United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600/3-79-052
May 1979
Research and Development
oEPA
Photochemical
Aerosol Dynamics
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems are assessed for their long- and short-term influ-
ences. Investigations include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects. This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-79-052
May 1979
PHOTOCHEMICAL AEROSOL DYNAMICS
by
S. K. Friedlander
California Institute of Technology
Pasadena, CA 91125
Grant No. R802160
Project officer
Jack L. Durham
Atmospheric 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 publica-
tion. 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.
ii
-------
ABSTRACT
New data are reported on (1) the rate of formation of condensable chemi-
cal species by photochemical reactions (2) the effect of the reaction products
on the particle size distribution and (3) the distribution of reaction
products as a function of particle size.
Gas-to-particle conversion for cyclopentene, cyclohexene and 1,7-
octadiene, ranged from 5 to 39 percent of the initial gas-phase carbon con-
centrations. Size distribution data for cyclohexene were correlated by a
diffusion controlled growth law with a Kelvin cutoff diameter at about 0.25 ym.
In polluted atmospheres, some new particle formation takes place as a
result of homogeneous gas phase reactions even though an aerosol is already
present. To explain the results of laboratory studies of this phenomenon,
classical nucleation theory must be modified to take into account the scaveng-
ing of clusters by the aerosol.
Using a new low pressure impactor, the first measurements have been made
of the distributions of sulfate and nitrate with respect to particle size for
dp < 0.25 ym. In Pasadena, the data for sulfate often show a peak in the
mass distribution for 0.6 < dp < 1.0 ym; less often, a peak is observed near
0.1 ym, consistent with laboratory data for aerosols formed by homogeneous
gas phase reactions.
This report was submitted in fulfillment of Grant No. R802160 by the
California Institute of Technology, under the sponsorship of the Environmental
Protection Agency. This work covers the period April 1, 1973 to March 31, 1978,
111
-------
CONTENTS
Abstract li:L
Figures vi
Tables vii
I. Introduction 1
2. Conclusions 3
3. Nucleation Processes A
Kinetics of homogeneous nucleation 4
Experimental test of the theory 5
4. PhotochemicalProcesses 8
Photochemically generated condensable species:
sulfates and nitrates 8
Photochemically generated condensable species:
organics
S. Particle Growth 14
Heterogeneous condensation . 14
Growth laws 16
Measurement of growth laws: organic aerosols 17
6. New particle formation 22
Effect of pre-existing aerosol 22
7. Distribution of chemical species with respect to
particle size 27
Distribution of sulfur with respect to particle size . . 27
Aerosol growth dynamics 30
Deviation between theory and experiment: sulfates ... 31
Distribution of nitrate with respect to particle size, . 32
35
References
Appendix
A. Project Dissertations and Published Papers 37
-------
FIGURES
Number Page
1 The droplet current for supersaturated water vapor at
T=300°K calculated from (3) 6
2 Cyclohexene concentration as a function of the integral
of the ozone concentration over time 11
3 Number concentrations or particles larger than a given
diameter vs diameter at various times in a smog chamber
experiment 19
4 Particle growth rates between the fourth and fifth size
distribution measurements for the data of Fig. 3 20
5 Aerosol colume distributions for the data of Fig. 3 21
6 Distribution of sulfur with respect to particle size for the
ambient aerosol in Pasadena, California, measured with a
low pressure impactor 28
7 The distribution of sulfur with respect to particle size for
a balloon chamber aerosol, measured with the low pressure
impactor 29
8 Aerosol nitrate distributions with respect to particle size
for different locations at various times of day 33
vi
-------
TABLES
Number Page
1 Mechanisms of Gas-to-Particle Conversion.
2 Room Temperature Reaction Rate Constants for Hydrocarbons
Relevant to Smog Chamber Experiments* 10
3 Fraction of the Initial Hydrocarbon Converted to Aerosol
Organic Carbon 12
4 Aerosol Organic Carbon (AOC) Formation Rates 13
5 Growth Laws for Gas-to-Particle Conversion. ... 17
vii
-------
SECTION 1
INTRODUCTION
About 30 percent (annual average) of the aerosol mass in the Los Angeles
atmosphere results from gas-to-particle conversion in the atmosphere (Hidy
and Friedlander, 1971). The percentage is much greater on days of strong
photochemical activity (Grosjean and Friedlander, 1975). This material plays.
a major role in visibility degradation since it tends to accumulate in the
0.1 to 1.0 ym size range which is most efficient for scattering light.
