EPA-600/2-81-161
September 1981
SOURCE RESOLUTION OF POLYCYCLIC AROMATIC
~~HYDROCARBONS IN THE LOS ANGELES ATMOSPHERE
Application of a Chemical Species Balance Method
with First Order Chemical Decay
by
Marc Maurice Duval and S. K. Friedlander
Dept. of Chemical, Nuclear, and Thermal Engineering
University of California, Los Angeles
Los Angeles, California 90024
Grant No. R806404-02S1
Project Officer
Stanley L. Kopczynski
' Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences
Research Laboratory, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
TECHNICAL REVIEW
This research report was accepted as a master's thesis and thus
has been technically reviewed by academic peers.
11
-------
Abstract
The chemical element balance method has been extended to chemi-
cally reactive components of the atmospheric.,aerosol. The data used
for source emissions nrd ambient concentrations were taken fion the
literature. Contributions of automobile and refinery emissions to
selected airborne polycyclic aromatic hydrocarbons (PAH) concentra-
tions were determined for 13 sites in the Los Angeles basin over the
period 1970-75. Automobile emissions were predominant except near
a region with a large concentration of refineries. The average devi-
ations of the calculated PAH concentrations, obtained by adding the
contributions from the separate sources, from the measured concentra-
tions were within -1H to 71.
The data on automobile emissions were averaged over one hundred
cars registered since October 1, 1971, and representative of the
German automobile fleet. They were assumed to hold for the Los Angeles
fleet. The source resolution could be improved by performing the
•
same type of analysis on the Los Angeles fleet.
New data on refinery emissions and rates of atmospheric degra-
dation of benzo(a)pyene, anthanthrene, benzo(ghi)perylene and benzo-
fluoranthenes were obtained as a result of the analysis. Benzo(a)-
pyrene and anthanthrene were found to be the most reactive species.
An average residence time for aerosol particles in Los Angeles,
as well as lead to PAH ratios fron automobile emissions, are also
reported.
The literature on PAH emissions from major combustion sources
was reviewed. Agreement and discrepancies among investigators were
evaluated.
111
-------
TABLE OF CONTENTS
ABSTRACT 111
LIST OF TABLES V11
LIST OF ABBREVIATIONS V111
I. INTRODUCTION 1
I.I Air Pollution Modeling 1
1.2 Properties of PAH 4
II. THEORY 4
II.1 Chemical Element Balance Method 6
II.2 Extension of the Chemical Element Balance Method
to Reactive Species 7
II.3 Relationship of the Decay Factor to the Reaction
Rate Coefficient '. . 10
II.3.1 CSTR 12
II.3.2 PFR 13
III. SOURCE CONCENTRATIONS OF PAH 14
III.l Coal Combustion 14
III.2 Coke Production 15
II 1.3 Incineration 16
III.4 Wood Combustion 18
III.5 Open Burning 18
III.6 Gasoline-powered Cars 19
III.7 Diesel-powered Cars 20
IV. SOURCE RESOLUTION ANALYSIS FOR LA PAH: RESULTS AND
DISCUSSION 24
IV.1 Autmobile Emissions--PAH Decay Factors 24
IV.2 Refinery Emissions 31
IV.3 Source Resolution -34
V. CONCLUSIONS AND FURTHER CONSIDERATIONS 43
REFERENCES 46
-------
APPENDIX A 52
APPENDIX B 53
APPENDIX C 54
-------
LIST OF TABLES
Table 1 Comparison of PAH to perylene emission ratios.
fron coke production 17
Table 2 Selected PAH to BghiP emission ratios fron
wood and peat combustion 18
Table 3 PAH pattern comparisons 26
Table 4 PAH decay factors 28
Table 5 Comparison of rates of degradation of selected PAH . 30
Table 6 Source concentration matrix 32
Table 7 Contributions of automobiles and refineries to
selected airborne PAH at various locations of
Los Angeles 37
Table 8 Contributions of automobiles and refineries to
selected airborne PAH at various locations of
Los Angeles 39
Table 9 Los Angeles sampling sites with freeway network ... 40
Table 10 Los Angeles County divided into 13 areas (exluding
sparsely populated northern portion) 41
Table 11 Average lead to PAH ratios from automobile emissions
and per cent deviations of the calculated PAH con-
centrations from the measured concentrations .... 42
Table 12 Pattern of Refinery Contributions at Site 3 54
V11
-------
LIST OF ABBREVIATIONS
A/F air to fuel ratio
Amu atomic mass unit
ANT anthracene
ANTH anthanthrene
APCD Air Pollution Control District
AQMD Air Quality Management District
BaA benzo(a)anthracene
BaP benzo(a)pyrene
BbF benzofbjfluoranthene
BeP benzo(e)pyrene
BkF benzo(k)fluoranthene
BFL benzofluoranthenes » benzo(b)fluoranthene + benzo(j)fluoranthene
+ benzo(k)fluoranthene
BghiP benzo(ghi)perylene
CO carbon monbxide
COR coronene
CSTR continuous stirred tank reactor
EM engine modification
FLT fluoranthene
HC hydrocarbons
INP indeno(l,2,3-cd)pyrene
NO nitrogen oxides
PAH polycyclic aromatic hydrocarbons
PER perylene
VH1 "\
-------
PFR plug flow reactor
PHT phenanthrene
PYR pyrene
TSP total suspended participates
FID flame ionization detector
GC gas chromatography
MS mass spectrometry
-------
I. INTRODUCTION
I.I Air Pollution Modeling
Relating source emissions to air quality can provide a rational
solution to the problea of source apportionment. Two methods of
assessment have been developed separately in the past (NAS, 1980):
Dispersion models and receptor models. Such predictive techniques
can be used to control exposures of populations to airborne pollu-
tants. They may also be used to foresee the impact of new tech-
»
nologies or shifts in technologies such as coal or wood gasifica-
tion and dieselization of the automobile fleet. Thus, these
methods are of considerable importance in air quality management.
Dispersion models are mathematical expressions which simulate
the transport and dispersion of emissions in the atmosphere. The
input variables include emission and meteorological data, initial
and boundary conditions (Turner, 1979). The output gives an esti-
mate of concentrations of pollutants over an area for various
periods of time. Accuracy of these models, evaluated by computing
the deviations between estimated and measured ambient concentra-
tions, is generally controlled by the emission and meteorological
data uncertainties.
Receptor models are based on empirical relationships between
some known source characteristics and ambient concentrations mea-
sured at the sampling sites. Factor analysis and chemical element"
balances have shown to be the most powerful approaches.
-------
Factor analysis hao been recently reviewed by Gordon (1980b) and
Hopke (1980). The objective of the technique is to determine the
*
common factors (sources of emissions) which best account for the devi-
ations in atmospheric measurements. For, example, Hopke et al. (1976),
using a set of 18 elements taken in 90 samples from the Boston area,
showed that six common factors were responsible for 77.S\ of the
total variance in the systen. They are sea salt, oil, auto emissions,
soil (mixed with coal), refuse incineration and a sixth factor which
could not be associated with a particular source.
Unlike factor analysis, the chemical clement balance method
assumes a priori that certain classes of sources are responsible for
ambient concentrations of elements measured at the receptor. Further-
more, it is assumed that each source under consideration emits a
characteristic and conservative set of elements (Friedlander, 1973).
The method was first applied by Miller et al. (1972) to determine the
contributions of sea salt, soil, automobile emissions and oil fly ash
to the Pasadena aerosol. Contributions were traced by means of a
mass balance on the aerosol, weighted by the mass fraction of elements
present in the sources at the point of emission. Friedlander (1973)
further developed the method and extended it to an overdetermined
problem using a least-squares fit weighted by errors in the airborne
measurements. This approach has been recently refined introducing
errors in the source compositions as well as in atmospheric measure-
«
ments (Watson, 1979; Dunker, 1979).
The chemical element balance method has been applied to several
parts of the United States. Gatz (1975) estimated the contributions
2 :
-------
of soil, automobile emissions, fuel oil burning, cement manufactur-
ing, iron and steel manufacturing and coal burning at various loca-
tions of the Chicago area. Kowalczyck et al. (1978) performed a
similar analysis of the Washington, D.C. aerosol. Assuming soil,
marine, coal, oil, refuse ai>d~Tiotor vehicles to be responsible for
the aerosol, they were able to account for about 801 of the total
suspended particulates (TSP). Dzubay (1980) in his study of the
St. Louis aerosol, distinguished between the fine and the coarse
fraction. Assuming that seven sources were responsible for the aero-
sol burden, he was able to account for 78% of the fine fraction and
96\ of the coarse fraction.
A major weakness of the chemical element method is its diffi-
culty to account for non-stable compounds. For example, secondary
materials, such as sulfates and nitrates, which can be included in
factor analysis must be handled separately in chemical element
balances (Friedlander, 1973; Gartrell and Friedlander, 1975). As
shown in this work, the methodology can be extended to a reactive
part of ambient aerosols, assuming it follows first order decay
laws. Polycyclic aromatic hydrocarbons (PAH) which account for a
small mass fraction of urban aerosols (^.1 mg/g of TSP) are of
special interest (Dong et al., 1976; Gordon, 1976 for example). As
shown later, to account for chemical decay it is necessary to develop
hybrid models incorporating some elements of receptor and dispersion
A
modeling.
3
-------
1.2 Properties of PAH
PAH are by-products of the combustion of carbonaceous matter
and are believed to be formed by a free radical mechanism (Critten-
den and Long, 1978). They are emitted as vapors and because of their
high melting and boiling points, are subsequently adsorbed onto the
particulate phase and primarily soot (MAS, 1972).
Several studies on the size-distribution of PAH in urban air
have shown that most of their mass is associated with smaller
particles (<3 urn) and are consequently in the respirable range
(Albagli et al., 1974; Katz and Chan, 1980 for example).
