixtures and in actuality cannot have a Henry's Law constant
Table 1-26. Experimentally and theoretically derived Henry's
Law constants (H) and liquid phase mass
transfer coefficients (K )
OL
Aroclor
Theoretical*
Maximum
Minimum.
1221**
1016**
1242***
1254***
> m/hr
5.7 x 10
6.7 x 10
-2
-2
9.63 x 10
3.18 x 10
3.47 x 10-
8.44 x 10
H, atm-m3/mol
-3
4
5
*(Ref. 42).
**(Ref. 22).
***This study.
8.04 x 10";
7.31 x 10";
4.40 x 10~5
1.98 x 10~5
1221**
1016**
1242***
1254***
5.
2.
73
76
x 10
x 10
-4
-3
2.
1.
1.
2.
59
02
08
94
x
x
X
X
10~b
10-5
10~6
10~7
2.
2.
1.
6.
16 x
27 x
37 x
89 x
10~b
10"°
10~8
10"8
1-52
-------
1-7, FLUX CALCULATIONS
There are several processes that have to be considered when
estimating an atmospheric flux of PCBs to Lake Michigan. Fluxes both
to and from the lake must be investigated to obtain a net atmospheric
input. Wet and dry deposition represent fluxes of airborne PCBs to the
lake's surface. Dry deposition includes particulates and possibly
vapor. Revolatilization and bubble ejection of particulate matter must
be included as potential fluxes out of the lake. Bubble ejection is
a major natural source of atmospheric particulate matter over the
ocean (47,48); however, its significance for Lake Michigan has not
been investigated. Revolatilization of PCBs has been theoretically
investigated (5,42) but experimental evidence for the direction of
the PCS vapor flux is lacking. An experiment was described in the
results and discussion section which investigated the controlling
phase of the PCB vapor flux. The following sections deal with each
of these aspects for the purpose of estimating a net flux of PCBs to
Lake Michigan.
Washout Coefficients and the Wet Flux
Washout coefficients can be utilized to determine both the wet
flux of PCBs (using air concentration data) and the size of the
particulates that PCBs are associated with. The washout coefficient
for vapor depends upon the partitioning of the constituent between
vapor and water, essentially H, while that for a particulate depends
upon the collision efficiency of the raindrop and particle, which is
a function of diffusion, interception and impaction processes (49).
Since precipitation capture processes for the two phases are different,
the vapor/particulate partitioning is of extreme importance. The
extent of this partitioning is uncertain since it is not known if
the PCBs that pass through a glass fiber filter are small particulates
or vapor. Until further experimental evidence is obtained, it will
be assumed in some of the following arguments that the vapor/particulate
partitioning is operationally defined by the collection method.
If washout coefficients are to be employed to indirectly determine
particle size, the scavenged vapor component of precipitation must be
determined. Stracha et al. (50) monitored the PCB concentration
throughout a single precipitation event. They found most of the
deposition occurring in the early portion of the rain, which could
possibly indicate particulate washout. If a gas forms a simple solution,
molecular diffusion to the drop is fast enough for equilibrium to
occur within the atmosphere (49). Using this argument, if PCBs exist
partially as vapor in the atmosphere and are gas phase controlled,
the washout of PCB vapor could possibly show the same scavenging
profile as that of particulate PCB.
The following relation can be used to calculate gas scavenging
by rain (49):
1-53
-------
W = «POCb Eq. (5)
wnere
W is the wet flux of the contaminant
« is the washout coefficient which is RT/H
PQ is the annual amount of rainfall
C"b is the average contaminant concentration in air
A Henry's Law constant for Aroclor 1242 of 1.08 x 10~6 atm-m3/mol
was found in this study (Table 1-26). H can be converted to dimensionless
form by dividing by RT (2.45 x 10~2 atm-m3/mol at 25°C). Substituting
« = 2.27 x 104, po = 74 mm/yr (51), and Cb = 0.65 ng/m3 (this study)
into Eq. (5), the amount of Aroclor 1242 vapor scavenged annually
over Lake Michigan was calculated to be 635 kg/yr. Murphy and
Rzeszutko (2) reported that an average of 66% of the total PCBs in
rain collected in Chicago passed through a glass fiber filter. Stracha
et al. (50) also found that a majority of PCBs in rain collected
around Lake Superior was in "soluble" form.
A vapor phase washout coefficient can be calculated from the
precipitation data of Murphy and Rzeszutko (2). It will be assumed
that the filtered portion of the rain samples contained only scavenged
PCB vapor. They found that an average of 66% of the PCBs in their
Chicago area samples was in soluble form, 54% of that fraction was
Aroclor 1242. Applying these averages to the total wet flux (4800 kg/yr),
a scavenged Aroclor 1242 vapor flux of 1711 kg/yr can be calculated.
The following equation can be used to calculate the washout coefficient
(r) for Aroclor 1242 vapor (49):
r = W/p0Cb Eq. (6)
where
r = oc = RT/H Eq. (7)
Using Murphy and Rzeszutko's (2) value for the average Aroclor 1242
vapor concentration in Chicago air (C"b = 6.44 ng/m3), a washout
coefficient of 6.18 x 10^ was obtained. Applying Eq. (7), H^242 was
found to be 3.96 x 10~" atm-m3/mol, a remarkable similarity to the
value experimentally determined in this study. Table 1-27 contains
the atmospheric H values for Aroclors 1242 and 1254 along with the
laboratory values.
Atkins and Eggleton (52) reported values for atmospheric and
laboratory values of pesticide washout ratios (Table 1-28). The
laboratory washout coefficients they obtained fory-BHC, p, p'-DDT
and Dieldrin, which they termed partition coefficients, were actually
the reciprocal of the Henry's Law constant. The atmospheric values
were calculated from air and rain concentrations but unfortunately
the samples were not taken concurrently. They found good agreement
between the laboratory and atmospheric values for Y-BHC and Dieldrin
but somewhat poorer agreement for p, p'-DDT. They attributed the
disagreement to a larger proportion of airborne DDT being present in
the particulate phase. Their study, while not giving entirely
conclusive evidence, may indicate that utilizing the reciprocal of H
1-54
-------
Table 1-27. Vapor phase washout coefficients and Henry's constants for Aroclors 1242 and 1254.
Samples taken by (2)
M
Ln
Ui
Aroclor
1242
1254
Scavenged vapor
flux, kg/yr
1711
1045
Vapor
concentration, ng/m3
6.44
0.97
Washout Henry's constant, atm-m3/mol
coefficient Atmos. Lab.
6.18 x 10^ 3.96 x 10"!? 1.08 x 10"^
2.50 x 104 9.80 x 10~7 2.94 x 10~7
Table 1-28. Atmospheric and laboratory determined washout ratios for pesticides (52)
Pesticide
Concentration,* parts per 1012
AirRain
Washout ratio, rain:air
Atmospheric Laboratory
y-BHC
p.p'.-DDT
Dieldrin
5 to 11
3
20
66
194
48
6 to 13
65
2.4
14
2.3
0.65
-------
for vapor phase washout coefficients of pesticides and PCBs may be
valid.
Using three different assumptions, more than one washout coefficient
can be calculated from the air and rain data reported by Murphy and
Rzeszutko (2). Assumption 1 is that all the PCBs in air are associated
with particulate matter. Assumption 2 is that all the PCBs in rain
are due to scavenged particulates but only the non-filterable fraction
of an air sample is composed of particulate associated PCBs.
Assumption 3 is that air and rain samples contain both a particulate /
and vapor phase component which are operationally defined by filtration
through a glass fiber filter. Using the latter assumption, both a
particulate and vapor phase washout coefficient can be determined.
Murphy and Rzeszutko also measured washout coefficients by concurrently
sampling air and rain during 2 precipitation events. These coefficients
were based on total PCB concentrations in air and rain. Equation (6)
can be used to calculate washout coefficients.
By using assumption 1, the total PCB wet flux to Lake Michigan
(4800 kg/yr) and the average total PCB concentration in Chicago air
(7.7 ng/m3) a washout coefficient of 1.45 x 10^ was calculated. A
washout coefficient of 3.84 x 10^ was obtained by employing assumption 2,
the average concentration of particulate associated PCBs in Chicago air
(0.29 ng/m^) and the total PCB wet flux. A particulate phase washout
coefficient of 1.31 x 10-" was calculated by using assumption 3, the
non-filterable fraction of the wet flux (0.34 x 4800 kg/yr) and an
average air concentration of particulate associated PCBs of 0.29 ng/m-5.
By using assumption 3, the filtered fraction of the wet flux (0.66 x
4800 kg/yr) and an average vapor concentration in air of 7.48 ng/trH,
a vapor phase washout coefficient of 9.86 x 1CH was obtained. Murphy
and Rzeszutko (2) also determined PCB washout coefficients (g PCB/L of
rain per g PCB/L of air) of 13 and 37 for two separate rain events.
An average value of 2.08 x 10^ was obtained by converting the average
coefficient to a dimensionless constant using a conversion factor
(10^ cm3/m3 per 1.2 x 103 g/cm3) to correct for the density of moist
air. The calculated particulate PCB washout coefficients are in the
same range as certain trace metals, such as lead, which has a washout
coefficient of 2.4 x 105 (53).
Using the data presented by Slinn et al. (49) and the calculated
washout coefficients, the size of the particles the PCBs are associated
with can be estimated. From their graph of particle size versus
collision efficiency of rain droplets (Fig. 1-14), a washout coefficient
(r) can be determined for each particle size (Table 1-29) by using
the following relation:
r = ch^/R^ Eq. (8)
where
c is an empirical constant which equals 0.5
hw is the effective height from which the pollutant is removed
by wet processes (500 m)
Eg is the collision efficiency
RJUQ is the drop radius (0.1 mm).
1-56
-------
10
-1
o
10
-2
o
u
10
10
-3
10
-5
10
-6
DIFFUSION
I/?
1+0.4 Re1^ Sc
DATA:
INTERCEPTION
1/2 ,.1/31
K 4
vRe
-1/2
1MPACT10N
JL^3/2
ADAM AND SEMON1N (1970), R-0.5mm
SOOD AND JACKSON (1972), R- 0.5mm*
KERKER AND HAMPL (1974], R- 1.58mmt
* DROP JUST REACHED V,
I R-O.lmm
R- 1mm
t DROPS NOT AT V
t
V4"
R-0.5mm „
Pe- Re Sc
Sc-v/D
Re- RV./v
S-Vfr/R
— C-2/3-S*
S^- 12/10 + 1/12 /*. !l+Re)
1 + In, (1+Re)
E, NO SLIP
Re«100, R- 1.58mm
10
-3
I 1 l_l
i i i i ,i<
,-1
10 - 10 * 10
RADIUS OF UNIT DENSITY SPHERES, a (pm)
0
10'
Fig. 1-14. Collision efficiency against particle size (49).
1-57
-------
Table 1-29. Collision efficiencies and washout ratios for
different particle sizes
Particle
diameter, ym
10
2
1
0.2
0.1
0.02
0.01
0.002
Collision
efficiency,* E
5.0 x 1CT1
3.1 x 10~3
1.2 x 10~3
6.2 x 10~4
9.3 x 10~4
7.9 x 10~3
2.2 x 10~2
2.4 x 10"1
Washout
ratio, y
1.25 x 106
1.75 x 103
3.03 x 103
1.55 x 103
2.33 x 103
1.98 x 104
5.5 x 104
6.0 x 105
*For 0.1 mm drop size.
Table 1-30. PCB washout coefficients and particle sizes
Washout coefficient
Particle diameter range, ym
1.
3.
2.
1.
45
84
08
31
x
x
X
X
10
10
10
10
4*
5**
4***
5*
0
0
0.02
.002
0.01
.002
to
to
to
to
0.
0.
0.
0.
1
01
02
01
*Calculated by assumption 1. **Calculated by assumption 2.
***Determined from average washout ratio for two rains and
assumption 1.
^Calculated using assumption 3.
Table 1-31. Predicted PCB concentrations in precipitation
Washout
coefficients
PCBs in air, ng/m3
L. Michigan
L. Superior
PCBs in precipitation, ng/L
L. Michigan L. Superior
1.45 x IO4*
3.85 x IO5**
2.08 x IO4**
1.31 x IO5*
9.86 x IO3*
1.0
0.13
1.0
0.13
0.87
0.62
0.08
0.62
0.08
0.54
15
50
21
17 ) 26***
9 )
9
31
16
10
5
) 15***
*,**,***,* as for Table 1-30.
***Total PCB in rain, ng/L.
**Using 20 yg/m3 TSP.
1-58
-------
Employing a drop radius of 0.1 mm fit the PCB washout coefficients
better than the 1 mm size. The washout coefficients for PCBs and
their corresponding particle sizes can be found in summarized form
in Table 1-30. From the calculated particle sizes it is apparent that
PCBs are associated with Aitken particles (d < 0.2y), which are the
smallest of particles present in the atmosphere. Gatz (57) obtained
a curve for washout ratios versus mass median diameter of certain
particulate associated trace metals. Washout coefficients of the same
magnitude as those calculated for PCBs were obtained for particles
having mass median diameters
-------
Using vapor and particulate phase washout coefficients of
9.86 x 1(P and 1.31 x 105, a Lake Michigan wet flux can be calculated.
If it is assumed that the average air concentration of particulate
and vapor phase PCBs is 0.13 and 0.87 ng/m^, respectively (this study),
then the wet flux can be calculated as 1101 kg/yr with 369 kg/yr
being due to scavenged vapor.
The calculated washout coefficients indicate that PCBs are
associated with the smallest particles present in the atmosphere,
probably in the range of 0.002 to O.ly in diameter. The washout
coefficients may also indicate that atmospheric PCB vapor does exist
but the extent of the vapor/particulate partitioning cannot be
estimated. Until the extent of the partitioning is known, the best
method to estimate wet fluxes from air concentrations is to employ
particulate and vapor phase washout coefficients. The calculations
made in this section were based on data reported in several studies
which may produce large errors in the final results; nevertheless, they
indicate that more data must be obtained to determine the relationships
between vapor and particulate phase PCBs and their capture by pre-
cipitation.
Particulate Flux
The following equation can be used to calculate a particulate
flux (60):
F = VdCb Eq. (9)
where
\f
-------
10
v/1
E
u* zo u MOcrn)
(cm s ) (cm) (m s"*) .
A 11 Q002 2.2*
Q— 44 Q02 7.2* *
t-0 117 Ql 13.8*
X 40 ~0.05 ~g**
*SEHMELAND SLITTER (1974)
**MOLLER AND SHUMANN (1970)
o
o
o
a,
in
lu
10
-2
\
10
-2
10
10
PARTICLE Dl A/METER, pm
Fig. 1-15. Deposition velocity against particle size (61).
1-61
-------
PCS evaporation rates from Lake Michigan have been made (5). Due to
the scarcity of experimental evidence for liquid or gas phase control
of PCBs, the experimental data presented in Part 1-6 can be interpreted
in two ways. The magnitude of the air/water partition coefficients,
which indicated gas phase control, can be used to estimate the flux
of PCBs to the lake. Liquid phase control could be assumed and the
experimentally determined KQ^ used to estimate the revolatilization
rate of PCBs.