Chemically, it includes sulfate, nitrate and organic compounds, some of which
are of public health concern.
Gas-to-particle conversion may result from homogeneous gas phase processes
or it may be controlled by processes in the aerosol phase. Gas phase processes,
either physical or chemical, can produce a supersaturated state which then
collapses by aerosol formation. Physical processes include adiabatic expansion,
mixing with cool air, and radiative or conductive cooling.
Gas phase chemical reactions such as the oxidation of S02 to sulfuric
acid or the reaction of ozone with certain olefins may also generate condens-
able products. Even if the gas is not saturated with respect to one of the
products in its pure form, condensation may take place by the formation or
droplets composed of binary solutions.
Once a condensable species has been formed in the gas phase and the
system is in a non-equilibrium state, it may pass toward equilibrium either
by the generation of new nuclei or by condensation on existing particles.
When condensation takes place on existing particles, self-nucleation is
suppressed; the existing aerosol scavenges monomer and clusters,and nucleation
theory must be modified to predict rates of new particle formation as shown in
this report.
Alternatively, molecules from the gas may react on the particle surface
or in a droplet. The process can be considered to consist of two steps in
series. If the gas-to-particle transport step is rapid compared with the con-
version step, the rate of particle growth will be controlled by the rate of
chemical conversion in the particulate phase. This is probably the case for
S02 conversion in droplets in stack plumes when solar radiation is not strong.
Table 1 summarizes the various gas-to-particle conversion processes.
Of particular interest in this report are (1) the formation of conden-
sable species by gas-phase photochemical reactions (2) the effect of the con-
densable products on the particle size distribution and (3) the distribution
-------
of reaction products as a function of particle size. The discussion begins
with a brief review of classical nucleation theory; later it is shown how
this theory must be modified for photochemically generated pollution aerosols.
TABLE 1. MECHANISMS OF GAS-TO-PARTICLE CONVERSION
I. Homogeneous Nucleation
A. Physical processes producing supersaturation
1. Adiabatic expansion
2. Mixing
3. Conductive cooling
4. Radiative cooling
B. Gas phase chemical reaction
1. Single condensable species (classical theory)
2. Multicomponent condensation (hetermolecular theory)
II. Heterogeneous Condensation
A. Transport limited
1. Diffusion, d ȣ
2. Molecular bombardment, d «i
B. Surface controlled chemical reaction or adsorption
C. Particulate phase controlled chemical reaction
-------
SECTION 2
CONCLUSIONS
The secondary aerosol formed in photochemical smog is composed of sul-
fates, nitrates and organic compounds. Of special interest are the rate of
formation of aerosol material, the particle size distribution of the result-
ing aerosol and the distribution of the different chemical species with
respect to particle size.
The results of new laboratory studies of the conversion of
four organic gases to aerosol are discussed. The percentage conversion to
aerosol depends on two factors—the rate of reaction of the organic aerosol
precursor and the physicochemical properties of the reaction products. Cyclo-
pentene is a particularly efficient aerosol precursor; more than a third of
the carbon present in the gas phase was converted to aerosol material. The
condensable reaction products diffuse to the surfaces of particles in the
reacting gas mixture. A critical particle diameter of 0.25 ym below which no
growth occurs was observed in experiments with cyclohexene.
A novel experimental study was made of new particle formation in a photo-
chemical ly reacting gas containing S02 with controlled amounts of aerosol
present. Classical nucleation theory was not able to explain the experi-
mentally observed rates of new particle formation. A "nucleation barrier"-
free model came much closer to predicting observed rates and is justifiable
on theoretical grounds. More experimental and theoretical studies of this
phenomenon are needed because of its fundamental importance to aerosol
dynamics.
Using a specially designed low pressure impactor and a flash-volatiliza-
tion analysis method, distributions of sulfate and nitrate have been measured
for particles as small as 0.05 ym. In laboratory studies with an aerosol
present, sulfate tends to accumulate in the size range around 0.1 ym. Ambient
measurements in Pasadena usually show a sulfate peak in the 0.6 to 1.0 ym size
range but a peak at smaller sizes is sometimes found. More research is needed
to explain this discrepancy.