In the United States, the total emission rate of benzo(a)-
pyrene (BaP), the most studied PAH because of its notable carcino-
genicity, was estimated to approach 1300 tons/year in the early
seventies (Su3ss, 1976). Refuse and open burning, heating and
power generation and coke production were estimated to account for
almost 98\ of the total emissions while vehicles were believed to
be responsible for less than 2%. It is evident that depending on
the sampling site, these figures can vary significantly. In par-
ticular, automobile emissions are likely to dominate the overall
spectrum of emissions in large cities.
Due to the recent development of high resolution analytical
techniques, great improvements in the accuracy of PAH measurements
have been made. It has been shown that filter losses can alter
significantly the accuracy of measurements. Pupp et al. (1974)
calculated the equilibrium vapor concentrations of pyrene, benzo-
(a)anthracene (BaA), BaP, Benzo(e)pyrene (BeP), benzo(ghi)perylene
-------
(BghiP), and coronene (COR). They concluded that equilibrium vapor
concentrations can be considered as a measure of the collection
losses upper limits. The equilibrium vapor concentration of BaA
(228 amu) is almost 2 ug/m at 25*C which is approximately three
orders of magnitude greater than the measured atmospheric BaA
concentrations. As molecular weights increase, the corresponding
equilibrium vapor concentrations "decrease. The equilibrium vapor
concentrations of BaP (252 amu) and coronene (300 amu) at 2S°C
are respectively less than 0.1 yg/rn and 0.1 ng/m indicating that,
in the case of BaP, collection losses from atmospheric samples
can be important. De Wiest and Rondia (1976) measured the particu-
late and gas phase BaP of the Liege aerosol. The reported gas phase
BaP concentrations were always less than 15% of the total BaP
concentrations at temperatures not exceeding 25°C, but reached 44%
at 41°C. Miguel and Friedlander (1978) found no measurable BaP in
the gas phase in their measurements of the Pasadena aerosol from
October 1976 to March 1977.
Based on these results, atmospheric data will be analyzed for
the most commonly reported PAH whose molecular weights are greater
or equal to 252 ami.
-------
II. THEORY
II.1 Chemical Element Balance Method
The chemical element balance method applied to species i can
be considered to represent the equation of the center of mass of
species i:
c. .m.
where p. is the mass concentration of species i measured at the
receptor site; c., is the mass fraction of element i present in
the source m. at the point of emission. The source contribution,
c..m,, is the mass concentration of element i from source m. at the
point of measurement. It is assumed that species i is conservative.
If a source is missing in equation (II-l), the contribution of the
others will be overestimated.
Sources are classified into classes such as automobiles, power
plants and so on. Thus, aside from errors in the measurements, we
induce statistical errors in the determination of the c..'s. These
errors can be minimized by introducing more equations (elements) than
unknowns (sources). The overdetermined system can be solved by the
least-squares fitting technique (Friedlander, 1973). Suppose we
measure n elements which are known to be emitted by p sources.
Equation (II-l) can be rewritten with a matrix notation as follows^:
[P] = [C][M]
where [P] is a n x 1 matrix whose generic term on the ith row is
-------
p.; [C] is a n x p matrix whose generic term on the ith row and jth
column is c, ,; [M] is a p x 1 matrix whose generic tern on the ith
row is », . In order to proceed with the least-squares fitting, the
following assumptions are made:
1. The errors affecting the measurement of the Pi's are
normally distributed and uncorrelated.
2. The measured values of the source concentrations are exact.
3. The set of measurements of p.'s is the most probable set
(maximun likelihood principle).
•
If the two first conditions are met, the probability, P, of observing
a set of values between p.,...p and p.+dp.,...p +dp is (Young, 1962)
PJ" pn
where ]Cc'-5n< represents the exact value of p. obtained in the ab-
j-1 1J 3 i
sence of error in the measurements; a is the standard deviation in
Pi
the measurement of p., i=l...n.
Assuming the measured set of values of Pi's represent the
roost probable set, it is necessary to maximize equation (II-2). In
other words we have to choose the values of the BJ'S which miniaize
the argument of the exponential function in (II-2):
pi
Setting the derivative of X with respect to each n. equal to zero
-------
yields the result (the derivation is shown in Appendix A):
[D][M] - [T] (II-3)
where
[D] is a p x p matrix whose generic element on the jth row
and mth column is: •.
n c, .c.
pi
[M] is a p x 1 matrix whose generic term on the jth row is m.
[T] is a p x 1 matrix whose generic term on the jth row is:
A cijpi
pi
ii
The source contributions can be found by solving equation (II-3).
If the errors in the measurements of the p.'s are not known it can
be assumed that they belong to the same infinite parent distribu-
tion, that is:
..a « o i»l...n
pi
In this case, the value of o need not be precise since it can be
eliminated from equation (II-3).
This method has been extended by Watson (1979) to cases where
the uncertainties in the source concentrations are known. Assun-^
ing that these uncertainties are normally distributed and uncorre-
lated, especially to uncertainties in the p^'s, an analysis similar
-------
to the one discussed above can be performed. However, it will not
be included in the present treatment since uncertainties in refinery
emissions were not known, (see section IV.2).
II.2 Extension of the Chemical Element Balance Method to Reactive
Species
Mass fractions of species under consideration at the point of
emission are not always available. Species of interest may react
in the atmosphere. In these cases, the chemical element balance
method must be reformulated. Equation (II-l) can be rewritten as
follows:
uVu CII-4)
where the dimensionless decay factor, a.,, is the fraction of species
i emitted from source j remaining in the aerosol at the receptor
Y«
site and x,, is the dimensionless ratio of mass of species i tn, the
reference species 1 in the emission from source j. The source con-
tribution, y. ., is the mass concentration of reference species 1
from source j at the point of measurement.
It is convenient to choose as a reference species a non-reactive
component of the aerosol, preferably in the same family or group under
study.
The overdetermined system of n equations with p unknowns
(n > p) can be solved by the same technique discussed in Section II-l.
Assuming that errors in the measurements are normally distri-
buted and uncorrelated, that the decay factors and source coraposi-
-------
tions are exact and applying the maximum likelihood principle, the
solution to the least squares fitting is:
[Z][Y] « [Q]
where:
[Z] is a p x p matrix whose generic tern on the jth row and
mth column is
n a. .x. .a. x.
v-> ij ij im in
[Y] is a p x 1 matrix whose generic term on the jth row is y.,.
[Q] is a p x 1 matrix whose generic term on the jth row is:
a 2
p
The same conclusions noted in section II-l can be drawn when errors
in the measurements are not known.
II. 3 Relationship of the Decay Factor to the Reaction Rate Coefficient
The decay factor, a. ,, accounts for the mass fraction of reacting
species i emitted from source j and remaining in the aerosol at the
receptor site. It is unity for non-reactive species. In the case of
reactive species it can be obtained computing the ratio of atmospheric
concentrations found at the receptor site to the corresponding concen-
*
trations measured at the point of emission (see section IV'.'l-) . 'As shown
below, if the atmospheric reaction of species i follows first order
10
-------
kinetics, the decay factor can be related to the reaction rate
coefficient.
Let p,. be the mass concentration of species i of the aerosol
(or gas) emitted from source j. The rate of decay of species i of
the aerosol is given by:
dp
-at1 - -Vu
where k. is the reaction rate constant for species i. Integrating
from the time of release from the source (t « 0) to the time of col-
lection at the receptor (t - T) yields to:
-±>- « exp(-k.T)
Pij,o *
where p., is the mass concentration of species i emitted fron
source j and measured at the point of emission.
In general for any sample taken at the receptor site there
will be a residence time distribution g(f) such as:
df - g(T)di
where df is the fraction of the material sampled with atmospheric
residence times between T and T+dT. The amount of species i emitted
from source j remaining at the sampling point is obtained by inte-
gration over all residence times:
,./.
11 •
-------
Further analysis requires the introduction of the residence time
distribution g(r). As a first approximation, the region of interest
can be assumed to behave like a continuous stirred tanfc reactor (CSTR)
or- a plug flow reactor (PFR) at steady, state. Therefore, we reduce
the p sources to a unique source and ignore the decay factor depen-
dency on the spatial distribution of the sources of emissions.
II. 3.1 CSTR
In such reactors, upon introduction, the reaction immedi-
ately reaches uniform concentrations determined by the reactor
volume and feed flow rate. The concentration of the species in the
stream leaving the reactor are equal to the reactor concentrations
(Carberry, 1976). The exponential residence time distribution is
given by:
i>
g(T) - £ exp(- J)
where 8, the average residence time, is given by
where V is the reactor volume and Q the volumetric feed rate. Sub-
stituting in equation (II-5) and integrating yields to:
where a. is the decay factor specific to species i and p. is the
mass concentration of species i measured at the point of emission.
12 .
-------
II.3.2 PFR
In this type of reactor, all entering molecules have the
same residence time which is given by the ratio of the reactor
volume to the volumetric feed rate:
13
-------
III. SOURCE CONCENTRATIONS OF PAH
As discussed in the preceding section, source resolution
through the chemical element balance method is based on the knowledge
of two types of data: ambient measurements at the receptor sites
(p.*s) and source concentration matrix (c,,'s) of emissions measured
at the point of emission. Furthermore, since the method cannot
account for each source but instead types of sources (automobile
emissions, coke production and so on), it is necessar/ to develop a
source concentration matrix which is statistically representative
of the types of sources under study. For this purpose, a literature
review of PAH emissions from major combustion sources and an evalua-
tion of agreement and discrepancies between investigators is pre-
sented in this section. "
III.l Coal Combustion
Junk and Ford (1980) reviewed the literature on organic emis-
sions, including PAH, from coal combustion, waste incineration and
coal/refuse combustion. In the case of coal combustion emissions,
sources will be separated into home heating and power generation.
PAH emissions are higher when the combustion is incomplete
and/or non uniform (NAS, 1972). Thus, emissions from coal-fired
power plants are expected to be less than those from residential
furnaces, for the same amount of fuel burned. Hangebrauck et al.«
(1967) have studied PAH emissions from different units depending on
the unit size, its operating conditions, fuel and design parameters.