Assuming gas phase control and the microlayer to be a perfect
PCB adsorber, Bidelman et al. (21) calculated the vapor flux to the
North Atlantic Ocean using the following equation:
F = kGCb Eq. (10)
where
JCQ is the gas phase mass transfer coefficient
C]., is the average PCB concentration in air
This relation could be utilized to obtain a maximum value for the
influx of vapor phase PCBs to Lake Michigan. Using the values of
kg for Aroclors 1242 and 1254 (Table 1-22) and the average vapor
concentrations of 0.65 and 0.22 ng/m , respectively, the average flux
of Aroclor 1242 was found to be 2619 kg/yr while that of Aroclor 1254
was 788 kg/yr (Table 1-32). The total PCB vapor flux to the lake,
then, is 3407 kg/yr with a range of 1724 to 4347 kg/yr.
A more rigorous approach to calculating the vaoor flux to Lake
Michigan would be to use the following modified version of the flux
equation from Liss and Slater (44) :
N = KOG(CH-P)/RT Eq. (11)
where
is the overall gas phase mass transfer coefficient
C is the soluble concentration of PCBs
H is the Henry's Law constant
P is the partial pressure of PCBs
RT is 2.45 x 1Q-3 atm-m3/mol at 25 °C
K was calculated from the following relation (44) :
1/K__ = 1/k., + H/kT Eq. (12)
U(jr (j L,
where
kg is the gas phase mass transfer coefficient
k, is the liquid phase mass transfer coefficient
Maximum and minimum values of KQQ and H (Table 1-33) determined from
the experimental values of KQL (Part 1-6) were used to determine the
vapor flux (Table 1-34). The maximum average value was found to be
1-62
-------
Table 1-32. Vapor flux to Lake Michigan (F
Type of
value
Average
Maximum
Range
Minimum
Aroclor
1242
1254
1242
1254
1242
1254
kg, m/hr
7.93
7.05
7.93
7.05
7.93
7.05
cb,
ng Aroclor/m3
0.65
0.22
0.83
0.28
0.33
0.11
F,
kg Aroclor/yr
2619
788
3344
1003
1330
394
F,
kg total PCB/yr
3407
4347
1724
Table 1-33. Henry's constants (H) and gas phase mass transfer coefficients
(KQG) for Aroclors
Arochlor
H, atm-m3/mol
Maximum
Minimum
Maximum
m/hr
Minimum
1221
1061
1242
1254
2.59 x 10
1.02 x 10
1.08 x 10
2.94 x 10
-6
-5
-6
-7
2.16 x 10
2.27 x
1.37 x
10
10
6.89 x 10
-6
-6
-7
-8
9.10
7.64
7.85
7.03
9.12
8.00
7.88
7.04
1-63
-------
Table 1-34. Vapor flux to Lake Michigan [N = KQG(CH-P)/RT]
H
I
ON
•P-
Type of
value
Average
Maximum
Range
Minimum
KQG» H>
Aroclor m/hr atm-m3/mol
1242 7.85 1.08 x 10~6
1254 7.03 2.94 x 10~7
1242 7.88 1.37 x 10~7
1254 7.04 6.89 x 10~8
1242
1254
1242
1254
1242
1254
1242
1254
C,* P,**
mol/m3 atm
7.66 x W~9 5.58 x 10~14
2.57 x 10 8 1.51 x 10~14
6.60 x 10~14
2.33 x 10~14
3.00 x 10~14
6.17 x 10~15
N,
kg Aroclor/yr
2019
360
2335
636
2457
751
2774
1028
923
66*4=
1235
210
N,
kg total PCB/yr
2379***
2379***
2971*
2971*
3208***
3208***
38024=
38024=
857***
857***
1025+
10254=
*(Ref. 2).
**This study.
***Calculated using maximum values of Hand
4=Calculated using minimum values of Hand
4=4=Flux out of lake.
-------
2335 kg of Aroclor 1242/yr and 636 kg of Aroclor 1254/yr for a total
vapor flux of 2971 kg/yr.
Deposition velocities of 0.04 cm/sec for Y-ZEC, Dieldrin and
p,p'-DDT vapors on turf in wind tunnel experiments have been reported
by Atkins and Eggleton (52). While the deposition velocity for a
water surface would not be expected to be the same as that for turf,
the value could be used to give a lower limit of the vapor flux.
Using this deposition velocity and an average vapor phase concentration
of 0.87 ng/np, an annual flux of 637 kg was obtained.
Another way to interpret the vapor flux would be to assume liquid
phase control of PCBs. The liquid phase control model of Mackay and
Leinonen (42) can be used to calculate the revolatilization rate of
PCBs from Lake Michigan by utilizing the following equation:
N = KOL(C-P/H) Eq. (13)
Using the KQL and H values they theoretically determined and the
concentration of PCBs in air and water for Lake Michigan, a revola-
tilization rate of 3.43 x 10^ kg/yr was obtained. The KQL values
determined in this study can also be used to calculate a revolatilization
rate (Table 1-35). Since the KQL values determined in this study are
actually a combination revolatilization and desorption rate, the total
PCB concentration of PCBs in Lake Michigan were used in the calculation.
A maximum flux of 3132 kg/yr was calculated.
Cohen et al. (63) found that the liquid phase mass transfer
coefficients of benzene and toluene were a function of the wind and
stirring speed in an experimental wind-wave tank. The two relations
they obtained were:
kT = 11.4 Re*0'195 - 5.0 0 rpm Eq. (14)
Li
kL = 11.4 Re*0-195 - 4.1 540 rpm Eq. (15)
where
Re* is the roughness Reynolds number.
Re* is defined by the following relation:
Re* = zQU*A)a Eq. (16)
where
ZQ is the effective roughness height
U* is the friction velocity
\)a is the kinematic viscosity (1.506 x 10~1 cm^/sec at 18°C).
1-65
-------
Table 1-35. Revolatilization rates from Lake Michigan
Aroclor
1242
1254
1242
1254
1242
1254
KQL, H,
m/hr atm-m3/mol
5.73 x 10~4
2.76 x 10~3
3.47 x 10~4
8.44 x 10~5
4.40 x 10~5
1.98 x 10~5
C,*
mol/m3
7.66 x 10~9
2.57 x 10~8
4.75 x 10~8
6.82 x 10~8
P,** N,
atm kg Aroclor/yr
5.58 x 10~14 57,200
1.51 x 10~14 286,000
2,175
957
277
224
N,
kg total PCB/yr
343,200***
343,200***
3,1324=
3,1324=
5014=4=
5014=4=
*(Ref. 2).
**This study.
***(Ref. 42).
4=Calculated using maximum values of KQ-^ in this study.
4=4=Calculated using minimum values of KQL in this study.
-------
Values of ZQ and U* were obtained from a graph presented in their study
and used along with "V to calculate k-r for different wind speeds. It
was then assumed that the profile of the wind speed versus PCB mass
transfer coefficient curve would be the same as that of the benzene
and toluene curves. Using this assumption and the KQL values obtained
in this study, PCB liquid phase mass transfer coefficients were
calculated for different wind speeds (Table 1-36). Since the stirring
speed in the experiment presented in this study (Part 1-6) was unknown,
540 rpm was assumed. The exact dependence of stirring speed on kL has
not been determined (63). Revolatilization rates as a function of
wind speed were calculated using Eq. (13) (Table 1-37). Wind speed
data taken periodically over 3 summers on Lake Michigan showed an
average of 4.7 m/sec (35). Under these wind speed conditions the
PCB revolatilization rate for Lake Michigan would be 1.72 x 10^ to
1.13 x 106 kg/yr.
The validity of the revolatilization rates can be checked by
determining if they can account for the enrichment observed in the
microlayer. Elzerman (64) reported microlayer concentrations on
Lake Michigan of 78 to 1090 ng/L with concentrations in the bulk water
being 5 to 64 ng/L. The screen technique he employed to collect the
microlayer is thought to sample the top 300y of the water surface.
Knowing this fact, the amount of PCB in 1 m^ of surface area can be
calculated. The wet flux (4800 kg/yr) and particulate flux (1189
V.g/yr) were combined to give the downward flux of PCBs. The revol-
atilization rates under quiescent conditions were used as the upward
flux. The residence time of the microlayer was calculated by
subtracting the time involved in depositing an amount of PCB equal
to that in the microlayer from the time involved in stripping that
amount (Table 1-38). It is apparent that the microlayer would not
exist under the conditions of the KQL calculated by Mackay and
Leinonen (42). The evaporation rate estimated by Versar (5) and the
maximum and minimum KQT values determined in this study are more
realistic in that they give residence times >1 hr.
Two interpretations have been given to elucidate the direction
and magnitude of the PCB vapor flux. Gas phase control of PCBs is a
new interpretation. Besides the experiment performed in this study
and that of Hetling et al. (22), experimental evidence is lacking.
Further experimentation must involve the determination of accurate
Henry's Law constants for PCBs. Liquid phase control of PCBs has been
hypothesized (42). The KQ^ values determined in this study show that
adsorption of PCBs to particulate matter slows the rate of revolatiliza-
tion. Extrapolating the KQ^ values to Lake Michigan may involve certain
errors. The desorption rate of the particulate matter used in this
study may not be the same as that of Lake Michigan particulates.
Also, all the PCBs in Lake Michigan water may be adsorbed to particulate
matter and unless desorption kinetics are rapid and the revolatilization
rate is faster than the 'sedimentation rate, revolatilization may not
occur. Another uncertainty is the effect of the microlayer on
revolatilization rates. Factors which may affect the revolatilization
rate must be further explored.
1-67
-------
Table 1-36.
against wind speed
Wind speed
cm/sec
0
300
600
1150
KL>*
cm/hr
2.00
2.97
8.38
30.50
KOL(1242), cm/hr
Maximum Minimum
3.47 x 10~2 4.40 x 10~3
1.00 9.88 x lO'1
6.41 6.38
28.50 28.50
KOL(1254),
Maximum
8.44 x 10~3
9.78 x 10"1
6.39
28.50
cm/hr
Minimum
1.98 x 10" 3
9.72 x 10"1
6.38
28.50
*(Ref. 63).
Table 1-38. Microlayer residence times against revolatization rates
Revolatilization
rate, kg/yr
343,000**
1,160***
3,132*
501**
Deposition
rate, kg/yr
5,989
PCB in
microlayer,* g/m2
2.34 x 10~8
Residence
time, hr
0
8
2
22
*Calculated using minimum observed concentration (78 ng/L) (64).
**Estimate based on data in (Ref. 42).
***(Ref. 5).
^Calculated from maximum KQL in this study.
^Calculated from minimum K-T in this study.
I~68
-------
Table 1-37. Revoiatilization rates from Lake Michigan against wind speed
Wind speed, KQL*
Aroclor cm/sec
1242
300
1254
1242
600
1254
1242
1150
1254
m/hr
0.0099
0.0097
0.064
0.064
0.285
0.285
H,* C,** ps*** N,
atm-m3/mol mol/m3 atm kg Aroclor/yr
5.73 x 10~4 4.75 x 10~8 5.58 x 10~14 6.22 x 104
2.76 x 10~3 6.82 x 10~8 1.31 x 10~14 1.10 x 10^
4.02 x 105
7.24 x 10;?
1.79 x 106
3.22 x 105
N
kg total PCB/yr
1.72 x 105
1.13 x 106
2.11 x 106
*(Ref. 42).
**(Ref. 2).
***This study.
-------
Bubble Ejection
Bubble ejection of particulate matter must be considered as a
potentially important water to air transport process. The ocean is
a major natural source of atmo&pheric particulate matter (47,48). The
process which produces the particulates is bubbles breaking at the
air/water interface (65). These bubbles strip off a layer of
0.025 to 0.75 pm of the air/water interface. The composition of the
scavenged material depends upon the composition of both the surface
microlayer and the bubble skin. It has been estimated that the ocean
produces 10^-5 to 10^" g/yr of atmospheric sea salt particles (47,48).
Estimates of this kind have not been performed for freshwater lakes.
To estimate the amount of PCBs injected into the atmosphere
from Lake Michigan, the composition and enrichment of the microlayer
must be known. Andren et al. (66) perpared a review paper dealing
with some of the physical and chemical aspects of surface microlayers
in freshwater lakes. Elzerman (64) performed a study on surface
organic microlayers in Lake Michigan. He determined the enrichment of
PCBs in the microlayer in several locations finding an average microlayer
concentration of 245 ng of Aroclor 1254/L while that of the bulk
water was 22 ng of 1254/L. The maximum enrichment observed was under
conditions where the microlayer and bulk water concentrations were
664 ng/L and 5.2 ng/L, respectively. The actual depth sampled by
the screen technique is thought to be about 300y (64). Most freshwater
surface microlayers are 10 to 30 nm thick (67). Knowing the concen-
tration in the bulkwater, the concentration in the microlayer and
the actual depth sampled, a simple calculation can be performed to
obtain the concentration in the actual surface microlayer. The average
concentration in a 30 nm thick microlayer was calculated to be 2.2 ppm
while that of the maximum enrichment case was 6.7 ppm.
To estimate the flux of particulate PCBs from Lake Michigan to the
atmosphere, the particulate flux for the ocean must be recalculated
to compensate for the differences in area of the two environments.
Due to the salinity differences of the two bodies of water, a salt
correction must also be applied since the composition of the ejected
material over the ocean would be expected to be more enriched in
salt than the material ejected from Lake Michigan. The average total
dissolved solids concentration of Lake Michigan is 157 mg/L (68)
while the average salinity of the ocean is 35 g/L, the salt correction
factor is then 0.157/35. The estimated amount of particulate matter
ejected from Lake Michigan was calculated to be 7 x 10** to 7 x 10^ g/yr.
An assumption must then be made that the PCB concentration of the
microlayer is similar to that of the material ejected. This is logical
since a jet droplet carries away part of the microlayer (65), although
other factors may dilute or enhance the concentration. The composition
of the bubble skin also contributes to the amount of material ejected.
Wallace and Duce (69) have shown that bubbles may also adsorb materials
in the water column before they reach the surface. Employing a
concentration of PCBs in the microlayer of 2.2 ppm, a range of the
annual flux of PCBs of 2 to 16 kg was obtained. If the maximum micro-
layer enrichment values are used, a flux of 5 to 46 kg/yr is cal-
culated. If the microlayer depth is 10 nm the fluxes would be three
times as great as the ones calculated.
1-70
-------
These calculations would seem to indicate that the ejection of
PCBs into the atmosphere from Lake Michigan is not as important a
process as it could be in the ocean. The numbers obtained should be
considered a rough estimate since the significance of bubbles bursting
in freshwater lakes has not been investigated.
Photolysis
Photolysis must be considered as a degradative pathway which may
alter the amount of PCS deposition. Photolysis of PCBs in water as
well as in a number of solvents has been reported (36,37,38,70). The
significance of PCB photolysis in the atmosphere has not been inves-
tigated. The main reaction of photolysis of PCBs is stepwise dechlor-
ination. If extrapolation from a solvent media to the atmosphere
could be made, this would mean photolysis of airborne PCBs would
change the composition of an air sample with the extent of the process
depending upon the residence time in air.
Bunce et al. (70) performed some calculations to estimate the
impact of PCB photolysis in Lake Erie. The quantum yield of PCBs
must be known in order to adapt their model to calculate the impact
of photolysis in the atmosphere. The quantum yields they reported
were determined"in solvents such as acetonitrile-water and isooctane.
Information is not available for the determination of quantum yields
in the atmosphere. Solvents can influence the rate of photolysis.