Nitrate distributions were measured at a coastal, an urban and an agri-
cultural site further inland. The coastal nitrate was found in larger part-
icles, probably as a result of droplet phase reaction in the marine aerosol.
The measurements made inland showed a nitrate peak in the submicron range. A
bimodal nitrate distribution was observed at the urban site located between
the other two. The dynamics of nitrate containing aerosols remains poorly
understood; both theoretical and experimental studies are needed.
-------
SECTION 3
NUCLEATION PROCESSES
KINETICS OF HOMOGENEOUS NUCLEATION : A REVIEW
The vapor pressure over a small droplet is greater than that over a flat
surface of the same liquid (Kelvin effect), this leads in nucleation theory
to the existence of a critical droplet diameter, d*, related to the saturation
ratib as follows: "
d* - '"-m
p ~ KTAttS
where a is the surface tension of the droplet liquid, v^ is the molecular
volume* k Boltzmann's constant, and T the absolute temperature. If the
saturation ratio, S, is fixed, droplets smaller than d* tend to evaporate
while larger Ones grow. Calculations based on the obslrvations of Wilson
and subsequent measurements by many other experimenters indicate values Of
d* mahy times greater than the diameter of a single water molecule, about
2?8 A. Hdw, then* does Condensation take place in systems which have been
freed £fOm condensation nuclei by previous expansion or filtration?
Molecular clusters are always present even in an unsaturated gas as a
result of collisions among the molecules. When a system becomes super-
saturatedj these clusers increase in concentration and pass through the
critical size d* by attachment of single molecules. The formation of stable
nuclei relieves^the supersaturation in the gas. Since condensation nuclei
are generated by the vapor itself* the process is khowtt as homogeneous nuclea-
tion Or self-nucleation.
The rate of formation of clusters containing g monomer molecules is
called the particle (or droplet) current. It is denoted by the symbol 1 and
has dimensions of particles per cm-* per seeOhd. According to the theory Of
homogeneous nucleation (Frenkel, 1955) the particle current can be expressed
as the sum Of two terms:
D
r, 3n v 8A$ ,.,
1 = - Dv3V * kf "17 h (2)
The first term on the right hand side represents the diffusion of particles
through particle size (v) sfjace. the coefficient of diffusion is given by
the relation ird2y
D = PI P*
v (2tnnkt)
-------
D
where PJ is the partial pressure of condensing molecules of mass m.
has dimensions of (particle volume) 2 sec'-'-. The second term represents par-
ticle transport through v space under the influence of the potential energy.
A$ = (Ay) g + ird a
P
where Ap is the chemical potential difference between the gas and droplet
phases per molecule. As a result of the diffusional term in (2), particles
originally present in a given size range will tend to spread over a wider
size range as growth occurs.
If the particle current I is independent of cluster size, the following
expression can be derived (Frenkel, 1955):
I = 2
(2rrmkT)
__
1/2
ov
2/3 -,
m
1/2
3 2
exp -
3(kT)3(TnS)2
(3)
The first term in brackets is the monomer flux (molecules per unit area per
unit time) and the second is proportional to the monomer surface area per
unit volume of gas. Their product has the same dimensions as I, number per
unit volume per unit time. The group crv£/ -VkT is dimension less. The last
term includes an exponential factor, the "nucleation barrier," a term much
less than unity when the saturation ratio is small (nearly unity) -
The droplet current calculated from (3) is shown in Figure 1 for water
vapor at a temperature of 300°K. Order of magnitude increases in I result
from small increases in S, primarily because of the dependence on In S in
the argument of the exponential function. Although a supersaturated vapor
is always unstable, the rate of generation of stable nuclei is negligible
for small values of S. When 1=1 drop/cm^ sec, particle formation can be
conveniently observed experimentally. The corresponding value of S is
designated the critical saturation ratio, S
EXPERIMENTAL TEST OF THE THEORY
It is difficult to carry out experimental studies capable of verifying
the theory. Measurements of the nuclei size spectrum would constitute a 0
sensitive check, but instruments capable of measurement in the 10 to 100 A
size range have not been available. Most experimental tests have involved
measurements of the saturation ratio at which condensation occurs using an
expansion (Wilson) cloud chamber. Data collected with the chamber are
difficult to interpret because of the unsteady nature of the expansion
process. These investigations are reviewed by Mason (1971).