14
-------
They showed that emissions from hand-fired residential furnaces were
much larger than emissions froa any other type of unit. This result
is in agreement with the observations of Natusch (1976). Ratios
of emissions with respect to coronene presented by Hangebrauck et al.,
were wot reproducible even within the same type of unit probably
because of the many variables affecting the PAH formation. Never-
theless, for all types of units, emission rates of higher molecular
weights (greater or equal to 252 amu) were less important than
those of lower molecular weight PAH. Similar patterns were reported
by Lee et al. (1977). Recent measurements (Bennett et al., 1979)
on emissions from three coal-fired power plants showed that emissions
of PAH heavier than 226 amu were generally smaller than the minimum
detection limit (0.1-0.2 ng/Na ). Natusch (1978) suggested that
most of the measured PAH in the stack were still present! as vapors
since their mass concentrations were, in the aerosol phase, larger in
tho vicinity of the coal-fired power plant than in the stack itself.
Concentrations measured in the plume at a distance ranging from zero
to five miles away from the stack were approximately two orders of
magnitude higher than concentrations reported by Bennett et al. for
PAH heavier than 226 amu.
III.2 Coke Production
Particulate and gaseous phase emissions from coke ovens have
been measured for their PAH content (BjjJrseth et al., 1978). Samples
were taken on top of the coke oven batteries in Spring and Fall 1976
before and after reconstruction of a Norwegian coke plant. \ rignifi-
15
-------
cant mass fraction of the phenanthrene (178 arau) and anthracene
(178 amu) s any led were shown to be present in the gas phase f'vS-S
Traces of BaP and BeP (^0-2%) were also found in the gas phase.
Table 1 shows a comparison of Bjjirseth's emission ratios with
respect to perylene (PER), with the results of Btoddin et si. (1977)
and Lao et al. (1975). Table 1 also shows some aspects of the samp-
ling conditions used by the different investigators. The fit be-
tween BjfSrseth and Broddin is good while discrepancies with the
results of Lao are noted for the BeP, ANTH and BghiP perylene ratios.
III.3 Incineration
Emissions are likely to depend on the type of material burned.
Hangebrauck et al. (1967) examined PAH emissions from two municipal
and two commercial incinerators burning wastes from households,
grocery stores and restaurants. Mass fractions of emissions were
one to two orders of magnitude higher for the smaller units (commer-
cial incinerators) than for the larger ones. Pyrene, fluoranthene
and phenanthrene were predominant in all types of units. The same
result was observed by BjfJrseth et al. (1978) and Lao et al. (1975)
in the case of coke oven emissions. However, a municipal incinera-
tor burning an average composition by weight of 32% paper, 18% fine
dust and cinder, 15% vegetable and putrescibles, 9% metal, 8% tex-
tile and wood, 7.5% glass and ceramic and 5% plastic and rubber, was
shown not to emit significantly hi'her emission rates of fluoranthene
and pyrene than other PAH reported (Davies et al., 1976).
16
-------
Table 1. Comparison of PAH to peiylene emission ratios
from coke production.*1
Reference PAH/PER
BaP BeP BFL ANTH BghiP COR
BjfSrseth 5.0* 2.9 3.6 1.3
et al.
(1978)
Broddin
et al.
(1977)
Lao
et al.
(1975)
4.6 2.2 2.9 1.8
10.7
1.4
3.0
3.1
3.7 0.1 3.8 0.1
5.5 0.0 7.9 0.0
3.3 0.0 3.6 0.0
3.3 0.0 4.1 0.0
1.5
0.6
1.2
0.0
1.4
1.1
Sampling conditions;
sampling point, type of
filter used, method of
extraction and analysis.
0.3-0.5 m above the floor,
on top of the coke oven
batteries. Acropore fil-
ters. Soxhlet extraction.
GC/MS analysis.
8 n away from the top of
the coke oven batteries.
Anderson cascade impactor
equipped with glass fiber
filters. Soxhlet extrac-
tion. GC/MS analysis.
No information on the
sampling point. Glass
fiber filters or silver
membrane filters. Soxhlet
extraction. GC/UV, GC/MS,
GC/FID analysis.
aPAH lighter than 252 amu are not included in this table. Perylene
was chosen for reference since it was reported by all investigators.
Samples were taken before modification in coka plant.
^Samples were taken after modification.
17
-------
III.4 Wood Combustion
Emissions resulting fron the combustion of wood and peat in
a hot water boiler were measured for their PAH composition (Alsberj
and Stenberg, 1979). Peat combustion emissions were approximately
one order of magnitude higher than wood combustion emissions due to
a less complete and less uniform combustion. As shown in Table 2,
emission ratios of selected PAH with respect to BghiP were
generally found similar for both fuels. Phenanthrene, pyrene and
fluoranthene were predominant while only traces of coronene were
detected.
Table 2. Selected PAH to BghiP emission ratios fron wood
and peat combustion. Adapted from Alsberg and
Stenberg (1978).a
PAH
" BghiP PHT ANT FLT PYR BaA BaP BeP PER INP
Fuel
Wood6
Peat
55
45
4.2 23.5 19 3.4 1.4 2.7 0.2 0.8
3.7 19.4 16 4.7 0.4 2.4 0.2 1.5
aData on coronene were not usable. BghiP (276 amu) was chosen for ref-
erence. BghiP is believed to be one of the least reactive PAH (see
Table 5).
Arithmetic mean over 3 experiments.
III.5 Open Burning
Several types of open burning have been surveyed for their
PAH emissions: municipal refuse, landscape refuse, automobile coi$-
ponents (Hangebrauck et al., 1967; Cerstle and Kemnitz, 1967), auto-
mobile tires (Hangebrauck et al., 1967). Both investigators used
18
-------
the same sampling train, and emission ratios with respect to a species
of reference are in excellent agreement. Pyrene and fluoranthene
emissions were predominant. Emissions of anthracene and phenanthrene
were surprisingly low when compared to BaP emissions. This does not
seen to be the pattern for emissions from other sources.
Forest fires can become a locally important source of PAH.
MacMahon and Tsoukalas (1978), simulating a pine needle fire, have
reported emissions factors for PAH species from anthracene to BghiP.
Coronene was not detected. The mass fractions of the 18 PAH reported
depended on the type of fire (i.e. with the wind or against the wind)
and on the mass concentration of fuel burned. Emission ratios with
respect to a reference species were not reproducible.
III.6 Gasoline-powered Cars
Although diesel emissions per mile, may be larger, of all
mobile sources gasoline-powered vehicles are likely to be the pre-
dominant source of PAH emissions and to contribute a significant
amount of the PAH found in urban atmospheres. Emissions depend on
many variables and the following presentation will differentiate
vehicle characteristic effects from fuel composition effects (HAS,
1972).
Vehicle characteristic effects. When the mixture is rich
(i.e. cold start) combustion is less complete than during lean
combustion and PAH emissions are expected to be higher. Williams
and Swarin (1979) showed that average emission rates from seven
gasoline, non-catalyst cars were more than six times higher in cold
19
-------
starts. In the vicinity of the stoichiometric air to fuel ratio
(A/F « 14.5) PAH emissions were shown to be constant (Begeman and
Co^ucd, 1970; Pedersen et al., 1980). Begeman and Colucci also
reported thirty times greater emissions of BaA and BaP at A/F " 10
than at 14. The BaA to BaP emission ratio varied from 3.1 to 3.7,
showing a maximum at A/F ™ 14, when the air to fuel ratio was kept
between 12 and 16. For the reasons stated above, PAH emissions vary
according to the driving mode. Begeman and Colucci (1970) showed
that BaP emissions were at a minimum of 3.3 pg/g of tar for a 30 mph
cruise, reached 16 Pg/g for a 50 mph cruise and peaked at 21.4 pg/g
during acceleration. The BaA to BaP emission ratios were respec-
tively 7, 5.4 and 3.9. Over-consumption of oil (9.1 qt/1000 miles
instead of 0.26 qt/1000 miles) based on a 2000 mile run on a 1968
(EM) vehicle, was responsible for about a tenfold increase in BaP
emissions while the BaA to BaP emission ratios decreased respectively
from 4.5 and 5.6 to 1.6 and 0.9. In the same fashion, Begeman and
Colucci (1970) showed that, for a city driving cycle, an eightfold
increase in the oil consumption (200 miles/qt instead of 1600 miles/
qt) induced a tenfold increase in BaP emissions and an eightfold
increase in BaA emissions. Statistically, mileage did not affect
the BaA and BeP to BaP emission ratios fron a 1968 (EM) and a 1970
(EM) vehicle (Gross, 1974).
Fuel composition effects. Fuel lead content was shown to
have no significant effect on PAH emissions (Gross, 1971; Pedersen
et al., 1980). It is generally believed that an increase in the
20
-------
fuel PAH content increases PAH emissions (Pedersen et al., 1980;
Newhall et al., 1973; Gross, 1972) but seems not to affedt'signifi-
cantly PAH emission profiles (emission ratios with respect to a
species of reference). A fuel BaP content increase of 3 ppa caused
a 32\ increase in BaA emissions and an &\ increase in BaP emissions
from a 1970 (EM) vehicle (Gross, 1972). Newhall et al. (1973)
testing 1969, 1970 and 1972 vehicles did not show statistically
significant differences in the BaA to BaP emission ratio when
using a fuel with no PAH and a fuel containing 0.8 and 2.2 ppm
of BaP and BaA, respectively. Similarly, at a 30 mph cruise,
the BaA to BaP emission ratio varied respectively froa 7.3 to
6.8, at constant BaP fuel content (1.1-1.2 ppm), when the BaA
fuel content was changed from zero to 1.3 ppm (Begeman and Colucci,
•'1970). Newhall et al. (1973) showed that increasing the C,-C0
o o
fuel aromatic content from 8\ to 38.1% increased BaA and BaP
emissions from 1969, 1970 and 1972 vehicles by a factor of one to
3.5 depending on the vehicle while the BaA and BaP emission ratios,
within the same vehicle, were not significantly affected. However,
the type of fuel aromatic fraction, at a 40 vol.% constant aromatic
content, was shown to strongly influence PAH emission profiles;
for example, the BaA to BghiP emission ratios were approximately
four and nine for an o-xylene and a C.+C.- aromatic fraction,
respectively (Pedersen et al., 1980).