Ruzo et al. (38) found different rates of tetrachlorobiphenyl photo-
lysis in different solvents. They also found that the presence of
02 decreased the photolysis rate, sometimes by a factor of four.
Herring et al. (36) found that Aroclor 1254 degraded most rapidly in
hexane, then water, and slowest in benzene. Since the solvent plays
an important role in the rate of photolysis, it is extremely difficult
to extrapolate rates, taken in any other media except air, to the
atmospheric environment. Studies on the photolysis of PCBs under
simulated atmospheric conditions must be performed before a reasonable
estimate of its importance can be made.
The fluxes both to and from Lake Michigan have been estimated in
the preceding discussion. Models for both the liquid and gas phase
control of PCBs have been discussed. Table 1-39 contains net atmos-
pheric input estimates to Lake Michigan based upon the two models.
The Lake Michigan estimates show that vapor deposition or revolatilization,
depending upon which model is the correct one, are the most important
fluxes.
Combining the data obtained in this study with the input data
reported by Murphy and Rzeszutko (Table 1-1), it is apparent that the
atmosphere is now the major source of PCBs to Lake Michigan (Table 1-40).
as also concluded by Murphy and Reszutko (2). The known industrial
discharges prior to 1975 are an order of magnitude greater than the
annual atmospheric input. Further experimentation must be performed
to determine which phase controls the vapor flux of PCBs so more
accurate estimates can be made.
1-71
-------
Table 1-39. Estimated net atmsopheric inputs to Lake Michigan based
on gas and liquid phase control model
Flux
Particulate
Vapor
Wet
Bubble ejection
Revolatilization
Net atmospheric input
Gas phase control,
kg PCB/yr
1189
2675
4800** 1101***
9
0
8655
Liquid phase control
kg PCB/yr
1189
0
4800** 1101***
9
3132+
2848
*Average using maximum and minimum values of H found in this study.
**Data taken from (2).
***Estimate based on air concentrations found in this study.
+Based on KQ-^ values determined in this study under quiescent conditions,
Table 1-40. Inputs of PCBs (kg/yr) to Lake Michigan*
Sources
Industrial discharges
Atmospheric
Streams and wastewater
Total
Prior
LPC
25,000
2,848
750
28,598
to 1975
GPC
25,000
8,655
750
34,405
LPC
2,848
750
3,598
1977
GPC
8,655
750
9,405
*Data in Table 1-1 incorporated into estimates. LPC and GPC are
liquid and gas phase control, respectively.
1-72
-------
REFERENCES-I
1. Peel, D. A. Organolchlorine Residues in Antarctic Snow. Nature 254:324-
325, 1975.
2. Murphy, T. J. and C. P. Rzeszutko. Polychlorinated Biphenyls in
Precipitation in the Lake Michigan Basin. EPA Grant-803915.
Environmental Research Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Duluth, Minnesota, 1977.
3. Nisbet, I. C. T. and A. F. Sarofim. Rates and Routes of Transport of
PCBs in the Environment. Environ. Health Persp. 1:21-38, 1972.
4. Hutzinger, 0., S. Safe and V. Zitko. The Chemistry of PCBs. CRC Press,
Inc., Cleveland, Ohio, 1972.
5. Versar, Inc. A First Order Mass Balance Model for the Sources,
Distribution and Fate of PCBs in the Environment. Submitted to U.S.
Environmental Protection Agency, Office of Toxic Substances, Special
Projects Branch, Washington, D.C., 1977.
6. Wallnofer, P. R., M. Kariger and 0. Hutzinger. The Solubilities of
Twenty-One Chlorobiphenyls in Water. Analabs, Inc., Research Notes 13:3,
14-16, 1973.
7. Haque, R. and D. W. Schmedding. A Method of Measuring the Water
Solubility of Hydrophobic Chemicals: Solubility of Five Polychlorinated
Biphenyls. Bull, of Environ. Contam. and Toxicol. 14:13-18, 1975.,
8. Mackay, D. and A. W. Wolkoff. Rate of Evaporation of Low-Solubility
Contaminants from Water Bodies to Atmosphere. Environ. Sci. Technol.
7(7):611-614, 1973.
9. Mieure, J. P., 0. Hicks, R. G. Kaley and V. W. Saeger. Characterization
of Polychlorinated Biphenyls. In: Proc. of the National Conf. on Poly-
chlorinated Biphenyls. U.S. Environmental Protection Agency Report No.
EPA-560/6-75-004, pp. 84-93, 1975.
10. Miles, J. W., L. E. Fetzer and G. W. Pearce. Collection and
Determination of Trace Quantities of Pesticides in Air. Environ. Sci.
Technol. 4(5):420-425, 1970.
11. Bidelman, T. F. and C. E. Olney. High Volume Collection of Atmospheric
Polychlorinated Biphenyls. Bull. Environ. Contain, and Toxicol. 11:5,
442-450, 1974a.
1-73
-------
12. Harvey, G. R. and W. G. Stelnhauer. Atmospheric Transport of
Polychlorobiphenyls to the North Atlantic. Atmos. Environ. 8:777-782,
1974.
13. Environmental Science and Engineering, Inc. Evaluation of a Method for
the Analysis of Airborne Polychlorinated Biphenyls. P. 0. Box 13454,
University Station, Gainesville, Florida, 1976.
14. Lewis, R. G., A. R. Brown and M. D. Jackson. Evaluation of Polyurethane
Foam for Sampling of Pesticies, Polychlorinated Biphenyls and
Polychlorinated Naphthalenes in Ambient Air. Anal. Chem. 49(12):1668-
1672, 1977.
15. Risebrough, R. W., R. J. Huggett, J. J. Griffin and E. D. Goldberg.
Pesticides: Transatlantic Movements in the Northeast Trades. Science
159:1233-1235, 1968.
16. Seba, D. B. and J. M. Prospero. Pesticides in the Lower Atmosphere of
the Northern Equatorial Atlantic Ocean. Atmos. Environ. 5:1043-1050,
1971.
17. Sodergren, A. Chlorinated Hydrocarbon Residues in Airborne Fallout.
Nature 236:395-397, 1972.
18. Bengston, S. A. and A. Sodergren. DDT and PCB Residues in Airborne
Fallout and Animals in Iceland. Ambio 3:84-86, 1974.
19. Young, D. R., D. J. McDermott and T. C. Hessen. Polychlorinated Biphenyl
Inputs to the Southern California Bight. Southern California Coastal
Water Research Project, 1500 East Imperial Highway, El Segundo,
California, 1975.
20. Bidelman, T. F. and C. E. Olney. Chlorinated Hydrocarbons in the
Sargasso Sea Atmosphere and Surface Water. Science 182:516-518, 1974b.
21. Bidelman, T. F. Personal Communication, 1978.
22. Hetling, L., E. Horn and J. Tofflemire. Summary of Hudson River PCB
Study Results. New York State Department of Environmental Conservation,
Bureau of Water Research, 50 Wolf Road, Albany, New York, 1978.
23. Bidelman, T. F., C. P. Rice and C. E. Olney. High Molecular Weight
Chlorinated Hydrocarbons in the Air and Sea: . McDermott and T. C.
Hessen. Polychlorinated Biphenyl Inputs to the Southern California
Bight. Southern California Coastal Water Research Project, 1500 East
Imperial Highway, El Segundo, California, 1975.
20. Bidelman, T. F. and C. E. Olney. Chlorinated Hydrocarbons in the
Sargasso Sea Atmosphere and Surface Water. Science 182:516-518, 1974b.
21. Bidelman, T. F. Personal Communication, 1978.
1-74
-------
22. Hetling, L., E. Horn and J. Tofflemire. Summary of Hudson River PCB
Study Results. New York State Department of Environmental Conservation,
Bureau of Water Research, 50 Wolf Road, Albany, New York, 1978.
23. Bidelman, T. F., C. P. Rice and C. E. Olney. High Molecular Weight
Chlorinated Hydrocarbons in the Air and Sea: Rates and Mechanisms of
Air/Sea Transfer. In: Marine Pollutant Transfer, H. L. Windom and R. A.
Duce (eds.). Lexington Books, Lexington, Massachusetts, 1977. pp. 323-
351.
24. McClure, V. E. and J. Lagrange. Deposition of Heavy Chlorinated
Hydrocarbons from the Atmosphere. Bull, of Environ. Contain, and toxicol.
17(2):219-224, 1977.
25. Webb, R. G. and A. C. McCall. Quantitative PCB Standards for Electron
Capture Gas Chromatography. J. Chrom. Sci. 11:366-373, 1973.
26. Mackay, D. Personal Communication, 1978.
27. Junge, C. E. Basic Considerations about Trace Constituents in the
Atmostphere as Related to the Fate of Global Pollutants. In: I. H.
Suffet (ed.). Proc. ACS Symposium on the Fate of Pollutants in the Air
and Water Environment. Wiley-Interscience, 1977. pp. 7-25.
28. Sodergren, A. Monitoring DDT and PCB in Airborne Fallout. In: F.
Coulston and F. Korte (eds.). Environmental Quality and Safety,
Supplement Volume III. Georg Thieme Publishers, 1975. pp. 803-810.
29. Junk, G. A., J. J. Richard, M. D. Greiser, D. Witiak, J. L. Witiak, M. D.
ArguellO, R. Viek, H. J. Svec, J. S. Fritz and G. V. Valder. Use of
Macroreticular Resins in the Analysis of Water for Trace Organic
Contaminants. J. Chromatog. 99:745-762, 1974.
30. Ketseridis, G., J. Hahn, R. Jaenicke and C. Junge. The Organic
Constituents of Atmospheric Particulate Matter. Atmos. Environ. 10:603-
610, 1976.
31. Armour, J. A. Quantitative Perchlorination of PCBs as a Method for
Confirmatory Residue Measurement and Analysis. J. Assoc. Off. Anal.
Chem. 56:987, 1973.
32. Burkhard, L. P. Personal Communcation, 1978.
33. Stolzenburg, T. and A. W. Andren. International Joint Commission
Menomonee River Pilot Watershed Study, Vol. 6. Atmospheric Chemistry.
U.S.-Environmental Protection Agency, 1978.
34. Thomas, J. F., M. Mukai and B. D. Tebbens. Fate of Airborne Benzo (a)
Pyrene. Environ. Sci. Technol. 2:33-39, 1968.
35. Strand, J. S. Personal Communication, 1978.
36. Herring, J. L., E. J. Hannan and D. D. Bills. UV Irradiation of Aroclor
1254. Bull. Environ. Contam. and Toxicol. 8(3):153-157, 1972.
1-75
-------
37. Hutzinger, 0., S. Safe and V. Zitko. Photochemical, Degradation of
Chlorobiphenyls (PCBs). Environ. Health Persp. 1:15-20, 1972.
38. Ruzo, L. 0., M. J. Zubik and R. D. Schuetz. Photochemistry of Bioactive
Compounds: Photoproducts and Kinetics of Polychlorinated Biphenyls. J.
Agric. Food Chem., 22(2):199-202, 1974.
>
39. Rondia, D. Sur La Volatilite Des Hydrocarbures Polycycliques. Int. J.
Air Water Pollut. 9:113-121, 1965.
40. Lockhart, L. B., Jr., R.L. Patterson, Jr. and W. L. Anderson.
Characteristics of Air Filter Media Used for Monitoring Airborne
Radioactivity. NRI Report 6054. U.S. Naval Research Laboratory,
Washington, B.C., 1964.
41. Pupp, C., R. C. Lao, J. J. Murray and R. F. Pottle. Equilibrium Vapour
Concentrations of Some Polycyclic Aromatic Hydrocarbons, As/0/-, and SeO?
and the Collection Efficiencies of These Air Pollutants. Atmos. Environ.
8:915-925, 1974.
42. Mackay, D. and P. J. Leinonen. Rate of Evaporation of Low-Solubility
Contaminants from Water Bodies to Atmosphere. Environ. Science and
Technol. 9(13):1178-1180, 1975.
43. Whitman, W. G. The Two-Film Theory of Gas Absorption. Chem. Metall.
Eng. 29:146, 1923.
44. lisa, P. S. and P. G. Slater. Fluxes of gases across the Air-Sea
Interface. Nature, 247:181-184, 1974.
45. Veith, G. Interdepartmental Task Force on PCBs. U.S. Department of
Commerce. COM-72-10419, 1972. p. 94.
46. Castellan, G. W. Physical Chemistry. Addison-Wesley Publishing Co.,
Inc. Reading, Massachusetts, 1971. 319 pp.
47. Eriksson, E. The Yearly Circulation of Chloride and Sulfur in Nature:
Meteorological, Geochemical, and Pedological Implications. Part I.
Tellus 11:375-403, 1959.
48. Eriksson, E. The Yearly Circulation of Chloride and Sulfur in Nature:
Meteorological, Geochemical, and Pedological Implications. Part II.
Tellus 12:63-109, 1960.
49. Slinn, W. G. N., L. Basse, B. B. Hicks, A. W. Hogan, D. Lai, P. S. Liss,
K. 0. Munnich, G. A. Sehmel and 0. Vittori. Some Aspects of the Transfer
of Atmosperhic Trace Constituents Past the Air-Sea Interface. Atmos.
Environ. 12(11):2055-2087, 1978.
50. Stracha, W. M. J., J. Hernault, W. M. Schertzer and F. C. Elder.
Organochlorines in Precipitation. Paper presented at the International
Symposium on the Analysis of Hydrocarbons and Halogenated Hydrocarbons,
Hamilton, Ontario, May 23-25, 1978.
1-76
-------
51. Jones, D. M. A. and D. D. Meredith. Great Lakes Hydrology by Months.
1946-65. Proc. 15th Conf. on Great Lakes Research, International
Association for Great Lakes Research, 1972. p. 477.
52. Atkins, D. H. F. and A. E. J. Eggleton. Studies of Atmospheric Wash-out
and Deposition of y~BHC, Dieldrin, and p-p DDT using Radio-Labelled
Pesticides. Nuclear Techniques in Environmental Pollution, International
Atomic Energy Agency, Vienna, Austria, 1971. pp. 521-523.
53. Cawse, P. A. A Survey of Atmospheric Trace Elements in the U.K. (1972-
73). United Kingdom Atomic Energy Authority, Harwell, England. AERE-
R7669, 1974. p. 78.
54. Adam, J. R. and R. G. Semonin. Collection Efficiency of Scavenging for
Submicron Particulates. Precipitation Scavenging—1970. AEC Symp.
Series 22, 1970.
55. Sood, S. K. and M. Z. Jackson. Scavenging Study of Snow and Ice
Crystals. Reports IITRC C6105-9, 1969 and 11TR1 C6105-18, 1972, ITT
Research Institute, Chicago, Illinois, 1972.
56. Kerker, M., J. Hampl, D. D. Cooke and E. Matijevic. Scavenging of
Aerosol Particles by a Falling Water Droplet. J. Atmos. Sci. 28:1211-
1221, 1974.
57. Gatz, D. F. Pollutant Aerosol Deposition into Southern Lake Michigan.
Water, Air, Soil Pollution 5:239-251, 1975.
58. Hollod, G. Personal Communication, 1978.
59. Gatz, D. F. and A. N. Dingle. Trace Substances in Rain Water:
Concentration Variations during Convective Rains, and Their
Interpretation. Tellus 23:14-27, 1971.