The diffusion cloud chamber is a particularly attractive experimental
system for the study of -nucleation kinetics; it is compact and produces a
well-defined, steady super saturation field. The chamber is cylindrical in
-------
25 r-
Figure 1. The droplet current for supersaturated water vapor at T = 300°K
calculated from (3). The critical saturation ratio, correspond-
ing to I = 1 cm""-* sec,
is about 3.1.
-------
shape, perhaps 30 cm in diameter and 4 cm high. A heated pool of liquid at
the bottom of the chamber evaporates into a stationary carrier gas, usually
hydrogen or helium. The vapor diffuses to the top of the chamber, which is
cooled, condenses and drains back into the pool at the bottom. Since the
vapor is denser than the carrier gas, the density is greatest at the bottom
of the chamber and the system is stable with respect to convection. Both
diffusion and heat transfer are one dimensional, with transport occurring
from the bottom to the top of the chamber. At some position in the chamber,
the temperature and vapor concentrations reach levels-corresponding to super-
saturation. The variation in the properties of the system are calculated by
numerical integration of the one dimensional equations for heat conduction
and mass diffusion. The saturation ratio is calculated from the computed
local partial pressure and vapor pressure.
The goal of an experiment is to set up a "critical" chamber state, that
is, a state which just produces nucleation at some height in the chamber where
the vapor is critically supersaturated and droplets are visible. This occurs
when the temperature difference across the chamber has been increased to the
point where a rain of drops forms at an approximately constant height.
Observed and predicted saturation ratios can then be compared.
Good agreement between theory and experiment has been obtained with this
system for a variety of organic compounds (Katz et al., 1975). For water,
agreement is poorer (Heist and Reiss, 1973). Such experiments offer the most
convincing support yet developed for the theory of homogeneous nucleation.
This brief review of theory and experiment for homogeneous nucleation
has set the stage for the more complex processes leading to aerosol formation
in polluted atmospheres.
-------
SECTION 4
PHOTOCHEMICAL PROCESSES
PHOTOCHEMICALLY GENERATED CONDENSABLE SPECIES: SULFATES AND NITRATES
In the experiments discussed in the last section, the supersaturated
state was produced by a set of physical processes. Gas phase chemical reac-
tions may also lead to the formation of condensable species and several may be
present simultaneously. Of special air pollution interest are reactions involv-
ing S0_, NO-NO2 and certain organic compounds.
The rate of S02 conversion to particulate sulfur (usually (^4)^804) in
the Los Angeles atmosphere, heavily polluted by the organic vapors and
nitrogen oxides in automobile exhaust, ranged from 1-13% per hour in measure-
ments reported by Roberts and Friedlander (1975). The S02 is probably oxi-
dized by certain free radicals present in photochemical smog. The most
important such radicals are thought to be the hydroxyl, OH:
S02 + OH + M •* HS03 + M
the hydroperoxyl, HO-:
S0_ + H0_ -> SO + OH
and peroxyalkyl, R02:
S0_ + R0_ -*• SO- + RO
where R represents an alkyl group. The conversion of HSOj formed in the first
reaction, to ^804 is believed to occur rapidly but the mechanism has not
been established. Rate constants for reactions of the last type are not well
known and more data are needed to assess their importance for the atmospheric
oxidation of S02- Calculations based on the kinetics of known reactions
involving S02 have not been able to account for S02 oxidation rates measured
in smog chamber studies (Sander and Seinfeld, 1976).
Particulate nitrates may be formed in the atmosphere by the gas-phase
oxidation of NO to N02 and then to nitric acid. Nitrogen oxides are pro-
duced in combustion processes and emitted into the atmosphere in large
quantities by power plants and automobiles. Calvert (1973) considers oxida-
tion by the hydroxyl radical the principal mode of formation
HO + N02 + M -> HON02 + M
8
-------
where M represents a third body. The aerosol nitrate is usually present as
the ammonium salt in polluted atmospheres. This form may result from gas
phase reaction of nitric acid with ammonia or from reaction in the droplet
phase.
If only a single condensable species is formed by chemical reaction, it
may condense by homogeneous nucleation in the absence of existing particles.
In other cases, two or more condensable species may be present simultaneously
in the gas. If these form solution droplets, nucleation can take place at
partial pressures much lower than those required for nucleation of pure vapors.