Grimmer and Hildebrandt (1975) investigated PAH emissions
froa 100 passenger cars registered since October 1, 1971, in West
Germany. A detailed analysis of their data, which were those used
21
-------
for automobile emission profiles in our source resolution, is pre-
sented in section IV.1.
III.7 Diesel-powered Cars
The impact of a possible dieselization of the United States
passenger vehicle fleet is currently of interest. At the present
time, emissions from diesel engines are unlikely to account for a
significant amount of atmospheric PAH found in urban areas. A 1975
inventory of BaP emissions in the Los Angeles atmosphere (Abrott
et al., 1978) pointed out that the estimated daily load from diesel
engines was less than 61 of the total emissions listed. Gasoline-
powered vehicles accounted for more than 84t.
BaP emission rates were found to be similar in cold starts
for gasoline-noncatalyst cars and diesel cars (Williams et al.,
1979). They averaged 5.1 yg/mile for seven gasoline-noncatalyst
cars and 3.9 pg/mile for one 1978 diesel car. The diesel car was
a 1978, 5.7 L, 8 cylinders burning a diesel fuel No. 2 containing
0.11 ppm of BaP. The noncatalyst, gasoline cars were burning an
unleaded fuel containing 1.04 ppm BaP. The average hot start
emission rates were higher for the 1978 diesel car than for the
seven gasoline-noncatalyst cars, 1.9 yg/aile and 0.8 pg/mile, respec-
tively.
The BaA and BaP emission rates from one diesel engine (a six
cylinder Mack ENDT-675, four cycle, turbo charged engine with direct
injection) burning a composite of ten commercial N°2 diesel fuels,
were determined at no load, half load and full load (Brickmeyer and
22'. I
-------
Spindt, 1978). Emission rates of BaP, averaged over five runs,
varied fron 8 to 25 ugAg of fuel burned. They were minimua at
half load. The BaA to BaP emission ratios were 2.7, 3.9 and 1.1 at
idle, half load and full load, respectively.
Kaschani (1979) presented a gas chronatogran of PAH in
diesel exhaust gas showing that pyrene, fluoranthene and phenanthrene
were the predominant species. Coronene, BaP and perylene peaks were
notably smaller than BeP, BghiP and chrysene peaks.
23.
-------
IV. SOURCE RESOLUTION ANALYSIS FOR LA PAH:
RESULTS AND DISCUSSION
IV.1 Automobile Emissions—PAH Decay Factors
In the preceding section, PAH emissions and emission profiles
froa gasoline-powered cars were shown to depend on many variables.
In order to compute a PAH emission profile representative of auto-
mobile emissions, a survey type of study was needed.
Grimmer and Hildebrandt (1975) surveyed PAH emissions fron
100 passenger cars representative of the German automobile fleet
(twenty most common models and five cars per model). The selec-
tion of the models chosen'/was made by the percentage they held of
the total number of first registrations from January to May 1972.
These 20 models represent 67.07 percent of all newly registered
passenger cars." All vehicles chosen were registered since October
1, 1971. They were supplied by dealers, car rentals or private
owners. These cars were tested simulating a city driving mode
(Europa-Test). They were submitted to a technical control before
being tested. In particular, "compression, spark plugs, ignition
timing and contact angles, CO and HC content during idling and
tightness of the exhaust system were checked. If necessary, igni-
tion timing and contact angles were corrected. The CO concentration,
when too high, was regularly reduced to 2 to 3 vol.%, or, when too
low, increased to these values." The crankcase oil was not changed
and the code of fuel used was ERF/G1. No information was given on
24-
-------
the characteristics of the various models tested. Reproducibility
of the chemical analysis was checked by analyzing 10 times the sane
exhaust condensate. The variation coefficients of single PAH were
within 2.8 to 5.8\.
A statistical evaluation of PAH to coronene emission ratios
was performed for PAH molecular weights starting from 252 amu since
these emission profiles are to be compared with ambient PAH concen-
trations. The choice of 252 amu has been discussed in the intro-
ductory section. Our analysis included BaP, BeP, BFL, ANTH, BghiP
and IMP. Emission ratios with respect to coronene are shown in
Table 3. The standard deviation of each PAH to coronene ratio
was* within 15* to 41% of the arithmetic mean of emission ratios
from the 20 models tested. This result suggests that PAH emission
ratios are not strongly dependent on the model tested. In support
of this last statement, results were checked for consistency with
atmospheric samples taken in Cincinnati (Sawicki, 1962) and Los
Angeles (Gordon, 1980a) at sites which are likely to be dominated by
automobile emissions (Table 3). In the analysis which follows,
the arithmetic mean of emission profile calculated from the data of
Grimmer and Hildebrandt was assumed typical of automobile emissions
in Los Angeles over the period 1970-75. It was further assumed
that the deviation between the data of Gordon (1980a) and Grimmer
et al. (1975) is due to a first order decay in all species of in-
terest. The assumption of first order decay is supported by the
results of Katz et al. (1979) and Lane and Katz (1977). They per-
formed experiments on simulated atmospheric degradation of BaP, BeP,
25
-------
Table 3. PAH Pattern Comparisons
N>
O%
Reference
Sampling
Point
PAH
COR
BaP
BeP
BFL
INP
BghiP
ANTH
Grimmer and
Hildebrandt
(19730
Tailpipe of
cars 0
0.50±0.14
0.51±0.12
0.55±0.13
0.40±0.09
1.58±0.24
0.2910.12
Gordon
(1980a)
•Site 1 a
(LA)
0.27±0.11 b
0.23±0.06 o
0.23±0.06 d
0.4810.08
0.5610.09
0.5310.05
0.54±0.07 /
0.70±0.17 g
1.3210.24
1.21±0.17
1.39±0.21
0.05±0.03
0.06±0.02
0.07±0.02
Sawicki
(1962)
Tailpipe of
cars
0.71
0.55
--
--
3.0
0.24
Sawicki
(1962)
Cincinnati
downtown
garage)
(January)
1.3
0.78
--
—
2.08
0.32
Sawicki
(1962)
Cincinnati
Auto Safety
Lane
(January)
0.41
0.62
—
~
2.62
0.52
aSite 1 (see Table 9) was near a freeway junction, nearest freeway 0.12 mile, with heavy traffic
and was not close to any known stationary coubustion source at the time of sampling.
' Arithmetic mean within 68% of 12 samples taken on a monthly basis from 6/71 to 6/72, 12/72 to
12/73 and 2/74 to 1/75, respectively.
*Arithmetic mean within 68%.
•'Samples were taken on a monthly basis from 12/72 to 12/73.
^Samples were taken on a monthly basis from 2/74 to 1/75.
-------
BkF and BbF observing first order decay laws. Falk et al. (1956)
observed a 16% degradation of coronene after a four hour irradia- .
tion under a strongly oxidizing synthetic smog (^30 ppn oxidant).
Light intensity did not seen to affect the stability of coronene
(Barofsky and Baun, 1976; Falk et al., 1956). Based on these
results, coronene was assumed to be a stable species in our analysis.
Computing the ratios of Gordon's to Grimmer1s emission pro-
files with respect to coronene provided decay factors for each PAH.
Indeno (1,2,3-cd) pyrene, (IMP), was not included since its ratio
to coronene at the receptor site was greater than the same ratio
at the point of emission. This discrepancy could not be explained.
Decay factors were first calculated on a quarter basis for
samples taken in downtown Los Angeles from June 1971 to June 1972,
December 1972.to December 1973, and February 1974 to January 1975 o
Due to stronger photochemical conditions, PAH decay factors were
expected to be higher during the second and third quarters of the
year. Since that was not found to be the case after statistical
analysis of the data, the data were randomized and solved for their
arithmetic mean. A propagation of errors analysis was performed
assuming uncorrelated errors (Bevington, 1969). Results showed a
stronger reactivity for BaP and ANTH than for BeP, BghiP and BFL
(Table 4).
An eleven hour average atmospheric residence time based on
lead concentration and morning inversion height was computed (Ap-
pendix B). PAH were assumed not to react after deposition in the
filter media since they were no longer exposed to solar radia-
27
-------
Table 4. PAH Decay Factors0
PAH
BaP
BeP
BFL
BghiP
ANTH
Decay Factor
0.4810.21
1.0410.292*
0.98±0.26
0.83±0.19
0.2110.11
Arithmetic mean within 68* confidence.
BeP was assumed a stable species in the
source resolution analysis.
28
-------
tion which probably accounts for most of their removal (NAS, 1972).
This assumption is supported by the fact that Miguel and Friedlander
(1978), sampling with glass fiber filters and a low pressure impactor
during 72 hours froa October 1976 to March 1977. and Gordon (1980a)
operating with high volume samplers for a three week period from
August 1974 to January 1975, have found similar BaP to coronene ratios
for nearby sampling sites (0.14 to 0.22 and 0.26, respectively).
Reaction rate constants were calculated assuming that atmos-
pheric residence times were equal to those for a continuous stirred
tank reactor (CSTR) or a plug flow reactor (PFR). As expected, reac-
tivity is greater for the CSTR model than the PFR; the difference
being less marked for less reactive species such as BghiP and BFL than
for more reactive species such as BaP and ANTH.
Results were compared with data reported 4^.n the literature for
PAH degradation in laboratory experiments or simulated atmospheres
(Table 5). The df.ta shown are not self-consistent. For example,
since PAH react readily with ozone or other oxidant (NAS, 1972), BaP
reactivity is expected to increase with the oxidant content of the
reactants. However, this is not thj case when the results of Tebbens
et al. (1971) and Falk et al. (19S6) are compared with each other.
Similarly, it is believed that PAH degradation is .less important in
the absence of irradiation. But irradiation, all other parameters
being equal, did not affect significantly the reactivity of BaP and
BeP while a tenfold increase in reactivity was observed in the cas"e
of BbF and BkF (Katz et al., 1979; Lane and Katz, 1977). On the
other hand, Korfmacher et al. (1980) carrying out experiments on
29
-------
Table 5. Comparison of rates of degradation of
selected PAH. - —
Refercaca
Fall et al.