60. Chamberlain, A. C. Transport of Lycopodium Spores and Other Small
Particles to Rough Surfaces. Proc. R. Soc. London, Series A 296, 1966.
pp. 45-70.
61. Sehmel, G. A. and S. L. Sutter. Particle Deposition Rates on a Water
Surface as a Function of Particle Diameter and Air Velocity. J. Rech.
Atmos. 8:911-920, 1974.
62. Holler, U. and G. Schumann. Mechanisms of Transport from the Atmosphere
to the Earth's Surface. J. Geophys. Res. 75(15):3013-3019, 1970.
63. Cohen, Y., W. Cocchio and D. Mackay. Laboratory Study of Liquid-Phase
controlled Volatilization Rates in Presence of Wind Waves. Environ. Sci.
Technol. 12(5):553-558.
64. Elzerman, A. W. Surface Microlayer-Microcontaminant Interactions in
Freshwater Lakes. University of Wisconsin-Madison, Ph.D. Thesis, 1976.
65. Maclntyre, F. Flow Patterns in Breaking Bubbles. J. Geophys. Res.
77:4211-4228, 1972.
1-77
-------
66. Andren, A. W., A. W. Elzerman and D. E. Armstrong. Chemical and Physical
Aspects of Surface Organic Microlayers in Freshwater Lakes. J. of Great
Lakes Res. 2:101-110, 1976.
67. Baier, R. E., D. W. Goupil, S. Perlmutter and R. King. Dominant Chemical
Composition of Sea-Surface Films, Natural Slicks, and Foams. J. Rech.
Atmos. 8:571, 1974.
68. Torrey, M. S. Environmental Status of the Lake Michigan Region, Vol.
3. Chemistry of Lake Michigan. Argonne National Laboratory/ES-40, 1976.
pp. 366-367.
69. Wallace, G. T., Jr. and R. A. Duce. Concentration of Particulate Trace
Metals and Particulate Organic Carbon in Marine Surface Waters by a
Bubble Flotation Mechanism. Marine Chem. 3:157-181, 1975.
70. Bunce, N. J., Y. Kumar and B. C. Brownlee. An Assessment of the Impact
of Solar Degradation of Polychlorinated Biphenyls in the Aquatic
Environment. Chemosphere 2:155-164, 1978.
1-78
-------
APPENDIX I-A
XAD-2 CLEANUP PROCEDURE
1. Dry XAD-2 for 24 hr at 60°C.
2. Clean XAD-2 by soxhlet extraction using petroleum ether for 72 hr,
changing the solvent every 24 hr.
3. Dry XAD-2 overnight at 60°C.
4. Weigh out 70 g quantities of clean resin and store in glass jars
with foil lined caps.
SAMPLE CLEANUP PROCEDURE
1. Soxhlet extract XAD-2 for 24 hr with 500 ml of petroleum ether.
2. Concentrate XAD-2 extract to 10 ml using a Kuderna-Danish concen-
trator.
3. Dry extract through disposable pipette containing sodium sulfate.
4. Concentrate to 2 ml using a slow stream of purified nitrogen.
5. Transfer to 4 g alumina (6% w:w 1^0) column (1 cm ID) topped with
a small amount of sodium sulfate.
6. Drain column to top of sodium sulfate level collecting the eluate
in an appropriate receiving vessel.
7. Add more petroleum ether to collect PCB fraction.
8. Total volume collected should be 13 ml.
9. Concentrate to a suitable volume with a slow stream of purified
nitrogen.
10. To soxhlet extract filter, cut into strips small enough to fit
in the thimble of a 250 ml soxhlet extractor.
11. Soxhlet extract with petroleum ether for 24 hr.
12. Treat filter extract in the same manner as XAD-2 extract.
1-79
-------
PART II
PAHs IN AIR OVER LAKE MICHIGAN
J, W, STRAND
A, W, ANDREN
Il-i
-------
ABSTRACT
Aerosol sampling was conducted over Lake Michigan for a 3-yr period to
develop and adapt methods of analysis for measuring low concentrations of
polycyclic aromatic hydrocarbons (PAHs) and to determine the qualitative
nature and possible cycling of these compounds over the lake's surface.
Twelve PAHs were identified in the air over Lake Michigan. Their wet and dry
fluxes were approximately 105 kg/yr and relatively high levels were found for
two potent carcinogens—benz(a)anthracene and benz(a)pyrene. Neither bubble
ejection nor volatilization proved to be significant as pathways for PAH
return to the atmosphere. The atmosphere is postulated to act as a source and
the sediment as an eventual sink for these compounds.
Il-ii
-------
CONTENTS - PART II
Title Page II-i
Abstact Il-ii
Contents II-iii
Tables Il-iv
II-l. Introduction II-l
II-2. Conclusions II-2
II-3. Background II-3
Gas Chromatography of PAHs in Aerosols II-3
11-4. Experimental Methods II-7
Glass Fiber Filter Treatment II-7
Ship Aerosol Sampling 11-7
Gas Chromatographic Analysis II-9
GC Column Silanization and Operation II-9
II-5. Results and Discussion.. 11-13
PAH Concentrations and Fluxes to Lake Michigan 11-13
Bubble Ejection of PAHs from Lake Michigan
Microlayer Water 11-20
Volatilization of PAHs from Lake Michigan 11-21
References 11-25
II-iii
-------
TABLES
Number Page
II-l Chemical and physical properties and carcinogenicity of
PAHs 11-12
II-2 Abbreviations for polycyclic aromatic hydrocarbons 11-14
o
II-3 PAH concentrations (ng/m ) for Lake Michigan cruises 11-15
*\
II-4 Annual average ambient concentrations (ng/m ) of BaP. 11-14
II-5 PAH concentrations (ng/m ) for hi-volume size
fractionated samples 11-17
II-6 PAH concentrations (pg/L) for the Lake Michigan
microlayer samples 11-18
II-7 Dry flux of polycyclic aromatic hydrocarbons to
Lake Michigan 11-21
11-8 Wet flux of polycyclic aromatic hydrocarbons to
Lake Michigan 11-21
Il-iv
-------
II-l. INTRODUCTION
The increased use of fossil fuels as a source of energy has
resulted in an increase of organic combustion by-products in the air.
In heavily populated areas the concentration of these compounds
can reach undesirable levels (1). Although this pollution has been
recognized, little information exists on the composition and fate
of the organic fraction of aerosols. Once organic aerosols are
formed those physical, chemical and meteorological factors that can
influence their deposition from air to water, as well as other
pathways, have not been documented adequately and real values
to work with are meager or nonexistent. The total composition of
organic matter in aerosol particles has never been fully investigated
and no information, to date, exists on the organic fraction of
aerosols directly over Lake Michigan or the size fractionation
of polyaromatic hydrocarbons over Lake Michigan. This becomes
important in view of work done on aerosol trace element input and
the small drainage basin to surface area ratio of the lake.
The principal objective was to systematically develop and
adapt methods of sampling and analysis to measure low (ng/rn-^)
concentrations of 16 selected PAHs in aerosols directly over
Lake Michigan.
II-l
-------
II-2. CONCLUSIONS
Twelve polycyclic aromatic hydrocarbons were identified in
aerosols sampled directly over Lake Michigan. This constituted the
first measurements of this kind on a large inland lake. Concentra-
tions ranged from 0.1 to 4.2 ng/rn-^ which were found on single filters
exposed for <24 hr. These values correlated well with those from
shoreline data in the literature and added to the meager qualitative
and quantitative data obtained to date on PAHs in this unique
environment. The highest concentrations of these compounds were
found for the smallest PAH ring sizes, and may be an unexplained
lake effect on organic aerosols and vapor.
Concentrations of PAHs in the Lake Michigan microlayer ranged
from 0.15 to 0.45 yg/liter, representing on a relative scale,
10" times the concentration in air. This suggests that aerosols
are a source of these compounds and that the microlayer serves as
a repository until the PAHs are removed.
The wet and dry fluxes of individual polycyclic aromatics were
virtually the same and approximately an order of magnitude lower
than the total organic carbon flux to the lake. High levels of
two important aromatics, benz[a]anthracene and benz[a]pyrene,
emphasize the need for data on the atmospheric input of water
pollutants such as these known carcinogenic agents.
Neither bubble ejection nor volatilization of polycyclic
aromatics proved to be significant routes for the return of PAHs to
the atmosphere from the lake, compared to the inputs of wet and
dry deposition.
It appears that PAHs in aerosols originate from man-made
combustion processes. It is hypothesized that the controlling
mechanism is a gas-to-particle formation associated with particles
from 0.001 ym (molecular) to 1 ym (submicron). This conversion for
PAHs is made rapidly at ambient temperatures and pressures. By
the nature of the particle sizes PAHs are formed on, their residence
times are long and removal from air is principally by impaction.
Thus, the flux from air to water is slow but significant. The
atmosphere acts as a large source and the lake as a sink, with
little or no return. Once the PAHs reach the lake their build up
in the microlayer seems evident, until adsorption and sedimentation
to the lake bottom remove them from further cycling.
II-2
-------
II-3. BACKGROUND
Gas Chromatography of PAHs in Aerosols
Gas chromatography (GC) for the determination of PAHs in air
was used by Liberti et al. (2). The GC analysis of PAHs in other
matrices was reported previously. Extraction of urban aerosols was
achieved by dividing into cyclohexane, methanol-water and nitro-
methane fractions. This procedure allowed quantitative recovery of
PAHs in the nitromethane fraction and complete separation of
aliphatic and aromatic fractions. Subsequent separation of the
aromatic fraction was performed using a 35 m capillary column
coated with SE-30. Good separation of 19 PAHs was achieved.
However, isomer resolution was not completed. Analysis was carried
out with flame ionization detectors using 1,3,5 triphenylbenzene
as an internal standard.
Cantuti et al. (3) evaluated SE-30, SE-52 and XE-60 liquid
phases, temperature programming and the electron capture detector for
improved PAH analysis. The SE-52 and XE-60 columns provided mixture
separation of the difficult pairs benz[a]- and benz[e]pyrene, chrysene
and benz[a]anthracene. It was found that retention indices for
programmed runs were in agreement with isothermal values, indicating
programming could be used with reliability.
Other researchers have utilized the resolution of GC to analyze
PAHs in atmospheric samples (4,5).
Extensive evaluation of GC liquid phases was conducted with
PAH separation in mind. Apiezon L, silicon oil and polystyrene were
examined by Abraham and Marks (6) and JXR, OV-1, DC-200 and
Dexsil 300 were compared by Birchfield et al. (7). Of these phases,
Dexsil 300—a borane-silicone copolymer—showed best thermal stability,
but was incapable of resolving certain refractory pairs such as
benz[a]- from benz[e]pyrene. Others have found Dexsil 300 to be
a superior liquid phase. For example, Natusch (8) used 1.5% Dexsil
300 on 40/60 mesh Chromosorb W coupled with temperature programming
and flame ionization, which yielded a clean separation of 20 to 30
PAHs. Kubota et al. (9) used a 4.6 m x 3.2 mm glass column of
3% Dexsil 300 on 80/100 mesh Chromosorb W(HP), with programming
from 110 to 320°C, to elute 2 through 7 ring systems.
Because of the interest in PAHs as carcinogens, there has been
a proliferation of publications on their occurrence, distribution
and methodology.
II-3
-------
The chemical characteristics of the organic matter found in
air which is extracted by benzene in a soxhlet procedure have been
examined by infrared spectra, empirical formula, functional groups
and particle sizes (10,11). Variation of organic chemical content
with size was shown to occur. This finding is important because
it delineates the kinds of substances which will enter the human
lung.
Grimmer and Bohnke (12) used capillary column gas chromatography
to provide high resolution of over 100 PAHs in airborne dust
collected at Bochum and Marz, Federal Republic of Germany. The
acetone extract was partitioned successively between cyclohexane,
dimethylformamide and cyclohexane again, to give an enriched aromatic
fraction. Clean up of the complex mixture was facilitated with
silica and Sephadex LH-20 column chromatography prior to GC.
Analysis of fourteen PAHs in a large area of Los Angeles County
showed a correlation with inversion height (13). The higher the
median inversion height the lower the concentration of benz[aJpyrene,
or the larger the volume of air a pollutant has to disperse in,
the greater its dilution. The major source of benz[a]pyrene was
traffic with an average emission value of 0.43 ng/nr' (14,15).
Until recently, the lack of an effective separation scheme
was the most serious hindrance to the understanding of airborne
PAHs. Schemes"typically entailed liquid-liquid extraction, column-,
thin layer- (TLC) or paper-chromatography (PC) followed by
luminescence or absorption spectroscopy. These techniques met
with varying degrees of success due to serious limitations, including
loss of sample, tailing and loss of reproducibility.
It has been recognized that small differences in the structure
of PAHs cause great differences in their tumorgenicity. To remedy
this problem emphasis has been focused on high resolution columns.
Zoccolillo (16) developed a high efficiency packed column of
1% SE-52 on 60/80 mesh Chromasorb G that was comparable to a 35 m
capillary column with respect to resolution and time of analysis.
This was accomplished by selecting a solid support of low surface
area and minimizing the liquid phase coating. The column was
applied successfully to routine analysis of PAHs in airborne
particulates in Rome, Italy.
Doran and McTaggart (17) utilized glass capillary columns for
the routine analysis of benz[aJpyrene and benz[a]anthracene. High
temperature stationary phases such as the OV-series and the Dexsil
series were used to allow temperature programming with minimum
bleed.
While suspected previously, the complexity of airborne PAHs
has been made more clear with the use of high resolution capillary
columns. Lao et al. (18) and Lee et al. (19) reported >100 PAHs
in extracted airborne matter using capillary columns. A systematic
extraction scheme for airborne PAHs was described by Novotny
II-4
-------
et al. (20) involving several partitioning steps, lipophilic gel
fractionation, high resolution liquid chromatography (LC) on
chemically-bonded stationary phases, followed by capillary column
GC.
Using the sensitive flame ionization detector (FID) and high
resolution capillary columns, unique profiles of polyaromatic
compounds were obtained (21). This technique was used to screen
sources of atmospheric pollution.
Lao et al. (18) was the first to evaluate the potential of
combined GC-mass spectrometry (MS) for PAH analysis in airborne
pollutants. Mass spectra of more than 70 major PAHs in air were
obtained and compared to reference standards. It was found that MS
could be used as a sensitive detector for GC (22,23,24).
Lao et al. (25) extended their GC-MS work to a quadrupole
system for qualitative analysis and micro-processor controlled GC-FID
for quantitative analysis of PAHs. Comparisons of Dexsil 300, 400
and 410 packed columns showed small differences in retention times
or response factors.
Fourier transform nuclear magnetic resonance spectroscopy (NMR),
GC and MS were used to characterize PAHs and alkylated PAHs in air
pollution particulate matter over Gary, Indiana (19). Emphasis
was placed on hitherto unresolved toxicologically important isomers
by using glass capillary columns and NMR to locate positional isomers.
It was found that Gary gave a distinct "finger print" of PAHs and
concentrations that could be used as a tracer.
Recent publications on high pressure liquid chromatography
indicate separations of airborne PAHs comparable to GC (26,27).