The theory of this type of homogenous nucleation, known at heteromolecular
nucleation, is a modified version of the theory for a single component dis-
cussed in an earlier section (Reiss, 1950). An application of the theory to
the sulfuric acid-water and nitric acid-water systems has been made by Kiang
and Stauffer (1973) and Mirabel and Katz (1974).
PHOTOCHEMICALLY GENERATED CONDENSABLE SPECIES: ORGANICS
Certain diolefins and cyclic olefins are very efficient aerosol precur-
sors. Grosjean and Friedlander (1978) studied the formation of organic aerosols
using mixtures of NO, N02 (0.15 to 0.30 ppm) and olefins (0.5 to 2.0 ppm).
These were added to atmospheric air (Pasadena) and allowed to react in sun-
light for up to three hours in a large outdoor chamber constructed from 2 mil
FEP Teflon film located on the roof of the Keck Laboratories at Caltech. The
chamber surface was about 117 mr and the initial chamber volume ranged from
53 to 134 m3. Four olefins were studied: 2,3-dimethyl-2 butene, cyclopentene,
cyclohexene, and 1,7 octadiene.
Olefin concentration as a function of time was measured during several
of the cyclopentene, cyclohexene and 1,7-octadiene runs. Ozone appearance in
these systems always coincided with a marked increase in hydrocarbon con-
sumption.
The rate of olefin reaction can be written as a sum of biomolecular
reaction rate terms:
- d[HC]/dt - Zk± [Xi] [HC] (4)
i
where k^ are the second order rate constants of hydrocarbon attack by the
species X^, with Xj^ = 0, OH, 03 H02, etc. Assuming that only one species
reacts with the olefin:
- d[HC]/dt = k. [X.] [HC]
k.
or: log [HC] = log [HC]Q + j-^W f
Jo
dt
where [HC] is the initial hydrocarbon concentration.
-------
Therefore, the relative importance of ozone in the overall hydrocarbon
consumption can be tested further by plotting log [HC] as a function of
I [0^]dt, as is shown in Figure 2 for cyclohexene. With the exception of
•r O O
the first -^ 10 minutes during which an "excess rate" was observed, a good
linear relationship was observed throughout the experiment, with a slope
corresponding to a cyclohexene-ozone rate constant of 0.15 ppm~l min~l. The
reactivity of the four hydrocarbons (2,3-dimethyl-2 butene > cyclopentene >
cyclohexene > 1,7-octadiene) was consistent with their known rates of reac-
tion with ozone and with the hydroxyl radical OH (Table 2). It was difficult,
however, to assess the relative importance of the two species in these experi-
ments .
TABLE 2. ROOM TEMPERATURE REACTION RATE CONSTANTS FOR HYDROCARBONS
RELEVANT TO SMOG CHAMBER EXPERIMENTS* (GROSJEAN £ FRIEDLANDER
1978)
Hydrocarbon
03(a)
OH(b)
0(3P)(c)
N0(d)
2,3-dimethyl 2-butene
cyclopentene
cyclohexene
1,7-octadiene
1-hexene (g)
1 , 3-butadiene (g)
1510
813 No
169
ND
8.4
11.1
110
data (ND)
62.4
ND
31.2
68.5
5.11
1.79
1.62
ND
0.39
1.45
59
23
3.5
ND
1.7
4.3
* Original references given by Grosjean and Friedlander, 1978.
— 18 3 —1 —1
a) k_ in 10~ cm molecule sec
b) k_.u in 10 cm molecule sec
UH
•> , - ,«-H 3 _ , -1 -1
c) k in 10 cm molecule sec
d) Nitric oxide photooxidation rate, ppb min
Aerosol samples were analyzed for their organic carbon content, which
averaged only ^ 35 percent by weight of the total aerosol. Inorganic nitrate
was also produced in all runs and accounted for ^ 10 percent by weight of the
aerosol. Thus, the low carbon fraction indicates that the aerosol products
formed are highly oxygenated species averaging *» 40-45 percent carbon by-
weight. This is indeed the case for the major organic products identified in
these runs, for example adipic acid and 6-nitrato hexanoic acid (carbon =41
percent) from cyclohexene.
10
-------
0.3
0.2 -
ex
Q.