(19S«)
CaraUune and
Urohata (1SHS2
TeOb«ni rt al.
(196«)
rebbenx et al.
(1971)
i
larofsky and
Uua (1<76)
Lao* and Call*
(1977)
latl et al.*
(1979)
Ail
vort
Syitea
Static
Ijrstea
Static
tystea
Flo*
reactor
nF 0.01S
ML* 0.020
UP 0.0*1
AKTH O.S42
UP 0.091
UP •
tn. 0.002
llhiP O.Olt
AXTO 0.142
UP 0.047
UP 0
•n. 0.002
IlhlP 0.017
'llat* coDjtaatl shown were obtained froa the correspond in| FAH nalf-livct reported
authors. Other rate constant* were calculated assualnf first order decay.
by th*
30
-------
the degradation of PAH adsorbed onto the coal fly ash concluded that
their reactivity was dependent on the type of adsorbent. Thus, reac-
tion rates reported in Table 5 should be compared with caution.
IV.2 Refinery Emissions
The data used are those of Gordon and Bryan (1973) and Gordon
(1980a). Again, it was assumed that site 1 (Gordon and Bryan, 1973)
was dominated by automobile emissions. Site 3 (Gordon and Bryan,
1973) was near to and downwind froa a concentration of petroleun
refineries and chemical plants. The nearest freeway was within 1.27
miles and the traffic was moderate. Site 3 was not close to any
other known stationary combustion source at the time of sampling,
but it was assumed to receive contributions from refineries and
automobiles.
o
Gordon and Bryan (1973) measured the lead concentrations at
sites 1 and 3 from June 1971 to June 1972. Automobiles were assumed
to account for the lead concentrations found at these sites. Thus,
lead was used to trace automobile emissions at site 3 since site 1
was assumed to be totally dominated by automobile emissions. Re-
finery emissions were obtained by computing the difference between
each PAH concentration measured on a monthly basis from June 1971 to
June 1972, at site 3 (Gordon, 1980a), and the corresponding calculated
PAH concentration resulting from automobile emissions at that site
for the same period of time. Table 6 shows profiles of emissions m
with respect to BeP, from automobiles and refineries. It consti-
tutes the source concentration matrix that was used to solve the
31
-------
Table 6. Source Concentration Matrix
Source of Emissions PAH/;BeP !! :
BFL BaP BghiP ANTH . COR
Automobiles
(Grimmer et al., 1975) 1.08±0.36 0.98±0.36 3.10±0.87 0.57±0.27 1.96±0.46
a
Refineries" 1.43 3.85 2.46 2.12 0
aAll PAH were corrected for their decay.
b
Based on small coronene concentrations.
.23"
-------
chemical species balance. BeP is believed to be a stable species
(see Table 4). It was used as the reference species, instead of
coronene, since coronene emissions from refineries are small and the
*
accuracy of our result is uncertain in this case. However, the
calculated coronene emissions from refineries are generally one
order of magnitude smaller than other PAH emissions under considera-
tion. And this last result is in agreement with findings of Bennett
et al. (1979) who investigated PAH emissions from an oil fired
power plant burning a Venezuelan residual fuel.
Lead concentrations measured at sites 1 and 3 from December
1972 to December 1973, were also reported by Gordon (1980a). An
alternative approach, in order to compute the refinery emission pro-
files, was to average the profiles obtained from the data of June
1971 to June 1972 and December 1972 to December 1973. Results are
shown in Appendix C. Emission profiles are similar to those shown
in Table 6. Furthermore, when the chemical species balance is
solved using the new set of data for refinery emission profiles, it
yields almost identical results to those shown in Tables 7 and 8.
This approach was not taken into consideration since it provided
negative refinery emissions in the case of coronene.
Another type of analysis was performed by Gordon and Bryan
(1973). They showed that coronene correlated well with traffic den-
sity. And assuming that automobiles were responsible for the con-
centration of coronene measured at sites 1 and 2 (see Table 9 for
the site locations), they were able to determine the PAH concentra-
tions, resulting from non-automobile emissions, measured at site 3.
33
-------
Again, this method yielded almost identical results to those shown
in Tables 7 and 8, but was not selected since it assumed that refin-
eries did not emit coronene.
IV.3 Source Resolution
The chemical species balance was solved for 13 sites in the
Los Angeles basin. It was assumed that automobiles and refineries
were the only sources to contribute significantly to the coronene,
BaP, BeP. BFL, ANTH and BghiP measured at the stations. A set of
four, five or six equations, depending on how many species were mea-
sured at the different receptor sites, was solved according to the
least-squares fit technique. In all cases, errors in the measure-
ments were assumed equal since they were not known (see section II.1),
Uncertainties on the elements of the source concentration matrix
were ignored since they could not be determined in the case of re-
finery emissions. The atmospheric concentration vectors were based
on data reported by Henderson et al. (1975) and Gordon (1980a).
Results are shown in Tables 7 and 8 together with the atmospheric
concentration vectors. Table 9 shows the locations of the sampling
sites.
As shown in Table 7, refinery contributions were found to
range between zero and 13* at site 1, depending on the PAH, the
year or the period of year that samples were taken. In fact, there
should have been no contributions from refineries at that site, as.
a result of the calculations, since one of the constraints of our
analysis was to assume that site 1 received contributions fron auto-
34
-------
mobiles only. As stated previously (p. 27), avera£e decay factors,
for each PAH under.study, were found.by computing.the ratios of
Gordon's to Grimmcr's emission profiles. Gordon's emission pro-
files were the arithmetic mean over samples taken froa June 1971 to
June 1972, December 1972 to December 1973, February 1974 to January
1975. They deviated from their three year mean value when computed
on a quarterly or yearly basis. These deviations are responsible
for the residual refinery contributions found at site 1. The results
of the calculation indicate that the PAH concentrations at site 1
were indeed dominated by automobile emissions.
Site 2, near site 1, followed a similar pattern. The refinery
contributions peaked at site 3 where the fractions of BaP, BFL and
ANTH due to refinery emissions varied from 0.73 to 0.94. This
result is in agreement with the locations of refineries (Bryan, 1974).
Sites 4, 10 and 13 also received significant amounts of PAH emitted
by refineries. The BaP and ANTH mass fractions due to refineries
at these sites, varied from 0.15 to 0.27. Other stations, not in
the vicinity of refineries, were shown to be automobile dominated.
Contributions for sites 1, 2, 3 and 4 were calculated on a yearly
basis and for the second and third quarters of the corresponding years.
Results showed a similar pattern suggesting that home heating is not
an important source of PAH, at least in Los Angeles.
Inconsistent results were sometimes found (i.e. slightly nega-
tive refinery contributions) due to errors in the source concentrf-
tion matrix, decay factors and/or ambient concentration vectors.
In these cases, errors in ambient concentration vectors are most
35
-------
likely responsible for the inconsistencies since these were associ-
ated with the smaller concentrations. A tentative solution was per-
formed on a monthly basis but did not yield satisfactory results.
In this case, the atmospheric concentration vectors, based on a
single measurement, may have generated errors in the source resolu-
tion. The method also failed to yield reasonable results when the
source concentration vectors were based on the data of Gordon (1976).
Specifically, significant contributions of refineries were found
for regions 7 and 8 (see Table 10 for the region locations) whereas
no contribution of refineries was found for regions 9, 10 and 12.
The ratios of lead to PAH from automobile emissions, computed
whenever lead concentrations were available, are also shown in Tables
7 and 8. Their arithmetic means within the 68% confidence are shown
in Table 11 and are in agreement with emissions profiles computed
from the data of Grimmer and Hildebrandt. Table 11 also shows the
average deviations of the calculated PAH concentrations from the mea-
sured concentrations at the different receptor sites. These devia-
tions can be regarded as a measure of the accuracy of our results.
They are within -11 to +7 per cent of the measured concentrations.
36
-------
Table 7.
Contribution, of automobiles and refineries to selected airborne PAH
at various locations in Los Angeles.
trt
StCtlM
a
1
1
1
1
1
2
2
]
T««r/
P*ria4
Jua«
to
Oct.
JVMM
t*
J«M
DM.
t*
Doe.
*y
t*
Oct.
Foo.
to
JM.
JMM
to
Oct.
Jt»*
to
JMO
J«M
to
Oct.
71
71
71
72
72
7J
74
74
74
71
71
71
71
71
71
71
PAH Automobile
Contribution*
ni7.S
COR
liP
MP
IfhlP
ANTH
aw
MP
MP
IfhlP
ANTH
COR
MP
MP
ȤMP
ANTH
tfl
CM
I«P
MP
• iMP
ANTH
CM
MP
MP
IfhlP
ANTH
aw
MP
MP
IfhlP
ANTH
CM
MP
MP
IfhlP
ANTH
aw
MP
MP
IfhlP
ANTH
4. SO
1.01
2.10
$.90
0.21
4.1*
1.17
2.41
*.I7
O.KJ
4.57
1.10
2.11
$.99
0.21
2.4*
2.S2
0.60
1.2*
S.M
6.1$
4.0*
0.97
2.07
S.S2
0.2S
2.1*
O.S2
1.10
2. IS
0.11
2.M
0.6*
1.4*
1.71
0.17
1.01
0.2*
O.SS
1.42
0.07
Ud/PAH
b
»io-J
..
..
..
.-
-•
1.10
2.20
2.16
0.70
1.79
0.91
1.17
1.11
O.SI
1.21
1.70
.»
••
..
• •
...
•-
-.