It has been demonstrated that gas-liquid chromatography (28)
gave baseline separations of isomeric 3 to 5 ring PAHs using a
nematic liquid crystal N, W-bis (p-methoxybenzylidene)a,
a'-bi-p-toluidene (BMBT). This liquid phase also yielded unique
resolution of steroid epimers (29), polychlorinated biphenyls (30)
and methoxybenzanthracene isomers (31). Separations were superior
to any published to date on packed or capillary columns, however,
BMBT showed bleed at elevated temperatures and prolonged operation.
To remedy this, a new liquid crystal N, N'-b±s (p-butoxybenzylidene)a,
a'-bi-p-toluidene (BBBT) was synthesized and shown to have diminished
bleed with the same separation (32). It was found, however, that
analysis of high molecular weight PAHs (22 to 24 carbon atoms) on
BMBT or BBBT resulted in prolonged retention times and
peak broadening. A new high temperature liquid crystal
N, tf'-bis (p-phenylbenzylidene)a, a'-bi-p-toluidene (BPhBT),
compatible with GC-MS and extending the analysis of PAHs to 5 to 7
ring systems was synthesized (33). With BBBT and BPhBT, PAHs from
2 to 7 rings can be separated and analyzed by packed column chroma-
tography. Baseline separations of 22 to 24 carbon PAHs were
performed in <10 min with small injection volumes.
II-5
-------
Airborne particulate matter from an urban area was fractionated
and analyzed on a 5 m, 4% Dexsil 300 column and Finnigan 3100
mass spectrometer. Interpretation and comparison with standards
led to identification of more than 100 compounds (34).
Recent determinations of PAHs in the environment were achieved
with glass capillary GC on a 20 m column of SE-52 (35). The method
was applied to samples of recent lake and river sediments, river
particulates, street,dust and airborne particles.
A unique separation of airborne PAHs was achieved using a
2 hr Soxhlet extraction. The extract was concentrated without
separation and analyzed by GC using a high performance ultra-thin,
thermally treated Carbowax 20M column (36). In the past, use of
composite extracts from as many as 300 high volume filters was
reported. However, for the first time they were capable of analyzing
a single 24 hr filter (36). A rapid procedure for the analysis of
nearly 100 organic compounds was developed, allowing a series of
50 comparative survey analyses from a single site in <8 hr/sample.
For quantitative GC analysis of PAHs, pentatricontane (035) has
been used as an internal standard (37).
Pierce and Katz (38) examined the size dependence of PAHs in air.
Using Anderson cascade impactors, eight PAHs and two oxygenated
arenes were separated by TLC and analyzed by absorption and fluor-
escence spectrophotometry. The size distribution of PAHs followed
a log normal relationship, with the majority of PAHs associated with
particles <3 ym diameter. Little other information on particle or
mass distribution of PAHs in air is known. DeMaio and Corn (39)
showed that >75% of the weight of selected PAHs was associated with
particles <2.5 urn in diameter. Kertesz-Saringer et al. (40) reported
that from 70 to 90% of the total benz[a]pyrene was associated with
aerosols with a radius of <1 Urn. Thomas (41), however,
showed that the amount of benz[a]pyrene per unit weight of soot
obtained from the combustion of various fuels was constant fo*r all
particle size fractions. Other studies (38) showed that photo-
oxidation by-products of PAHs also are size dependent.
Only one paper has been published on PAHs in airborne matter
over a large body of water (42). Eighteen different PAHs with from
2 to 6 ring structures, were collected over the North Atlantic Ocean.
The ether soluble fraction of these aerosols was separted according
to a scheme yielding organic acids, phenols, organic bases, ali-
phatics, aromatics and neutral polar compounds (43).
Work on the total organic carbon content of aerosols over a
large (marine) body of water showed a decrease in the amount of
organic matter with increasing salt content (44,45).
II-6
-------
II-4. EXPERIMENTAL METHODS
Glass Fiber Filter Treatment
Glass fiber filters have a basic pH when untreated. This
poses a problem when used in weight measurements. Gases, such as
sulfur dioxide, can be adsorbed on the surface of the glass fiber
and, at a basic pH be oxidized to sulfate (46). To avoid this
weight anomoly, the glass filters were soaked in 6N f^SO^ prepared
from XAD-2-extracted water. To lower their organic blank, all
glass filters were then extracted with ether and combusted for 3 hr
at 450°C (43,47).
Glass fiber filters are hygroscopic to a certain extent although
less than other filter media. In order to minimize weight changes
with fluctuations in relative humidity (RH) or temperature the
filters were weighed before and after collection at constant
temperature and relative humidity (48).
Glass fiber filters were placed in the chamber, equilibrated
for 1 hr, and their mass determined by weighing on either an
analytical balance (±0.0001 g) or an electromagnetic ultratnicrobalance
(±0.0001 mg). The difference before and after collection yielded
particle mass.
Ship Aerosol Sampling
Sampling was conducted aboard several ships during a 3-yr
period and once from land on the shoreline of Lake Michigan. The
ships utilized were the University of Michigan's R/V Laurentian,
the University of Wisconsin Systems the Neeskay and the R/V Roger R.
Simons of the Environmental Protection Agency. Sampling was
conducted during the spring, summer and fall, although not at all
these times each year. Transects for the cruises were both north
and south of Milwaukee and stationary samplings in connection with
cruises under the direction of Dr. Herman Sievering of Governors
State College were also carried out.
Equipment including high volume aerosol samplers, high volume
cascade impactors and Delron cascade impactors were run in duplicate
and located on the bow of each ship ahead of smoke stacks, ventila-
tion systems and from 2 to 7.6 m above the water line. These
samplers were positioned such that they were as far apart as
possible. Precautions were taken to secure all samplers with rope,
tape and foul weather proofing which was water proof polyethylene
bags.
II-7
-------
It was found that inorganic contamination of filters was
possible from air recirculated by the pimp. The source of this
inorganic copper in filters is wear on the armature of the pump
(49). A study of the production of carbon aerosol by high
volume samplers from abrasive wear of motor brushes yielded
average concentrations of 150 ± 10 yg/m^ (50). This is an
appreciable value compared to the range of 1.7 to 19.8 yg/nH
found over Lake Michigan (this work). Recirculation was found
to be very low, i.e., <0.1%, and could be ignored in monitoring
environmental samples. Caution should be exercised in operating
two or more samplers in that they should have ample spacing or
exhaust tubes to preclude interference.
In a recent report (51) volatile polychlorinated naphthalenes
(PCNs) were found in the plastic parts of motor assemblies of
certain high volume samplers and contaminated the filters for
subsequent analysis. To avoid recirculation of PCNs, a collapsible
exhaust duct 3 m in length was connected to the housing of each
high volume sampler. This allowed the exhaust to be channeled away
from the intake (52). An evaluation of plastic components in our
samplers has shown no PCNs to be present. However, exhaust tubing
was mounted as a precaution.
To avoid .contamination of the filters from the smokestack of
the ship while sampling, the high volume samplers and Delron samplers
were turned off whenever the wind came within their arc. While
underway there was little problem with smokestack contamination.
However, when on station, samplers were often turned off.
Flows for the Delron cascade impactors were corrected by
calibration in the laboratory against a mercury manometer. Flow
rates for the high volume samplers and high volume cascade samplers
were calibrated in the laboratory before each cruise with a top
loading orifice and water manometer.
It has been shown that some PAHs are sufficiently volatile to
evaporate during collection if prolonged sampling procedures and
slow flow rates are used. This is especially true for the lower-
molecular-weight hydrocarbons, that is, it is not possible to
collect anthrancene, phenanthrene, pyrene or fluoranthene efficiently.
It is evident that accuracy requires collecting samples at high
flow rates during short periods (about 24 to 48 hr) and analyzing
them in the laboratory as soon as possible. Samples were stored
in clean aluminum pouches, in the dark at 4°C. The organic fraction
of samples remains stable for up to 5 yr if treated in this way (53).
In a study on the stability of PAHs collected on glass fiber filters,
no tetracyclic and larger PAHs were lost with a 2-hr collection at a
flow rate of 1.2 m3/min followed by 2-hr of filtered air at 1.2
m-Vmin (54). The same study showed some tetracyclic PAH loss with
24 hr filtration.
II-8
-------
Gas Chroraatographic Analysis
A Hewlett-Packard 5730-A gas chromatograph equipped with a dual flame
ionization detector and a linear temperature programmer was used.
Chromatograms were generated on a 1 mv full scale, 1 cm/min chart speed linear
electronic strip chart recorder fitted with an electronic integrator.
Nitrogen carrier gas flow was regulated by calibrated Brooks mass flow
controllers. Air flow rate was maintained at 240 ml/min and hydrogen at 39
ml/min to optimize FID response (55). In-line screens and molecular sieve
traps were placed on all gas lines. The carrier gas line also was fitted with
an Oxisorb cartridge (Altech Association). Hydrogen and nitrogen were of
>99.99% purity. High temperature, low bleed HT-10 septa (Applied Science)
were used with on-column injection. Sample volumes ranging from 1 to 3 pL
were introduced using a Hamilton 701-N syringe.
GC Column Silanization and Operation
Before packing a gas chromatographic column, a water aspirator was
attached to one end by means of a disposable pipette and tubing. Spectrograde
chloroform and acetone (300 ml) were drawn through the columns to remove oil
and grease. A 10% solution of dimethyldichlorosilane (DMCS, Supelco) in glass
distilled toluene was added and allowed to stand for 10 min. The DMCS was
removed, the column rinsed with toluene, filled with methanol and allowed to
stand for 10 min. The methanol was removed by suction and the column rinsed
with fresh methanol followed by acetone. The reactions for deactivating
(silanizing) the surface sites on glass are:
r
1. J... + (CH,).
-HC1 CH3OH 3
.0Cl0Si 5» CH_-Si-CHQ 3> CH,-Si-CH, + HC1
Glass 3/2 3 I 3 3 J
H OH -2HCL
2. I | + (CH3)2Cl2Si
Graphitized ferrules were used on the inlet and outlet ends of each column.
Temperatures reported were those on instrument dials. Baseline drift and
column bleed were minimized by using a differential electrometer. The packing
(3% Dexsil 300 on 60/80 mesh Chromosorb 750, Analabs Inc.) was placed in a
stainless steel drying tube, pressurized with 40 psi N2 and blown into 4.6 m
columns. Air in the columns was flushed out for 30 min with No (60 psi
maximum flow), followed by conditioning from 70 to 320°C at 4°C/min overnight
under N2 flow. The packing 2.5% BBBT on 100/120 mesh Chromosorb W-HP was
prepared from a chloroform solvent slurry and packed in 1.3 m columns. The
II-9
-------
columns were conditioned from 190 to 270°C at l°C/min overnight under maximum
N2 flow. The packing 2.5% BPhBT on 100/120 mesh Chromosorb W-HP was prepared
in a boiling dimethyl formamide (DMF) solvent slurry and packed in 1.85 m
columns. The columns were conditioned from 160 to 270°C at l°C/min overnight
under maximum N£ flow. A 1 cm plug of Chromosorb W coated with 4% SE-30 was
packed on the column inlet side to retard deterioration. The packing—Ultra-
Bond Carbowax-20 M on 100/120 mesh Chromosorb W (RFR Corporation)—was packed
into a 1.85 m column. The- column was conditioned under maximum No flow, at
200°C overnight and then 240°C for 2 hr.
Several problems occurred in the packing and operation of the Dexsil 300
columns. Extensive bleed and packing separation occurred due to bad batches
of the liquid phase supplied by Analabs. New packing was ordered. A large
pressure drop was found across the 4.6 m 80/100 mesh columns. During
temperature programming, the flow rate dropped to 10 ml/min, causing extensive
baseline drift. A change from 80/100 mesh to 60/80 mesh solved this
problem. Although the solid support caused no contamination, the highest
purity support—Chromosorb 750—was used for the larger mesh size. As a
result, 3% Dexsil 300 on 60/80 mesh Chromosorb 750 provided no bleed and no
flow decrease during temperature programming.
The stringent requirements of resolution, stationary phase thermal
stability and detector sensitivity demanded by the presence of trace levels of
closely related, nonvolatile PAHs over Lake Michigan challenge the analytical
capabilities of GC. Only the most sensitive GC detector was used, the flame
ionization detector (FID). The advantage of the Hewlett-Packard FID was its
wide dynamic range (10 ) and sensitivity (5 x 10" amps full scale). The FID
signal was proportional to the .number of carbon atoms in similar oxidation
states, providing a quantitatively comparable detector response for most PAHs.
The "solvent flush" injection technique was used to insure that all
analytes entered the column. The method involves drawing a plug of solvent
into the syringe then the analytical solution. The solvent behind the
solution washes the barrel clean, sweeping all material onto the GC column.
Linear PAHs are more readily oxidized than non-linear (angular) PAHs.
Therefore, PAHs such as tetracene usually are not found in the atmosphere,
although angular compounds such as benz[a]anthracene and chrysene are. The
stability of PAHs suspended in the atmosphere depends on molecular structure,
the amount of available light and the presence of oxidizing pollutants. For
example, half-lives of <1 day to several days have been given for
benz[a]pyrene on soot in the presence of sunlight (1). A PAH within the soot
probably has a much longer half-life. Smaller PAHs would disappear in
sunlight more rapidly than larger ones and summer temperatures would
accelerate this loss. The PAHs absorb light between 350 and 450 nm and
photooxidize at these wavelengths (1). In order to minimize sample or
standard losses, yellow fluorescent lights were installed in the laboratory.
In analyzing PAHs, factors affecting the reproducibility of analysis
are: the solvent, adsorbent, laboratory environment and quality of reference
compounds. Errors related to these factors cannot be overcome with internal
standards and can be avoided only by purifying solvents and working in a
clean, ventilated laboratory.
11-10
-------
Sixteen PAHs were chosen as standards based on those PAHs most
prevalently found in aerosols and known to be carcinogenic.
Two solvents—cyclohexane and glass distilled methylene chloride—were
used for disolution of PAH standards. It was found that the PAHs were more
soluble in methylene chloride than in cyclohexane at the high concentrations
(100 ppm) utilized. The high dielectric constant of methylene chloride (8.93)
allowed immediate solvation of PAHs whereas several hours at room temperature
were required for PAH solution in cyclohexane with a dielectric constant of
2.22.
The PAHs used for qualitative and quantitative analysis were fluorene,
anthracene, phenanthrane, fluoranthene, pyrene, 2,3-benzofluorene,
triphenylene, benz[a]anthracene, chrysene, benz[a]pyrene, perylene, picene,
1,2,4,5-dibenzpyrene, o-phenylene pyrene, benzofghi]perylene and coronene.
Each compound was weighed (2.5 mg) and dissolved in 25 ml of solvent.
Standards of 10 and 1 ppm were made by diluting 1 ml of 100 ppm stock solution
to a 10 ml volume and 1 ml of 10 ppm standard to a 10 ml volume. The
chemical, physical and biological properties of the PAH standards are listed
in Table II-l.
During analysis, organic solvents are evaporated to concentrate the
PAHs. The loss of PAHs during this procedure was reduced or avoided by
working at pressures above 12 mm Hg and at a water bath temperature <45°C.
Contamination and systematic errors were recognized and corrected by including
blanks with standards and samples (56).
No universally acceptable quantitative values are available for the
precision and accuracy of measurement techniques for PAHs. Accuracy data
reported in the literature usually are based on the recovery of added material
or the recovery of an added radioactive standard bound in the ambient
atmospheric matrix, exactly as the compound of interest, still remains
unanswered. The accuracy of PAH analysis depends on collection techniques as
well as on analytical methodology. To date no standard reference materials
are available for trace level PAHs in natural matrices, thus precluding
absolute PAH analysis. It is more meaningful to talk about relative precision
and sensitivity. For PAH analysis a sensitivity of ± 0.1 ng/m and a relative
precision of ± 25% are considered reasonable (54).
11-11
-------
Table II-l. Chemical and physical properties and carcinogenicity* of PAHs
(1,57,58)
Compound ,
formula
Fluorene
C13H10
Anthracene
C14H10
Phenanthrene
C4H10
Fluoranthene
C16H10
Pyrene
C16H10
2,3-Benzo-fluorene
C17H12
Triphenylene
C18H12
Benz [a ]anthracene
C18H12
Chrysene
C18H12
Benz[a]pyrene
C20H12
Perylene
C20H12
Picene
1,2,4, 5-Dibenzpyrene
C24H14
O-phenylene-pyrere
C22H12
Benzo(ghi)perylena
C22H12
Coronene
C24H12
Molecular
Structure weight
CCQ 166.23
££0 178.24
£*£) 178.24
v9 202.26
JCO 202.26
C£CQ 216.29
jOO 228.30
GOU 228.30
QQ^ 228.30
OOU 252.32
88 232'32
^"v._^"v _^\ 278»36
\~f\~f\-S
COCk 302.38
COjU 276.34
&\J 276.34
Q!Q 300.36
Melting
point, °C
116-117
216.2-216.4
101
110
156
t
199
159-161
255-256
176.5-179.3
273-274
367-369
233-234
162.5-164
?
438-440
Boiling
point, °C
293-295
340
340
393
393
t
425
435
448
475
500
518-520
!
,
525
*Benz[a]pyrene is strongly carcinogenic; benz[a]anthracene and o-phenylene-pyrene are
weak carcinogens; chrysene, triphenylene, picene and 1,2,4,5-dibenzpyrene have
uncertain or unknown carcinogenicity; the remainder are not carcinogenic.
11-12
-------
II-5. RESULTS AND DISCUSSION
PAH Concentrations and Fluxes to Lake Michigan
The environmental role of PAHs in air is of particular concern because
they pose a potential health hazard to man. Although this work does not
address itself directly to the health aspects of PAHs, it does so indirectly
by focusing on sources and types of PAHs emitted, physical and chemical
characteristics in the atmosphere, and methods of collection, separation,
detection, and quantitation. PAHs are found in small detectable amounts in
air, water, and soil. PAH solubility determines the specific aquatic
interactions; biodegradation controls the soil interactions; and association
with particulate matter dictates the prevailing interaction in air. Many—but
not all—PAHs can be attributed to man's activities. However, those that are
formed by man originate exclusively from combustion processes. Emitted vapor
cools and condenses on particles already prsent or forms small particles of
pure condensate. The physical properties that influence PAHs include particle
size, surface area, shape and density for which little information exists.
The PAHs are subject to the same processes as airborne particles. They are
dispersed by turbulence, transported by wind, and removed from the atmosphere
by sedimentation, impaction, washout and rainout.
Most previous PAH measurements are for benz[a]pyrene alone. For this
reason a broad range of three to seven ring PAHs was examined.
PAHs are either cycled through or stored in the environment. Pathways of
cylcing were explored to determine if the lake acts as a sink for aerosol
depsotion and where their concentrations are highest.
Table H-2 lists abbreviations for the PAHs found in aerosols over Lake
Michigan and the PAHs found in microlayer water. These PAHs were chosen
because of their broad range of ring sizes, sources, synerglsm and biological
activity.
o
Measured values for single filters ranged from 0.1 ng/m for several
compounds to 4.2 ng/m for pyrene sampled in late October, 1975 (Table II-
3). Average annual airborne BaP concentrations from 1966 to 1970 in urban
areas of the United States National Air Sampling Network ranged from a low of
0.2 ng/m3 over Hawaii to a high of 29.5 ng/m at Altoona, Pennsylvania (54).
The same 5 yr study also produced nonurban annual average concentrations of
BaP which were generally an order of magnitude less than those for the urban
concentrations, i.e., 0.1 to 2.1 ng/m (59). These ranges fit well with
overall individual PAH compounds and the values found for BaP over Lake
Michigan, 0.3 to 1.8 ng/m . It is reasonable to believe that southern Lake
Michigan might lie somewhere between an urban and nonurban range, with urban
generated aerosols being transported, diluted and deposited to the lake. The
BaP values for cities located on Lake Michigan are listed for comparative
purposes (Table II-4). If the urban-industrial area on the western shore of
11-13
-------
Table II-2. Abbreviations for polycyclic
aromatic hydrocarbons
Abbreviation
Chemical name
Fl
Ph
An
Flu
Blf
Py
BaA
Per
Tri
BaP
0-Pp
B[ghi]P
Fluorene
Phenanthrene
Anthracene
Fluoranthene
2,3 Benzofluorene
Pyrene
Benz[a]anthracene
Perylene
Triphenylene
Benz[a]pyrene
0-phenylenepyrene
Benz[ghi]perylene
o
Table II-4. Annual average ambient concentrations (ng/m°) of BaP
Station
Chicago
East Chicago
Hammond
Grand Rapids
1966
3.3
6.8
3.9
ND
1967
3.0
5.7
2.5
ND
1968
3.1
1.9
2.1
1.4
1969
3.9
6.8
3.3
1.6
1970
2.0
5.3
1.7
0.8
ND is not determined.
11-14
-------
o
Table II-3. PAH concentrations (ng/m ) for Lake Michigan cruises
Compound
Fl
Ph
An
H Flu
M Blf
t_n
Py
BaA
Per
Tri
BaP
0-Pp
B[ghi]P
II
0.51
—
—
1.7
4.1
4.2
0.8
—
—
—
—
2.0
III
0.4
0.2
0.1
—
0.2
0.1
0.2
0.1
0.1
—
—
—
Ill
0.2
0.1
0.1
0.2
0.1
0.1
—
0.1
0.1
—
—
—
IV V
1.2
0.8
0.8
0.8 1.3
2.5
3.4
0.9
1.7
0.5
1.8
—
1.5
Cruise number
V VI VIII
2.2 0.6
0.8 — 0.8
0.8 — 0.8
1.2 — 1.3
1.5 — 1.3
1.8 0.2 0.4
2. 5 0. 3
—
0.4
—
0. 9
— — —
VIII
1.1
0.4
0.4
0.9
—
0.2
0.2
—
0.2
—
—
—
XI
2.2
1.0
1.0
1.2
1.1
0.5
—
—
0.3
—
—
—
XI XII
1.3
0.5
0.5
0.8 0.9
0.9
0.4
2.2 0.4
—
—
0.3
—
— —
*Relative precision ± 25% ranging from 1.05 to 0.03 ng/m3.
-------
the lake was the principal BaP source, the levels shown in Table II-4 could
produce those levels found over the lake. No PAH values were found in the
literature for on-lake sampling, and no other PAH values besides BaP were
found for shoreline comparison.
From samples collected on cruise number III which was conducted in the
northern part of the lake, it appears that definite lower concentrations of
PAHs exist compared to the southern portion of the lake. This fact correlates
well with total suspended particulate matter (TSP) and total organic carbon
(TOC) values which showed similar north to south gradients. Concentrations of
individual PAHs, collected consecutively, showed similar values. The sources,
meteorology and precision of sampling apparently were stable in order to
provide this data. A similar trend was seen for samples run simultaneously on
cruise number VIII.
Three to six-membered ring systems were detected over Lake Michigan with
the bulk of those being contained in 2 to 4-membered ring. It is not clear
whether this reflects a natural environmental portioning over the lake or an
analytical bias. Small rings elute from a GC column first, yielding sharper
peaks. Since the flame ionization detector responds to mass flow rate
(dm/dt), the narrower the band the greater the sensitivity, favoring the
detection of these low molecular weight compounds.
The major mode of PAH removal from the atmosphere is through chemical
reactions (59). These include photooxidation, and reaction with oxidants,
nitrogen oxides and sulfur oxides. Reactions of the large PAH ring systems
adsorbed on the surface of soot are rapid and sensitive to electrophilic
substitution and oxidation (1). Evidence exists that destruction of aromatics
on soot is accelerated by photochemical smog (53) and that SOn, SOo and H^SO,
react rapidly with PAHs on aerosols (1). Ozone is also known to react with
PAHs by oxidation of a double bond, the nuclear ring, or side chains (60).
All of these compounds are polar, thus water-soluble, and would not be found
in an aromatic enriched fraction. If the larger PAHs are selectively oxidized
and removed as polar organics in an analytical workup scheme, this might
explain the higher concentrations of the smaller rings found over Lake
Michigan. Chemical half-lives may be only hours or days under intense
sunlight, whereas in dry unpolluted atmospheric conditions, residence times of
particles <5 pm in diameter may exceed 100 hr (61).
These arguments are contrary to what would be expected for collection of
PAHs on filters, in that smaller rings are more volatile and the collection
process less efficient. Pupp et al. (62) reported losses of volatile PAHs
from filters for high equilibrium vapor concentrations and sublimation. If
the collection methods do not recover all ambient PAHs, then the atmospheric
concentrations would be lower than the true value. It is interesting to note
that one of the most volatile PCBs—Aroclor 1242—was found to associate in
highest concentration with particulate matter over Lake Michigan versus the
vapor phases (63). This fact and the higher levels of small ringed PAHs may
be an unexpected—and as yet unexplained—lake effect on organic aerosols and
vapor.
Size fractionated PAH sample values are presented in Table II-5. For the
May and June cruises similar compounds and similar concentrations were
observed on the same stage. No trend was seen for increasing concentrations
of PAHs on small-sized particles. However, the limited number of samples
11-16
-------
Table II-5. PAH concentrations (ng/m ) for hi-volume size fractionated
samples
Roger Simons
May 17, 1977
Roger Simons
June 11, 1977
Roger Simons
Aug. 14, 1977*
Stage 1
Stage 2
Stage 3
Stage 4
Stage 5
Backup stage
Fl
Blf
Py
BaA
An
Ph
BaA
Fl
An
Ph
An
Ph
Flu
Blf
Py
BaA
An
Ph
Flu
BaA
Fl
An
Ph
Flu
BaA
2
0
3
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
.4
.7
.0
.3
.5
.5
.2
.0
.3
.3
.8
.8
.9
.5
.3
.4
.5
.5
.4
.2
.1
.4
.4
.2
.2
Fl
Flu
Fl
An
Ph
Flu
Blf
Fl
Fl
An
Ph
Flu
BaA
Fl
Py
Fl
An
Ph
Flu
0
0
0
0
0
0
0
1
0
0
0
0
0
1
0
1
0
0
0
.5
.1
.5 Fl 0.04
.3 Tri 0.02
.3
.1
.2
.2
.2
.15
.15
.9
.2
.3 Blf 0.3
.1
»
.4
.3
.3
.3
*Combination of nine filters.
11-17
-------
obtained precludes any conclusions on size fractionation. Approximately 85 to
90% of the PAH content of sized aerosols collected in Toronto were associated
with particles <5 ym in diameter for the same size during the summer period
(38). It was evident that the dependency of PAH content on size varied
significantly with sampling locations, seasons and emission sources.
Particles having a diameter <5 ym in the respirable range are most likely to
be deposited in the pulmonary region of the respiratory tract. These values
agree with those of De Maio and Corn (39) who found particles <5 ym contained
more than 75% of the weight of selected PAHs. More recently, Kertes-Saringer
(40) reported that from 70 to 90% of the total BaP was associated with
aerosols of <1 ym in Bucharest, Hungary.
The concentrations of PAHs in Lake Michigan microlayer water are given in
Table II-6. Values range from 0.15 to 0.45 yg/L which are similar to the
range (0.05 to 3 yg/L) found for surface waters (64). Fewer PAHs were
detected in the microlayer compared to the atmospheric samples, and no PAHs
were seen in a 5 L bulk water sample with a 2 ng absolute limit of
detection. These concentrations represent—on a relative scale—values 10
times greater than the concentrations of PAHs found in air. This might
suggest an atmospheric route such as deposition to the lake surface as a
source of high PAH levels. Garrett (65) observed organic microlayers are
contained in a surface monomolecular slick which was roughly 2 x 10 ym
thick. In a calculation of the number of monolayers that the mid-lake total
organic matter of 4.3 yg/m3 would yield, Andren et al. (66) determined that
nearly 1000 layers would occur for this value. It was further stated that
this calculation suggests the importance of atmospheric deposition. However,
the proteinaceous nature of the microlayer also suggests that the lake itself
is a source of TOM. From the scanning electron microscope results, it is
evident that anthropogenic aerosol input occurs to Lake Michigan. Glassy
sphericals found in the air were similar to those found in Lake Michigan
microlayer samples (67). The similarity of these particles indicates that
they are formed from similar processes such as coal combustion. Annual
consumption of coal around southern Lake Michigan is approximately 21 x 10
Tonnes (68), thus fly ash is probably an important source of particles in the
microlayer.
Table II-6. PAH concentrations (yg/L) for the Lake Michigan microlayer
samples*
Cruise
Roger Simons June 8, 1977
Roger Simons June 19, 1977
Sample 1
Sample 2
Fl 0.35
An 0.15
Ph 0.15
Flu 0.35
An 0.45
Ph 0.45
Flu 0.3
*A bulk water sample of 5 L collected on August 3, 1976 showed no PAHs,
Relative precision of ± 25%.
11-18
-------
The presence of PAHs in the microlayer cannot be attributed solely to
atmospheric input. Certain bacteria and plants produce PAHs as biological bi-
products in their normal development (59), e.g., the alga CKlovella vulgav-is
is known to synthesize several PAHs (69). The presence of these tiny plants
and animals that live in the microlayer (neuston) could be the main source of
high PAH concentrations. Strains of Chlovella have been bred that are 80% oil
by weight (70). Once the PAHs have been synthesized, -in situ, they are likely
to remain in the microlayer. The proteinaceous wet surfactants in the
microlayer which are known to be glycoproteins and proteoglycons (70),
effectively trap organic molecules and, due to their hydrophobic nature, might
retain PAHs at the air-water interface. Another mechanism for the microlayer
enrichment of PAHs might be rising bubbles which are capable of transporting a
substantial portion of particulate matter to the surface (71). As wind-
induced bubbles move toward the surface, aasorptive scavenging occurs on
surface-active particulate organic carbon. Transport rates for particulate
organic carbon in the open ocean have proved to be rapid, with the bulk of the
collection occurring in the first minute (71). Bubble flotation as a
mechanism for PAH concentration in the microlayer may not be a valid
argument. Samples were collected under stable atmospheric and lake
conditions, minimizing bubble formation.
The appearance of the low molecular weight PAHs may be a reflection of
solubility and extraction efficiency. Acheson et al. (72) investigated
factors affecting the extraction and analysis of PAHs in water. These
included the initial concentration of the PAHs, the presence of suspended
solids and prolonged storage of the sample prior to analysis. In general
extraction efficiency of pyrene (a lower molecular weight PAH) and
benzo[ghi]perylene (a higher molecular weight PAH) were diminished if initial
concentrations were low, suspended solids high, and storage prolonged.