"
--
•-
0.17
1.11
1.1$
0.41
2.17
—
H.fln.17
Contributions
nf/n
0.02(0)
0.16(1)
0.01(7)
0.17(1)
0.01(9)
0.01(1)
0.09(0)
0.04(9)
0.10(0)
0.02(2)
0.01(2)
0.09(6)
0.05(2)
0.10(6)
0.07(1)
0.07(1)
0.00(6)
0.04(9)
0.02(7)
0.01(4)
0.01(2)
0.00(4)
0.01(6)
0.02(0)
0.04(0)
0.00(9)
0.00(4)
0.01(9)
0.02(1)
0.04(1)
0.00(9)
0.00(0)
0.00(2)
0.00(1)
0.00(2)
0.00(0)
0.29
3.30
1.2$
2.$$
o.$s
Totil (Ulcu-
culitcd FAN
Concentration*
nt/m*
2
0
1
S
0
4
1
2
S
0
2
0
1
2
0
. 2
0
1
I
0
1
2
1
1
0
.12
.24
.19
.01
.12
.17
.26
.51
.47
.12
.51
.20
.31
.10
.10
.SI
.51
.65
.12
.IS
.1*
.0*
.01
.09
.34 •
.2*
.16
.$*
.12
.17
.14
.16
.69
.46
.71
.17
.17
.56
.M
.97
.62
Meiiurwi
Cone mint Ion*
ni/»J
4.
1.
2.
6.
0.
4.
1.
2.
6.
0.
4.
1.
2.
S.
0.
2.
2.
0.
1.
I.
0.
J.
0.
1.
t.
0.
2.
0.
1.
1.
0.
J.
0.
1.
1.
0.
1.
2.
1.
I.
0.
6*
17
24
01
21
17
M
11
51
24
12
20
SI
II
11
S9
SI
61
19
10
If
77
91
97
47
27
19
II
1*
|j
12
94
74
44
71
14
61
bl
II
77
S*
\ D«vUtl««
fro.
MMMTOMHtl
• 1
. 9
• 4
0
• 19
0
- 7
« *
. 1
•11
^
«
-
•
»
•1
•
•
•
. }
• 4
- 1
• 1
- *
* |
»17
. j
- 7
* 1
» 1
•21
.1*
• 4
. 4
• 1
• 11
-------
Table 7 (continued)
tutlon
PAH
Autonobll*
Contribution* b
R»fl»«rjr
Contributions
mt/i
Tot«l Calcu-
cuUtod PAH
Concentration*
zlO
Motiuro*1
Concentration*
HI/.'
% PvrltUoo.
,,
s
1
4
4
JUM 71
to
Jwto 72
Hif 71
to
Oct. 71
Dec. 72
to
Doc. 71
JUM n'
to
Oct. n
Juno 71
to
Juno 72-
COR
UP
UP
IfhlP
AXTH
COR
UP
IfhlP
AKTH
in.
COR •
UP
UP
IfhlP
AM1M
•PL
COR
UP
UP
IfhlP
AN1X
COR
UP
ItP
IfhlP
AKTH
1.71
0.43
0.91
2.33
0.11
0.49
0.12
0.2S
0.64
0.03
0.26
1.49
0.34
0.76
1.95
0.09
O.M
0.11(0)
0.04(3)
0.09(2)
0.23(6)
0.01(1)
0.11(1) ~
0.04(5)
0.09(6)
0.24(7)
0.01(1)
1.11 0
2.21 2
2.16 1
0.70 2
3.79 0
2.47
4.14
4.M
1.57
1.64
4.41
1.37
2.71
2.61
0.17
4.73
2.41 .
-
-
2.66
S.32
5.7.1
1.61
9.SS
.33
.67
.45
.95
.64
.25
.03
.10
.25
.49
.S4
.X
.K
.12
.10
.67
.11
.00(2)
.01(3)
.00(7)
.01(5)
.00(3)
.00(1)
.01(0)
.00(5)
.01(1)
.00(2)
2.11
3.10
2.36
S.2I
0.7S
0.74
2.15
1.3S
2.19
0.52
1.10
1.14
3.16
2.21
S.OS
0.76
2.91
O.ll(J)
0.05(6)
0.09(9)
0.25(1)
0.01(4)
0.-1IC9)
0.05(5)
0.10(1)
0.25(1)
0.01(3)
2.12
3.14
2.29
S.30
0.72
1.01
2.23
1.76
2.54
0.46
2.20
2.12
3.20
2.60
4.64
0.74
1.41
0.204
0.064
0.104
0.234
0.012
0.211
0.063
0.109
0.240
0.011
0
- 1
* 3
0
• 4
•27
- 4
•23
»14
• 13
•If
-11
. 1
• 12
» 9 '
« 1
-19
•11
-11
- I
» 7
»17
-10
-11
- 7
» 1
.11
"l*o T»blo 9 for tko ilto location*.
*AJ1 PAH *orocoTTOcto4 for tbolr 4oc«r.
*D*vUtloiu trtm BOuwroomtti u» «uo to cooputotloMl orror *»i to tko f»et t]««t ttttj focten «ro W*o4l
uaplo* takoa froa JHM 71 to JMM 72. DocooJ>or 72 to Doeooi>or 71 nA Pobruorf 74 to Jonuor/ 71.
AofMt Itn (licioyloto *»t*). -
-------
Table 8. Contributions of automiblies and refineries to selected
airborne PAH at various locations in Los Angeles.
»title*
•
PAH
AutMobll*
Contribution
Lud/PAH
a
V5
zlO
-I
Contribution*
Bt/«J
Total Calcu-
culatW PAH
Concentration*
Cmcmtrttie
ni/«J
f ro-
IJ
M
lit
NU
11
t4/lt
10
11
con
IIP
MP
IfhlP
CO*
I»P
UP
IfhiP
con
UP
UP
U>1P
CM
UP
UP
»fMP
CM
ItP
UP
IfhlP
CM
UP
UP
»jhlP
CM
UP
UP
IfhlP
CM
UP
UP
IfhiP
0.*4
0.1S
0.32
0.11
0.97
0.24
0.49
1.21
0.77
0.11
O.J9
1.01
0.79
0.19
.40
.OS
.01
.24
.SI
.M
.It
.21
.44
.14
.70
.17
O.M
0.92
0.91
0.24
O.SO
1.29
l.SS
3.09
3.09
0.91
1.14
2.27
2.27
0.72
1.S2
3.01
3.00
O.M
1.11
2.33
2.31
0.74
1.31
2.S9
2.S9
0.12
1.47
2.91
2. It
0.92
1.70
3.40
3.31
1.07
1.12
2.24
2.20
0.71
-0.00(3)
-0.02(4)
-0.01(3)
-0.02(7)
0.00(2)
0.01(6)
0.00(8)
0.02(7)
-0.00(3)
-0.02(4)
-0.01(3)
-0.02(7}
-0.00(0)
-0.00(2)
-0.00(1)
-0.00(2)
0.00(1)
0.00(1)
0.00(4)
0.00(9)
-0.00(1)
-0.00(4)
-0.00(2)
-0.00(4)
0.00(1)
0.04(4)
0.03(4)
0.07(1)
0.01(1)
0.01(1)
0.04(1)
0.09(7)
0.64
0.13
0.31
O.M
0.97
0.26
O.SO
1.31
0.77
0.16
0.31
0.91
.79
.19
.40
.OS
.01
.25
.51
.3S
.16
.21
.44
.14
0.71
0.23
0.39
0.99
0.99
0.33
o.ss
1.39
0.74
0.11
0.31
0.74
1.03
0.26
0.41
1.2t
0.19
0.11
0.40
0.19
0.90
0.21
0.37
0.91
1.17
0.29
0.47
1.2S
1.00
0.21
0.4S
1.04
O.M
0.2S
0.3t
0.94
1.07
0.34
0.64
1.32
-;«
; -11
0
« i
• 4
0
» 4
• 4
-1>
-11
- 1
«!•
•12
-10
• •
• 7
•14
•14
» f
» 1
• 14
• 1
• 2
• !•
• 11
- 1
» 1
« (
• 7
• 1
-14
* »
'All PAH v«r« c*rnctW ft* tfc.lr
•t U. (1971). Tb« Bate dtowii »r» tk« "fitted BM* nl»»* ef four *Mklf
T*kl« J fw tk« lit* l*e*ti«w.
-------
Table 9v Los Angeles Sampling Sites with Freeway Network*1
TTaken from Gordon and Bryan (1973) and Henderson et al. (1975).
40
-------
Table 10. Los Angeles County divided into 13 areas
(excluding sparsely populated northern
' portion). Dots denote air sanpling sites.t
"Table from Gordon (1976),
41
-------
Table 11. Average lead to PAH ratios from automobile emis-
sions and percent deviations of calculated PAH
concentrations from measured concentrations.*1
Lead/PAH x 10 Deviations from
Measured
Concentrations
COR
BaP
BeP
BFLfc
BghiP
ANTH
1.40±0.47
2.77±0.98
2.73±0.96
2.09
0.87±0.31
4.97±3.21
-916
-7±6
-1±8
-11
+6±4
+7±12
Arithmetic mean within 68% confidence of data shown in
Tables 7 and 8. Data from Table 7 not covering an entire
year were ignored. The June 71-June 72 results for sites
1 and 3 were also ignored since they were biased by our
assumptions (see p. 31).
Arithmetic mean of the two values shown in Table 7 for
samples taken from Dec. 72 to Dec. 73 at sites 1 and 3.
42
-------
V. CONCLUSIONS AND FURTHER CONSIDERATIONS
The chemical element balance method has been modified to account
for reactive species assuming that they follow a first order decay
process. Application of the method to airborne PAH in the Los
Angeles basin has shown that automobile emissions generally predomi-
nate. Significant non-automobile contributions were found for sites
near refineries. Some of the assumptions made in this work can be
relaxed or eliminated as follows:
1. An'analysis similar to that of Grimmer and Hildebrandt (1975)
for the German automobile fleet should be performed for the
Los Angeles automobile fleet. This would ensure that the
values for automobile emissions used in the source concen-
tration matrix are representative.
2. Coronene was assumed to be a non-decaying species in carry-
ing out the source resolution. This assumption is not
needed if mass fractions of species with respect to the
TSP, present in the .'source at tho point of emission, are
known.
3. The dependence of the decay factors on the locations of
emission sources was ignored. This allowed dramatic sim-
plification in the decay factor calculations. However, it
must be taken as a first approximation which can be up-
graded by dividing the region of interest into grids to
account for the non-uniforn distribution of emissions. Such
: 43
-------
a model has been previously used to estimate the spatial
and temporal distribution of CO, HC and NO emissions
within Los Angeles (Reynolds et al., 1973; Roth et al.,
1974).