Suspended solids, however, had a greater effect on benzo[ghi]perylene than on
pyrene. This may be due to the higher solubility of small ring systems such
as pyrene in water and their more efficient recovery from water by an XAD-2
column. No quantitative information was given on the relative degree of ring
system adsorption on the glass wall of the collection flask or chemical or
biological degradation. However, it was recommended that extraction should be
carried out as soon as possible after sample collection. Because the
solubilities of most PAHs in water are extremely low, PAHs often have been
assumed to exist entirely on particles. Studies on the partitioning of 1^C-
anthracene between dissolved and particulate phases in natural waters showed
that a significant fraction—35 to 85%—is in solution (73). The
fractionation also was found to be highly dependent on yeast cell
concentration (used as a model for suspended organic matter) and suspended
organic solids encountered in natural waters but not as dependent on
adsorption to mineral surfaces. This soluble fraction of the low molecular
weight anthracene may explain the frequency of detection of this compound
(Table II-3).
The inconclusive results of size-fractionated PAHs for this work led to
the choice of a particle size of 1 ym as the most representative value. This
was done on the basis of Kertesz-Saringer's (40) investigation of the total
BaP concentration and the aerosol size it was associated with. This value is
probably a liberal estimate since BaP was correlated to particles <1 pm.
Deposition velocities (74) are almost the same for particles between 0.1 and
1 pm, thus PAH deposition should not be unduly distorted by small particles.
11-19
-------
For this particle diameter and a mean wind velocity for the Lake Michigan
cruises of 4.7 m/sec, the deposition velocity is 2 x 10 cm/sec. The V^ for
field measurements (75) of 1 pm diameter particles and a wind speed of 4.7
m/sec was determined to be 32 x 10 cm/sec. The two values were used as a
range to bracket those depositions which might be seen for environmental PAH
aerosols. For dry deposition, the mean concentrations of individual compounds
for all cruises was selected. The dry flux of individual PAH compounds is
listed in Table 11-7. The flux to the northern sector of Lake Michigan is
less than the flux to the southern sector for all compounds and generally the
range of concentrations was an order of magnitude lower than dry TOG flux to
the respective parts of Lake Michigan. The highest flux occurred for fluorene
in the northern part of the lake and the highest fluxes were seen for
benzotghi]perylene, perylene, 2,3-benzofluorene and fluorene in the southern
part of the lake. High levels of benz[a]anthracene were seen for northern and
southern parts of the lake and the potent carcinogen benzfa]pyrene also was
found at relatively high levels in the southern third of Lake Michigan.
Wet PAH deposition was calculated for a particle of mass median diameter
of 1.0 ym. The scavenging ratio for this size is 160 (76) and rainfall to the
lake was taken as 75.2 cm/yr (77).
The wet flux of individual PAH compounds to the southern and northern
parts of Lake Michigan, is given in Table II-8. The values are virtually the
same for the upper range of dry deposition and are similar to the trend seen
for total suspended paticulate matter.
Bubble Ejection of PAHs from Lake Michigan Microlayer Water
In an attempt to understand the importance of the return of PAHs to the
atmosphere, bubble ejection was explored as a possible source for atmospheric
PAHs. The surface area of the ocean was taken as 3.61 x 10 m (78), the
surface area of Lake Michigan as 5.8 x 10 m (79) and a range for the bubble
ejected organic carbon transported from the ocean to the atmosphere from 5 x
10 g/yr to 5 x 10 g/yr (80). Scaling these values for the amount of
organic matter (1.5 x organic carbon) (81), and knowing the surface area of
Lake Michigan, a range from 1.2 x 10 g/yr to 1.2 x 10 g/yr of bubble
ejected organic matter released from the microlayer can be calculated. This
assumes that the number of bubbles formed in the open ocean is the same as the
number of bubbles produced per unit area of Lake Michigan, and that the
organic carbon and the PAH concentrations in the microlayer are the same as
those ejected into the air. As the bubble bursts, it peals off the microlayer
(70). A microlayer thickness of 300 ym was assumed, and the calculated
organic matter in the microlayer was determined to be from 0.69 to 6.9 g/L.
Using the organic matter concentrations, the ejected organic matter in air and
the mean concentrations of PAHs in the microlayer (Table II-6), the bubble
ejected flux from water to air for selected PAHs is: fluorene—6.1 kg/yr;
anthracene—5.2 kg/yr; phenanthrene—5.2 kg/yr; and fluoranthene—5.7 kg/yr.
This hardly would be significant compared to the 10 kg/yr flux of wet and dry
PAH deposition from air to water. However, the organic matter flux from water
to air calculated to be from 1.2 x 10 kg/yr to 1.2 x 10 kg/yr more than
exceeds that of the organic matter dry flux from the air to the water at
11-20
-------
Table II-7. Dry flux of polycyclic aromatic hydrocarbons to Lake Michigan
Compound
Fl
Ph
An
Flu
Blf
Py
BaA
Per
Tri
BaP
0-Pp
B[ghi]P
Northern 2/3 of Lake,
kg/yr x 102
0.72
0.36
0.24
0.48
0.36
0.24
0.48
0.24
0.24
- 11.5
- 5.8
- 3.8
- 7.7
- 5.8
- 3.8
- 7.7
- 3.8
- 3.8
^^"""
Southern 1/3 of Lake,
kg/yr x 102
1.6
0.9
0.9
1.4
2.3
1.7
1.3
2.1
0.43
1.2
1.1
2.8
- 25
- 14
- 14
- 22
- 36
- 28
- 20
- 33
— 6.8
- 19
- 17
- 44
Table II-8. Wet flux of polycyclic aromatic hydrocarbons to Lake Michigan
Compound
Fl
Ph
An
Flu
Blf
Py
BaA
Per
Tri
BaP
0-Pp
B[ghi]P
Northern 2/3 of Lake,
kg/yr x 102
1.1
0.57
0.37
0.73
0.57
0.37
0.73
0.37
0.37
—
—
Southern 1/3 of Lake,
kg/yr x 102
2.4
1.3
1.3
2.0
3.5
2.6
1.8
3.1
0.65
1.8
1.7
3.3
11-21
-------
13.3 x 10 kg/yr and would become important during high wind conditions when
numerous bubbles are formed.
Volatilization of PAHs from Lake Michigan
The approximate rates of evaporation of low water soluble PAHs were
examined as a route for entrance back to the atmosphere and to compare this
with the deposition flux from air to water. It has been shown that PAHs have
high activity coefficients in water resulting in a high evaporation rate (82).
The theoretical basis for mass transfer across the air-water interface
is :
Hk k
where
r is overall mass transfer coefficient
H is Henry's Law constant, a distribution coefficient
(mole/Lair:
k is gas phase mass transfer, the rate transport away from
the air interface (m/hr)
K£ is liquid phase mass transfer, the rate of transport to the
water interface (m/hr)
The gas phase and liquid phase mass transfer are affected by mixing and
temperature. Wind influences kg, water turbulence k. and temperature
influences H most significantly, although each environmental property affects
the other parameter properties to some degree.
In order to take the action of wind and water velocities into account ,
the equations of Southworth (83) were utilized:
k = 1137.5 (V . . + V )(—=———=T77) (2)
g wind current mol. wt. PAH
V0.969
current , 32 .1/2
_ 0.673 (mol. wt. PAH}
K
for
Vwind _< 1.9 m/sec
11-22
-------
and
0.969
_, C1 current 32 Nl/2 ( 0.526(V . .-I
\ = 23'51 ,0.673 (mol. wt. PAH> (e Wind
K
where
Vwind ^s w*nc* velocity (ra/sec)
^current is current velocity (m/sec)
R is mixing depth (m)
The laboratory estimates of Henry's Law coefficient and estimates based on the
ratio of equilibrium vapor pressure to aqueous solubility were chosen for
anthracene (2.66) and benz[a]pyrene (2.1 x 10~5) (83). These PAHs were
selected to cover a wide molecular weight range with available coefficients.
The mean wind speed of 4.7 m/sec for Lake Michigan was used and a current
velocity at 2% of this wind speed or 0.094 m/sec was selected on the basis of
the Hutchinson (84) discussion of wind generated currents. A mixing depth of
1 m was assumed as the depth at which instantaneous mixing equilibria were
obtained.
On the basis of the above values, equations and assumptions, coefficients
were calculated for anthracene and benz[a]pyrene. For anthracene, the mass
transfer results were: kg = 1732 cm/hr; k^ = 1.04 cm/hr; and KL = 1.04 cm/hr;
and for benz[a]pyrene k = 1456 cm/hr; k- = 0.87 cm/hr; and KL = 2.95 x 10
cm/hr. The mass transfer rates for the two compounds are liquid phase
controlled, thus factors directly affecting k^, such as current velocity
dramatically affect volatilization. Overall mass transfer follows the
expected trend, with the highest rate for low molecular weight anthracene and
the lowest rate for higher molecular weight (five aromatic rings)
benz[a]pyrene.
The mass flux, N^, of PAHs across a phase boundary necessary to predict a
steady state exchange rate of these compounds between the atmosphere and the
water body can be written in terms of the bulk liquid concentration and the
partial pressure in the atmosphere:
Ni =
11-23
-------
where
r\
NJ is mass flux (mol/m -hr)
K^T is overall liquid transfer coefficient (m/hr)
C. is bulk liquid concentration (mol/m )
o
H^ is Henry's Law coefficient (atm m /mol)
Pi is Cipis/Cis
P^s is vapor pressure (atm)
Cig is solubility in water (mol/m3)
Substituting the calculated KL, the Henry's Law coefficient and 0.3 pg/L or
1.68 x 10~6 mole/m3 (Table 11-6) as an estimate of bulk liquid concentration
for anthracene, the flux from Lake Michigan to the atmosphere was
determined. The conversion from Southworth's (83) Henry's Law constants was
made by making a density correction and multiplying by RT for the correct
units:
833-H-RT . (6)
where
T is absolute temperature (298°K)
R is gas constant (8.2 x 10~5 atm m3/mol °K)
o
The Henry's Law constant was 54.1 atm m /mole for anthracene and the
indicating that no anthracene is leaving the lake by volatilization, although
the flux is in the opposite direction from atmospheric vapor to the lake. No
calculation of flux was made for benz[a]pyrene due to the lack of solubility
and vapor pressure data, however, it is unlikely that this higher molecular
weight compound would volatilize to a greater extent than anthracene. It
appears from these approximations that volatilization is not a significant
flux and that wet and dry deposition dominate input with little return to the
atmosphere from the lake.
11-24
-------
REFERENCES
1. National Academy, of Sciences. Particulate Polycyclic Organic Matter.
U.S. GPO, Washington, D.C., 1972.
2. Liberti, A., G. P. Cartonia and V. Cantuti. Gas Chromatographic
Determination of Polynuclear Hydrocarbons in Dust. J. Chromat. 15:141,
1964.
3. Cantuti, V., G. P. Cartoni, A. Liberti and A. G. Tarri. Improved
Evaluation of Polynuclear Hydrocarbons in Atmospheric Dust by Gas
Chromatography. J. Chromat. 17:60, 1965.
4. Brocco, D., V. Cantuti and G. P. Cartoni. Determination of Polycyclic
Hydrocarbons in Atmospheric Dust by a Combination of Thin Layer and Gas
Chromatography. J. Chromat. 49:66, 1970.
5. Arito, H., R. Soda and H. Matsushita. Gas-Chromatographic Determination
of Polynuclear Hydrocarbons in Particulate Air Pollutants. Ind. Health
5:243, 1967.
6. Abraham, M. H. and R. E. Marks. Gas Liquid Chromatography of Some
Condensed-Ring Aromatic Hydrocarbons of Lightly-Loaded Columns. J.
Chromat. 13:344, 1964.
7. Birchfield, H. P., R. J. Wheeler and J. B. Bernos. Fluorescence Detector
fo Analysis of Polynuclear Arenes by Gas Chromatography. Anal. Chem.
43:1976, 1971.
8. Natusch, D. F. S. Report for Colstrip Project. Colorado State Univ.,
1975.
9. Kubota, H., W. H. Griest and M. R. Guerin. Determination of Carcinogens
in Tobacco Smoke and Coal-Derived Samples of Trace Polynuclear Aromatic
Hydrocarbons. 9th Conf. Trace Substances in Environmental Health, Univ.
of Missouri, Columbia, 1975.
10. Cukor, P., L. L. Ciaccio, E. W. Lanning and R. L. Rubino. Some Chemical
and Physical Characteristics of Organic Fractions in Airborne Particulate
Matter. Environ. Sci. Technol. 6:441, 1972.
11. Ciaccio, L. L., R. L. Rubino and J. Flores. Composition of Organic
Constituents in Breathable Airborne Particulate Matter Near a Highway.
Environ. Sci. Technol. 8:935, 1974.
12. Grimmer, G. and H. Bohnke. Determination of Constituent PAHs in Airborne
Dust and Motor Vehicle Exhaust by Capillary GC. J. Assoc. Off. Anal.
Chem. 55:261, 1972.
11-25
-------
13. Gordon, R. J. Distribution of Airborne Polycyclic Aromatic Hydrocarbons
Throughout Los Angeles. Environ. Sci. Technol. 10:370, 1976.
14. Gordon, R. J. and R. J. Bryan. Patterns in Airborne Polynuclear
Hydrocarbon Concentrations at Four Los Angeles Sites. Environ. Sci.
Technol. 7:1050, 1973.
15. Henderson, B. E., R..J. Gordon, H. Menck, J. Soo Hoo, S. P. Martin and
M. C. Pike. Lung Cancer and Air Pollution in Southcentral Los Angeles
County. Am. J. Epidemiol. 9:113, 1975.
16. Zoccolillo, L., A. Liberti and D. Brocco. Determination of Polycyclic
Hydrocarbons in Air by Gas Chromatography with High Efficiency Packed
Columns. Atmos. Environ. 6:715, 1972.
17. Doran, T. and N. G. McTaggart. The Combined Use of High Efficiency
Liquid and Capillary Gas Chromatography for the Determination of
Polycyclic Aromatic Hydrocarbons in Automative Exhaust Condensates and
Other Hydrocarbon Mixtures. J. Chromat. Sci. 12:715, 1974.
18. Lao, R. C., R. S. Thomas, H. Oja and L. Dubois. Application of a Gas
Chromatograph Mass Spectrometer - Data Processor Combination to the
Analysis of the Polycyclic Aromatic Hydrocarbon Content in Airborne
Particulates. Anal. Chem. 45:908, 1973.
19. Lee, M. C., M. Novotny and K. D. Bartle. Gas Chromatography/Mass
Spectrometric and Nuclear Magnetic Resonance Determination of Polynuclear
Aromatic Hydrocarbons in Airborne Particulates. Anal. Chem. 48:1566,
1976.
20. Novotny, M., M. L. Lee and K. D. Bartle. The Methods for Fractionation,
Analytical Separation, and Identification of Polynuclear Aromatic
Hydrocarbons in Complex Mixtures. J. Chromat. Sci. 12:606, 1974.
21. Bartle, K. D., M. L. Lee and M. Novotny. High Resolution GLC Profiles of
Urban Air Pollutant Polynuclear Aromatic Hydrocarbons. Intern. J. Env.
Anal. Chem. 3:349, 1974.
22. Thomas, R. S. and J. L. Monkman. C. D. Report No. 215, Chemistry
Division, Technology Development Branch, Air Pollution Control
Directorate, Department of Environment, Ottawa, Canada.