The limiting assumptions made in the calculation of rates of PAH
degradation, aside from those stated above, were the following:
a. The eleven-hour average residence time of particles of
the Los Angeles aerosol, based on morning inversion height
and lead concentration, is representative of the average
residence time of the polycyclic organic matter.
b. PAH do not react after deposition in the filter media.
An assumption similar to assumption a has been made in the past.
Lemke et al. (1969) estimated an average residence time of one day
for air in the Los Angeles basin, based on a 500 n inversion heigh.£
and carbon monoxide concentration.
Assumption b has been justified previously. Errors induced
by this postulate can be minimized by using polytetrafluoroethylene
membrane filters thus, avoiding the catalytic action of glass and
silica surfaces as pointed out by Lee et al. (1980).
This work has also summarized much of the data on PAH emissions
from major combustion sources available in the literature. It was
shown that data, within the same type of source (i.e. coal-fired
power plants, incinerators and so on), were generally not in a good
agreement. Thus, the selection of the data to be included in the"
source concentration matrix often seemed speculative.
44
-------
In our analysis, representative automobile emission profiles,
obtained fron published data, have been used to deduce refinery
emission profiles. A sinilar strategy can be applied to infer
representative source concentrations from other classes of sources.
45
-------
REFERENCES
Abrott. T. J., Barcelona, M. H.. White, W. H., Friedlander, S. K.f
Morgan, J. J., "Human Dosage/Emission Source Relationships for
Benzo(a)pyrene and Chloroform in the Los Angeles Basin", Special
Report to the U.S. Environmental Protection Agency, California
Institute of Technology (1978).
Albagli, A., Oja, H., Dubois, L. , "Size-Distribution Pattern of Poly-
cyclic Aromatic Hydrocarbons in Airborne Particulates", Environ.
Lett., £, 241 (1974).
Alsberg, T. , Stenberg, U. , "Capillary GC-MS Analysis of PAH Emissions
from Combustion of Pdat and Hood in a Hot dater Boiler", Chemo-
sphere, 7_, 487-96 (1979).
APCD, Profile of Air Pollution Control, Air Pollution Control District
County of Los Angeles (1971).
AQMD, Personal Communication, South Coast Air Quality Management Dis-
trict, El Monte, California (1980).
Barofsky, D. F., Baum, E. J., "Exploratory Field Desorption Mass
Analysis of the Photoconversion of Adsorbed Polycyclic Aromatic
Hydrocarbons", Amer. Chem. Soc- J., 98, 8286-7 (1976).
Begeman, C. R., Colucci, J. M. , "Polynuclear Aromatic Hydrocarbon
Emissions from Automotive Engines", Society of Automotive
Engineers Paper No. 700469 (1970).
Bennett, R. L. , Knapp, K. T. , Jones, P. W., Wilkerson, J. E. , Strup,
P. E., "Measurement of Polynuclear Aromatic Hydrocarbons and
Other Hazardous Organic Compounds in Stack Gases", in Polynu-
clear Aromatic Hydrocarbons, P. H. Jones and P. Leber editors,
Ann Arbor Science Publishers, Inc., Ann Arbor, MJ, 419-28 (1979).
Bevington, P. R., Data Reduction and Error Analysis for the Physical
Sciences, McGraw-Hill, New York (1969).
Bj^rseth, A., BjfJrseth, 0., Fjeldstad, P. E., "Polycyclic Aromatic
Hydrocarbons in the Work Atmosphere II. Determination in a Coke
Plant", Scand. J. Work Environ. ft Health, £, 224-36 (1978).
Brickmeyer, B. A., Spindt, R. S., "Measurement of Polynuclear Aro-.
matic Hydrocarbons in Diesel Exhaust Gases", Society of Automo-
tive Engineers Paper No. 780115 (1978).
46
-------
Broddin, C., Van Vaeck, L., Van Cauwenberghe, K., "On the Size Dis-
tribution of Pol/cyclic Aromatic Hydrocarbon Containing Par-
ticles fron a Coke Oven Emission Source", Atmos. Environ., IIf
1061-4 (1977).
Bryan, R. J., "Point Source Emission Inventory for Los Angeles County*1,
Environmental Protection Agency, EPA-450/3-74-018 (1974).
Carbenry, J. J., Chemical and Catalytic Reaction Engineering, McGraw-
Hill, New York. Ch. 3'.(1976).
Crittenden, B. D., Long, R., "The Mechanisms of Formation of Poly-
nuclear Aromatic Hydrocarbon Compounds in Combustion Systems",
in Carcinogenesis--A Comprehensive Survey, R. I. Freudenthal and
P. Jones editors. Raven Press, New York, Vol. I, 209-23 (1976).
Davies, I. W., Harrison, R. M., Perry, R., Ratnayaka, D., Wellings,
R. A., "Municipal Incinerator as Source of Polynuclear Atomatic
Hydrocarbons in the Environment", Environ. Sci. Technol., 10,
451-3 (1976).
De Wiest, P., Rondia, D., "Sur la Validite" des Determinations du
Benzo(a)pyr£ne Atmospherique pendant les Hois d'Ete"", Atmos.
Environ., 1£, 487-9 (1976).
Dong, M., Locke, D. C., Ferrand, E., "High Pressure Liquid Chromato-
graphic Method for Routine Analysis of Major Parent Polycyclic
Aromatic Hydrocarbons in Suspended Particulate Matter", Anal.
Chem., 4£, 368-71 (1976).
Dunker, A. M., "A Method for Analysing Data on the Elemental Composi-
tion of Aerosols", presented to the American Chemical Society,
Washington, D.C., September 12, 1979.
Dzubay, T. G., "Chemical Element Balance Method Applied to Dicho-
tomous Sampler Data", in Proceedings of the Conference on Aero-
sols: Anthropogenic and Natural, Sources and Transport, New York,
Ann. N.Y. Acad. Sci., 338, 126-44 (1980).
Falk, H. L., Markul, I., Kotin, P., "Aromatic Hydrocarbons IV. Their
Fate Following Emission into the Atmosphere and Experimental
Exposure to Washed Air and Synthetic Smog", AMA Arch. Ind. Health,
1,3, 13-7 (1956).
v/ Friedlander, S. K., "Chemical Element Balances and Identification of
Air Pollution Sources", Environ. Sci. Technol., 7, 235-40 (1973).
/ Gatz, D. F., "Relative Contributions of Different Sources of Urban
Aerosols: Application of a New Estimation Method to Multiple
Sites in Chicago", Atraos. Environ., 9, 1-18 (1975).
47
-------
/ Gartrell, G., Jr., Friedlander, S. K., "Relating Paniculate Pollution
to Sources: The 1972 California Aerosol Characterization Study*1,
Atmos. Environ., £, 279-99 (1975).
Gerstle, R. W., Kemniti, D. A., "Atmospheric Emissions from Open
Burning", J. Air Pollut. Control Assoc., IT, S24-7 (1967).
Gordon, R. J., "Distribution of Airborne Polycyclic Aromatic Hydro-
carbons Throughout Los Angeles", Environ. Sci. Technol., 10^
370-3 (1976).
Gordon, R. J., Personal Communication (1980a).
/ Gordon, G. E., "Techniques for Treating Multielement Particulate Dat*
to Obtain Information on Sources: Overview", in Proceedings of
the Conference on Aerosols: Anthropogenic and Natural, Sources
and Transport, New York, Ann. N.Y. Acad. Sci., 338, 93-102 (1980b).
Gordon, R. J., Bryan, R. J., "Patterns in Airborne Polynuclear Hydro-
carbons Concentrations at Four Los Angeles Sites", Environ. Sci.
Technol., l_t 1050-3 (1973).
Grimmer, G., Hildebrandt, A., "Investigations on the Carcinogenic Bur-
den by Air Pollution in Man XII. Assessment of the Contribution
of Passenger Cars to Air Pollution by Carcinogenic Polycyclic
Hydrocarbons", Zbl. Bakt. Hyg., I. Abt. Grig. B 161, 104-24 (1975).
Gross, G. P., "Second Annual Report on Gasoline'Composition and Ve-
hicle Exhaust Gas Polynuclear Aromatic Content", CRC-APRAC Pro-
ject CAPE-6-68 (1971).
Gross, G. P., "The Effect of Fuel and Vehicle Variables on Polynuclear
Aromatic Hydrocarbon and Phenol Emissions", Society of Automotive
Engineers Paper No. 720210 (1972).
Gross, G. P.,"Aut6raotive Emissions of Polynuclear Aromatic Hydrocar-
bons", Society of Automotive Engineers Paper No. 740564 (1974).
Hangebrauck, R. P., Von Lehmden, D. J., Meeker, J. E., "Sources of
Polynuclear Hydrocarbons in the Atmosphere", U.S.D.H.E.N.,
Public Health Service Pub. No. 999-AP-33 (1967).
Henderson, B. E., Gordon, R. J., Menck, M., Soohoo, J., Martin, S. P.,
Pike, M. C., "Lung Cancer and Air Pollution in the Southcentral
Los Angeles County", Amer. J. Epidemiol., 101, 477-88 (1975).
Hopke, P. K., "Source Identification and Resolution Through Application
of Factor Analysis and Cluster Analysis", in Proceedings of the
Conference on Aerosols: Anthropogenic and Natural, Sources and
Transport, New York, Ann. N.Y. Acad. Sci., 338, 103-15 (1980).
48
-------
Hopke, P. K., Gladney, E. S., Gordon, G. E., Zoller, H. H., Jones, A. G.,
"The Use of Multivariate Analysis to Identify Sources of Selected
Elements in the Boston Urban Aerosol", At'mos. Environ., 10, 1015-25
(1976).
Huntzicker, J. J., Friedlander, S. K., Davidson, C. I., "Material Bal-
ance for Automobile Emitted Lead in Los Angeles Basin", Environ.
Sci. Technol., £, 448-57 (1975).