23. Monkman, J. L. and R. F. Pottie. Proceedings of the International
Symposium of the Identification and Measurement of Environmental
Pollutants. Ottawa, Canada.
24. Lao, R. C., R. S. Thomas and J. L. Monkman. Analysis of Environmental
Samples by Mass Spectrometry. Intern. J. Env. Anal. Chem. 1:187, 1972.
25. Lao, R. C., R. S. Thomas and J. L. Monkman. Computerized Gas
Chromatographic Mass Spectrometric Analysis of Polycyclic Aromatic
Hydrocarbons in Environmental Samples. J. Chromat. 112:681, 1975.
11-26
-------
26. Dong, M., D. C. Locke and E. Ferrand. High Pressure Liquid
Chromatographic Method for Routine Analysis of Major Parent Polycyclic
Aromatic Hydrocarbons in Suspended Particulate Matter. Anal. Chem.
48:368, 1976.
27. Fox, M. A. and S. W. Staley. Determination of Polycyclic Aromatic
Hydrocarbons in Atmospheric Particulate Matter in High Pressure Liquid
Chromatography Coupled with Fluorescence Techniques. Anal. Chem. 48:992,
1976.
28. Janini, G. M., K. Johnston and W. L. Zielinski, Jr. Use of a Nematic
Liquid Crystal for Gas-Liquid Chromatography Separation of Polyaromatic
Hydrocarbons. Anal. Chem. 47:670, 1975.
29. Zielinski, W. L., Jr., K. Johnston and G. M. Muschill. Nematic Liquid
Crystal for Gas-Liquid Chromatographic Separation of Steroid Epimers.
Anal. Chem. 48:907, 1976.
30. Zielinski, W. L., Jr. Liquid Crystals Anisotropic GLC Phases for Novel
High-Temperature Separations. Analabs Technical Bulletin, 1977.
31. Wiley, J. C., Jr., C. S. Menon, D. L. Fisher and J. E. Engel.
Metabolites of Polycyclic Aromatic Hydrocarbons. II. Isomeric K-Region
Phenols and Methyl Ethers of Benz[a]anthracene. Tetrahedron Letters
33:2811, 1975.
32. Janini, G. M., G. M. Muschill and W. L. Zielinski, Jr. N, N'-bis [p-
butoxy-benzylidinel-cx1 , a'-bi-p-toluidine: Thermally Stable Liquid
crystal for Unique Gas-Liquid Chromatography Separations of Polycyclic
Aromatic Hydrocarbons. Anal. Chem. 48:809, 1976.
33. Janini, G. M., G. M. Muschill, J. A. Schroen and W. L. Zielinski, Jr.
Gas-Liquid Chromatographic Evaluation and Gas-Chroinatographic/Mass
Spectrometric, Application of New High Temperature Liquid Crystal
Stationary Phases for Polycyclic Aromatic Hydrocarbon Separations. Anal.
Chem. 48:1879, 1976.
34. Cautreels, W. and K. van Cauwenberger. Determination of Organic
Compounds in Airborne Particulate Matter by Gas Chromatography - Mass
Spectrometry. Atm. Env. 10:447, 1976.
35. Giger, W. and C. Schaffner. Determination of Polycyclic Aromatic
Hydrocarbons in the Environment by Glass Capillary Gas Chromatography.
Anal. Chem. 50:243, 1978.
36. Karesek, F. W., D. W. Denney, K. W. Chan and R. E. Clement. Analysis of
Complex Organic Mixtures on Airborne Particulate Matter. Anal. Chem.
50:82, 1978.
37. Severson, R. F., M. E. Snook, R. F. Arrendale and 0. T. Chartyk. Gas
Chromatographic Quantitation of Polynuclear Aromatic Hydrocarbons in
Tobacco Smoke. Anal. Chem. 48:1866, 1976.
11-27
-------
38. Pierce, R. C. and M, Katz. Dependency of Polynuclear Aromatic
Hydrocarbon Content on Size Distribution of Atmospheric Aerosols.
Environ. Sci. Technol. 9:347, 1975.
39. De Maio, L. and M. Corn. Polynuclear Aromatic Hydrocarbons Associated
with Particulates in Pittsburgh Air. J. Air Pollut. Assoc. 16:677, 1966.
40. Kertesz-Saringer, M., E. Mescaras and T. Varkomji. On the Size
Distribution of Benzo[a]pyrene Containing Particles in Urban Air. Atm.
Env. 5:429, 1971.
41. Thomas, J. F., M. Mukai and B. D. Tebbens. Fate of Airborne
Benzo[a]pyrene. Environ. Sci. Technol. 2:33, 1968.
42. Kesteridis, G., J. Hahn, R. Jaessicke and C. Junge. The Organic
Constituents of Atmospheric Particulate Matter. Atm. Env. 10:603, 1976.
43. Kesteridis, G. and J. Hahn. Determination of Organic Constituents of
Aerosol Particles in Unpolluted Air. Z. Anal. Chem. 273:257, 1975.
44. Barger, W. R. and W. D. Garrett. Surface Active Organic Material in the
Marine Atmosphere. J. Geophys. Res. 75:4561, 1970.
45. Hoffman, E. J. and R. A. Duce. The Organic Carbon Content of Marine
Aerosols Collected on Bermuda. J. Geophys. Res. 70:4474, 1974.
46. Lee, R. J., Jr. and J. Wagman. A Sampling Anomaly in the Determination
of Atmospheric Sulfate Concentration. Am. Ind. Hygiene Assoc. J. 27:266,
1966.
47. Grosjean, D. and S. K. Friedlander. Gas-Particle Distribution Factors
for Organic Pollutants in the Los Angeles Atmosphere. 67th Annual
Meeting of the Air Pollution Control Association, Denver, Colorado, 1974.
48. Strand, J. W., T. Stolzenburg and A. W. Andren. A Constant Relative
Humidity-Temperature Chamber for the Accurate Weight Determination of Mr
Particulate Matter Collected on Filters. Atm. Env. 12:2027, 1978.
49. Hoffman, G. L. and R. A. Duce. Copper Contamination of Atmospheric
Particulate Samples Collected with Gelman Hurricane Air Samplers.
Environ. Sci. Technol. 5:1134, 1971.
50. Courtness, R. J. Production of Aerosols by High Volume Samplers. J. Air
Pollut. Control Assoc. 24:605, 1974.
51. Lewis, R. G. and N. J. Zimmerman. Danger of Re-circulation in H^-vols.
Anal. Qual. Control Newsl. 28. U.S. Environmental Protection Agency
Environmental Monitoring and Support Laboratory, Cincinnati, Ohio, 1976.
11-28
-------
52. Environmental Science and Engineering, Inc. Evaluation of a Method for
the Analysis of Airborne Polychlorinated Biphenyls. Report Prepared for
the Office of Toxic Substances, U.S. Environmental Protection Agency,
1976.
53. Falk, H. L., I. Markul and P. Kotin. Aromatic Hydrocarbons. IV. Their
Fate Following Emission into Atmosphere and Experimental Exposure to
Washed Mr and Synthetic Smog. Am. Med. Assoc. Arch. Ind. Health 13:13,
1956.
54. U.S. Environmental Protection Agency. Scientific and Technical
Assessment Report on Particulate Polycyclic Organic Matter. PB-241,
1975. p. 799.
55. Green, L. Operating Conditions for Optimum Performance of the Model
5711A Flame lonization Detector. Hewelett Packard Application Note, GC
2-73, 1973.
56. Hertz, H. S., W. E. May, S. A. Wise and S. N. Chesler. Trace Organic
Analysis. Anal. Chem. 50:429A, 1978.
57. World Health Organization. IARC Monographs on the Evaluation of
Carcinogenic Risk of the Chemical to Man, Certain Polycyclic Aromatic
Hydrocarbons and Heterocyclic Compounds. 1973.
58. Weast, R. C. (ed.) CRC Handbook of Chemistry and Physics. 50th Ed.
Chemical Rubber Co., Cleveland, Ohio, 1970.
59. Braunstein, H. M., E. D. Copenhaver and H. A. Pfuderer (eds.).
Environmental, Health, and Control Aspects of Coal Conversion: An
Information Overview. Oak Ridge National Laboratory, Oak Ridge,
Tennessee. ERl,ll:ORNL/EIS-95 UCL1, 41, 48, 90i, 95, 1977.
60. Mariconi, E. J. and L. Salce. Ozonation of Polycyclic Aromatics. XV.
Carcinogenicity and K- and/or L-Region Additivity Towards Ozone. Adv.
Chem. Ser. 77:65, 1968.
61. Falk, H. L., P. Kotin and A. Miller. Aromatic Polycyclic Hydrocarbons in
Polluted Air as Indicators of Carcinogenic Hazards. Int. J. Air Pollut.
2:201, 1960.
62. Pupp, C., R. C. Lao, J. J. Murray and R. F. Pottie. Equilibrium Vapour
Concentrations of Some Polycyclic Aromatic Hydrocarbons, As406 and SeOo
and the Collection Efficiencies of these Air Pollutants. Atmos. Environ.
8:915, 1974.
63. Doskey, P. V. Personal Communication. Univ. of Wisconsin, Madison, 1978.
64. Andelman, J. B. and M. J. Suess. Polynuclear Aromatic Hydrocarbons in
the Water Environment. Bull. World Health Org. 43:479, 1970.
65. Garrett, W. D. Damping of Capillary Waves at the Air-Sea Surface by
Oceanic Surface-Active Material. J. Mar. Res. 25:279, 1967.
11-29
-------
66. Andren, A. W., A. W. Elzerman and D. E. Armstrong. Chemical and Physical
Aspects of Surface Organic Microlayers in Freshwater Lakes. J. Great
Lakes Res. 2:101, 1976.
67. Elzerman, A. W. Surface Microlayer-Microcontaminate Interaction in
Freshwater Lakes. Ph.D. Thesis. Univ. of Wisconsin, Madison, 1976.
71 pp.
68. Winchester, J. W. and G. D. Nifong. Water Pollution in Lake Michigan by
Trace Elements from Pollution Aerosol Fallout. Water, Air, Soil Pollut.
1:50, 1971.
69. Borneff, J., F. Selenka, H. Kunete and A. Maximas. Experimental Studies
on the Formation of Polycyclic Aromatic Hydrocarbons in Plants. Environ.
Res. 2:22, 1968.
70. Maclntyre, F. The Top Millimeter of the Ocean. Sci. Am. 230:66, 1974.
71. Wallace, G. T., Jr. and R. A. Duce. Concentration of Particulate Trace
Metals and Particulate Organic Carbon in Marine Surface Waters by Bubble
Flotation Mechanism. Mar. Chem. 3:157, 1975.
72. Acheson, M. A., R. M. Harrison, R. Perry and R. A. Wellings. Factors
Affecting the Extraction and Analysis of Polynuclear Aromatic
Hydrocarbons in Water. Water Res. 10:207, 1976.
73. Herbes, S. E. Partitioning of Polycyclic Aromatic Hydrocarbons Between
Dissolved and Particulate Phases in Natural Waters. Water Res. 11:493,
1977.
74. Sehmel, G. A. and S. L. Sutter. Particle Deposition Rates on a Waters
Surface as a Function of Particle Diameter and Air Velocity. J. Res.
Atmos. 3:911, 1974.
75. Cawse, P. A. Report AERE-R7669, Environmental and Medical Sc-iences
Division, United Kingdom Atomic Energy Authority, Harwell, Oxford-Shire,
1974.
76. Gatz, D. F. Pollutant Aerosol Deposition into Southern Lake Michigan.
Wate, Air, Soil Pollut. 5:239, 1975.
77. Changnon, S. A., Jr. Precipitation Climatology of Lake Michigan Basin.
Illinois State Water Survey, Urbana, Bull. 52, 1968.
78. Johnson, M. W. and R. H. Fleming. The Oceans, Their Physics, Chemistry,
and General Biology. Prentice Hall, Englewood Cliffs, New Jersey, 1946.
79. Klein, D. H. Fluxes, Residence Times and Sources of Some Elements to
Lake Michigan. Water, Air, Soil Pollut. 4:3, 1975.
80. Duce, R. A. and E. J. Hoffman. Chemical Fractionation at the Air/Sea
Interface. Rev. Earth Planet. Sci. 4:187, 1976.
11-30
-------
81. Grosjean, D. Personal Communication as Cited in G. Gartrell, Jr. and S.
K. Friedlander. Atmos. Environ. 9:279, 1973.
82. MacKay, D. and P. J. Leinonen. Rate of Evaporation of Low-Solubility
Contaminants from Water Bodies to Atmosphere. Environ. Sci. Technol.
9:1178, 1975.
83. Southworth, G. R. Transport and Transformations of Anthracene in Natural
Waters: Process Rate Studies. Atmos. Environ. , 1978.
84. Hutchinson, G. E. A Treatise on Limnology. Vol. I. John Wiley and
Sons, -New York, 1957. 291 pp.
85. MacKay, D. and W. Y. Shiu. Aqueous Solubility of Polynuclear Aromatic
Hydrocarbons. J. Chem. Eng. Data 22:399, 1977.
11-31
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
TTD/\ nr\£; /; r7o_noa_T
4. TITLE AND SUBTITLE
Atmospheric Chemistry of PCBs and PAHs-Volume 9
7. AUTHOR(S)
A. 1. Andren, P. V. Doskey and J. W. Strand
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Wisconsin Water Resources Center
University of Wisconsin
1975 Willow Drive
Madison, Wisconsin 53706
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Great Lakes National Program Office
536 South Clark, Street, Room 932
3. RECIPIENT'S ACCESSION- NO.
5. REPORT DATE
March 1980
6. PERFORMING ORGANIZATION
8. PERFORMING ORGANIZATION
10. PROGRAM ELEMENT NO.
A42B2A
11. CONTRACT/GRANT NO.
R005142
CODE
13. TYPE OF REPORT AND PERIOD COVERED
Final Report 1974-1979
14. SPONSORING AGENCY CODE
U.S. EPA-GLNPO
15. SUPPLEMENTARY NOTES
Water Chemistry Program, Department of Civil and Environmental Engineering, University
of Wisconsin-Madison assisted.
16. ABSTRACT
The air over Lake Michigan was sampled during 1977 to develop a collection
method for PCBs and obtain data about their atmospheric transport and dry
deposition onto the lake. A resin, XAD-2, was the most efficient collection
medium for PCB vapor and was incorporated into standard high volume air
samples for the collection of particulate and vapor phase PCBs. PCB
concentrations in air samples taken over Lake Michigan were lower than those
taken from urban areas; i.e., Milwaukee. Aroclors 1242 and 1254 were the main
components of vapor phase PCBs while in some instances the particulate phase
PCBs contained Aroclor 1260. The particulate phase PCBs over Lake Michigan
contained a larger percentage of the more volatile mixtures than those
reported in urban areas such as Chicago and Milwaukee. PCBs tend to associate
with particulates 0.002 to 0.1 urn in diameter. The amount and organic carbon
content of the particulate phase appear to control vaporization and
revolatilization of PCBs.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Samples
Organic carbon
Atmospheric
Contaminant s
Bubble ejection
Hydrocarbon
Vaporization
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (ThisReport/
Document is available to the public through
the National Technical Information Service,
20. SECURITY CLASS (Thispage)
21. NO. OF
126
22. PRTCT
EPA Form 2220-1 (9-73)
Jf U. S. Government Printing Office 1981 750-808
------- |