Junk, G. A., Ford, C. S., "A Review of Organic Emissions from Selected
Combustion Processes", Chemosphere, 187-230 (1980).
Karatsune, M., Hirohata, T., "Decomposition of Polycyclic Aromatic
Hydrocarbons Under Laboratory Illuminations", Anal, of Carcino-
genic Air Pollutants, Monograph Pa. 9, National Cancer Inst.,
117-25 (1962).
Kaschani, D. T., "Bestimmung und Gehalt von PAJC in Dieselabgas von
Kraftfahrzeugen", Erdoel Kohle, Erdgas, Petrochem., 32, 572
(1979).
Kata, M., Chan, C., "Comparative Distribution of Eight Polycyclic
Aromatic Hydrocarbons in Airborne Particulates Collected by
Conventional High-Volume Sampling and by Size Fractionation",
Environ. Sci. Technol., 14, 838-43 (1980).
Katz, M., Chan, C., Tosine, H., Sakuma, T., "Relative Rates of Photo-
chemical and Biological Oxidation (in vitro) of Polynuclear
Aromatic Hydrocarbons", in Polynuclear Aromatic Hydrocarbons,
P. M. Jones and P. Leber editors, Ann Arbor Science Publishers,
Inc., Ann Arbor, MI, 171-89 (1979).
Korfmacher, W. A., Natusch, D. F. S., Taylor, R. S., Mamantov, G.,
Wehry, E. L., "Oxidative Transformations of Polycyclic Aromatic
Hydrocarbons Adsorbed on Coal Fly Ash", Science, 207, 763-5
(1980).
Kowalczyk, G. S., Choquette, C. E., Gordon, G. E., "Chemical Element
Balances and Identification of Air Pollution Sources in Wash-
ington, D.C.", Atmos. Environ., ^2, 1143-53 (1978).
Lane, D. A., Katz, M., "The Photoraodification of Benzo(a)pyrene,
Benzo(b)fluoranthene and Benzo(k)fluoranthene Under Simulated
Atmospheric Conditions", in Advances in Environmental Science
and Technology, I. H. Suffet editor, J. Wiley and Sons, New York,
Vol. 8, Pt. 2, 137-54 (1977).
•
Lao, R. C., Thomas, R. S., Monkman, J. L., "Computerized Gas Chromato-
graphic-Mass Spectrometric Analysis of Polycyclic Aromatic Hydro-
carbons in Environmental Samples", J. Chromat., 112, 681-700
(1975).
49
-------
Lee, F. S.-C., Pierson, W. R., Ezike, J., "The Problen of PAH Degrada-
tion during Filter Collection of Airborne Participates. An Evalu-
ation of Several Commonly Used Filter Media", in Proceedings of
the Fourth International Symposium on Polynuclear Aromatic Hydro-
carbons, Colombus, Ohio, 2-4 October 1979, Battelle Publishing
Inc. (1980).
Lee, H. L., Prado, G. P., Howard, J. B., Kites, R. A., "Source Identi-
fication of Urban Airborne Polycyclic Aromatic Hydrocarbons by
Gas Chromatograpic Mass Spectrometry and High Resolution Mass
Spectrometry", Biomed. Mass Spectron., 4_, 182-6 (1977).
Lemke, E. E., Thomas, C., Zwaicker, W. E., Profile of Air Pollution
in Los Angeles County. Los Angeles County Air Pollution Control
District (1969).
McMahon, C. K.; Tsoukalas, S. N., "Polynuclear Aromatic Hydrocarbons
in Forest Fire Smoke", in Carcinogenesis, A Comprehensive Survey,
P. W. Jones and R. I. Freudenthal editors. Raven Press, New York,
Vol. 3, 61-73 (1978).
Miguel, A. H., Friedlunder, S. K., "Distribution of Benzo(a)pyrene
and Coronene with Respect to Particle Size in Pasadena Aerosols
in the Submicron Range", Atmos. Environ., 12, 2407-13 (1978).
Miller, M. S., Friedlander, S. K., Hidy, G. M., "A Chemical Element
Balance for the Pasadena Aerosol", J. Colloid Interface Sci., 39,
165-76 (1972). °
NAS, Particulate Polycyclic Organic Matter, National Adacemy of
Sciences, Washington, D.C., Ch. 3-4-5 (1972).
/ NAS, Controlling Airborne Particles, National Academy of Sciences,
Washington, D.C., Ch. 5 (1980).
Natusch, D. F. S., "Characterization of Atmospheric Pollutants from
Power Plants", in Proc. 2nd Fed. Conf. Great Lakes, 114-29
(1976).
Natusch, D. F. S., "Potentially Carcinogenic Species Emitted to the
Atmosphere by Fossil-Fueled Power Plants", Environ. Health
Perspect., 22_, 79-9C (1978).
Newhall, H. K., Jentoft, R. E., Ballinger, P. R., "The Effect of Un-
leaded Fuel Composition on Polynuclear Aromatic Hydrocarbon
Emissions", Society of Automotive Engineers Paper No. 730834
(1973).
Pedersen, P. S., Ingwersen, J., Nielsen, T., Larson, E., "Effects of
Fuel, Lubricant and Engine Operating Parameters on the Emission
of Polycyclic Aromatic Hydrocarbons", Environ. Sci. Technol.,
JJ, 71-9 (1980).
50
-------
Pupp, C., Lao, R. C., Murray, J. J., Pottle, R. P., "Equilibriun
Vapor Concentrations of Some Pol/cyclic Aromatic Hydrocarbons,
As406 and ScO2 and the Collection Efficiencies of these Air Pol-
lutants", Atmos. Environ., 8, 915-25 (1974).
Reynolds, S. D., Roth, P. M., Seinfild, J. H., "Mathematical Modeling
of Photochemical Air Pollution—I. Formation of the Model*1, Atnos.
Environ., 7, 1033-61 (1973).
Roth, P. M., Roberts, P. J. H., Liu, M. K., Reynolds, S. D., Seinfeld,
J. H., "Mathematical Modeling of Photochemical Air Pollution—
II. A Model and Inventory of Pollutant Emissions", Atmos. Environ.,
£, 97-130 (1974).
Sawicki, E., Hauser, T. R., Elbert, W. C., Fox, F. T., Meeker, J. E.,
"Polynuclear Aromatic Hydrocarbon' Composition of the Atmosphere
in Some Large American Cities", Amer. Ind. Hyg. Assoc. J., 23,
137-44 (1962).
Suess, M. J., "The Environmental Load and Cycle of Polycyclic Aromatic
Hydrocarbons", Science Total Environ., 6, 239-50 (1976).
Tebbens, B. D., Thomas, J. F., Mukai, M., "Fate of Arenes Incorporated
with Airborne Soot", Amer. Ind. Hyg. Assoc. J., 27, 415-22 (1966).
Tebbens, B. D., Mukai, M., Thomas, J. F., "Fate of Arenes Incorporated
with Airborne Soot: Effect of Irradiation", Amer. Ind. Hyg. Assoc.
J., j!2_, 365-72 (1971).
Turner,, B. D., "Atmospheric Dispersion Modeling. A Critical Review",
J. Air Pollut. Control Assoc., 29_, 502-19 (1979).
Watson, J. G., Jr., "Chemical Element Balance Receptor Model Metho-
dology for Assessing the Sources of Fine and Total Suspended
Particulate Matter in Portland, Oregon", Ph.D. Thesis, Depart-
ment of Chemistry, Oregon Graduate Center, Beaverton (1979).
Williams, R. L., Swarin, S. J., "Benzo(a)pyrene Emissions fron Gaso-
line and Diesel Automobiles", Society of Autnotive Engineers
Paper No. 790419 (1979).
Young, H. D., Statistical Treatment of Experimental Data, McGraw-Hill,
New York, Ch. 4 (1962).
51
-------
APPENDIX A
The weighted sua of the squares of the differences between the
measured and the calculated concentrations can be developed as fol-
lows:
n
x- £<
i-l
—
i-l a * i-l j-1 a
pi pi
22 .
n p c.. m, n p-1 p c.,c, to.«
* " *f-* 2 */—'£-' s * 2
i-l 5^1 0 i-l j-i £=1 a
Differentiating X with respect to source m. and equating the result
to zero yields the solution:
n p.c. . n c,
0 - -2
-2
i-l o
pi Mi
JL Picij 4L P
i-l ov i«i na
i-l n-1
B
Generalizing this result to j«l...p provides the solution shown in
section II-l.
52
-------
APPENDIX B
The most frequently occurring morning inversion height, H, be-
tween the years 1950 and 1974 was 350 meters (AQMD, 1980. Tho
distance traveled by one vehicle was averaged at 10 miles/year.
The gasoline-powered vehicle registration in 1971 was 4.2x10
vehicles, the gasoline consumed that year was averaged at 8.7x10
gals/day (APCD, 1971). Thus, the average gas mileage, G, was
13.2 miles/gal.
The lead content of Los Angeles gasoline, L, for the year 1972
was averaged at 0.56 g/L (Huntzicker et al., 1975). The mass frac-
tion of airborne lead, a, per unit input estimated from Huntzicker
et al., was 0.32.
The traffic density, d, at site 1 (see Table 9 in text) be-
tween June 1971 and June 1972 was 2x10 vehicles per square mile
and per day while the lead ir^ss concentration, p, was 5.35 ug/»
(Gordon and Brian, 1973).
Based on these results, an average residence time, 8, can be
computed as follows:
0 - 16'4 (oxLKd/G) «
where 16.4 is a units conversion factor.
53
-------
APPENDIX C
Table 12. Pattern of Refinery Contributions at Site 3
PAH
BFL
BaP
BeP
BghiP
ANTH
CORl>
Sampling
June 71 -June 72
2.01
2.64
1.41
2.90
0.60
0.33
Period
Dec. 72-Dec. 73
2.39
2.63
1.38
1.82
0.58
-0.19
Aritha.
Mean
2.20
2.64
1.40
2.36
0.59
0.07
PAH/BePa
1.60
3.93
1.00
2.03
2.01
0.05
aAll PAH were corrected for their decay*
54
------- |