O
0»
5 0
-0.1 -
-0.2
/[O3]dt
11
-------
Filial aerosol organic carbon concentrations ranged from 5 to 39 percent
of the initial hydrocarbon concentrations (Table 3) for cyclopentene, cyclohexene
and 1,7-octadiene. Thus, most of the products resulting from photooxidation
of the olefinic hydrocarbons studied remain in the gas phase. On the other
hand, cyclopentene, cyclohexene and 1,7-octadiene are more potent organic
aerosol precursors than the hydrocarbon mix present in ambient Southern Cali-
fornia air where carbon-gas-particle distribution factors of only 1 to 3
percent are measured (Grosjean and Friedlander, 1975).
Aerosol organic carbon formation rates are listed in Table 4 and ranged
from 1.5 to 41 pig carbon
min
"-*-
In all runs these rates paralleled the
rate of formation of ozone and the product of the hydrocarbon and ozone con-
.centrations, i.e. , increased to the maximum values listed in Table 4 and then
decreased upon further irradiation. Thus, aerosol formation rates were found
to correlate in an approximately linear fashion with the product of the olefin
and ozone concentrations. An empirical relation of the general form: d (aero-
sol) /dt = afO-j] [HC] does not necessarily imply that organic aerosol is
produced only as a result of the ozone-olefin reaction. The relation may be
composite, i,e, may include terms representing the contribution of other
species including the hydroxyl radical.
Assuming this relation holds for urban atmospheres (not necessarily a
good assumption) one obtains, for typical ozone (0.1 ppm) and hydrocarbon
(10 ppb) concentrations, organic aerosol production rates of ^ 0.02, 0.15,
and 0.6 yg carbon m~3 min"* for 1,7-octadiene, cyclohexene and cyclopentene,
TABLE 3. FRACTION OF THE INITIAL HYDROCARBON CONVERTED TO AEROSOL
ORGANIC CARBON (GROSJEAN, 1978)
Hydrocarbon
Cyclohexene
Run No.
A07
A91
A08
A09
A92
A10
All
A12
A/B(a)(b)
0.10
0.06
0.5
0.06
0.07
0.06
0.14
0.17
Hydrocarbon
2,3-dimethyl-2-
butene
1,7-octadiene
Cyclopentene
Run NO.
E01
F04
F05
F91
F06
F01
0.0
0.11
0.13
0.07
0.33
a)
b)
A = aerosol organic carbon measured at the end of the run,
G = initial hydrocarbon concentration, converted from ppm to pg m~C
(A/G is a dimensionless number) .
A/G are presumably upper limits since some artifact aerosol is expected
to form during sampling by adsorption of volatile acidic organics on the
surface of the basic glass fiber filters.
12
-------
respectively. For comparison purposes, average rates of aerosol organic
carbon production in Southern California on smoggy days are of the order of
0.5 ug carbon m~3 min~l (estimated from the measured particulate organic
carbon levels in the eastern part of the Los Angeles Basin, 30 pg carbon m~
and assuming an air parcel transport time of 10 hours.
TABLE 4. AEROSOL ORGANIC CARBON (AOC) FORMATION RATES
Hydrocarbon
Run No.
d(AOC)/dt, Pg C m"3
average
min"1
maximum
Cyclohexene
Cyclopentene
1,7-octadiene
A12
G01
G03
F05
F91
F06
16
41
9.1
2.5
4.3
1.5
39
78
33
3.2
8.0
2.5
13
-------
SECTION 5
PARTICLE GROWTH
HETEROGENEOUS CONDENSATION
When large concentrations of particles are present and the saturation
ratio is low, condensation takes place on the existing particles without
formation of new nuclei. We call this process heterogeneous condensation.
Cloud formation in the atmosphere takes place in this way since supersatura-
tions are usually less than a few percent and nuclei concentrations are high.
The rate of heterogeneous condensation depends on the exchange of matter
and heat between a particle and the continuous phase. The extreme cases of a
particle much larger or much smaller than the mean free path of the suspend-
ing gas, Si, are easy to analyze. In the continuum range, (dp »&) diffusion
theory can be used to calculate the transport rate. For a single sphere in
an infinite medium, the steady state equation of diffusion in spherical co-
ordinates takes the form
D3r jJn
3t r23r
where n is the molecular concentration of the condensing species and D is its
coefficient of diffusion. The solution which satisfies the boundary condi-
tions n = nj, the concentration in equilibrium with the surface at r = d /2
and n = n1 at r = <» is "
n - n, d
v^t = * C6}
The rate of diffusionai condensation is given by
F = D[^M ud2 = 2 wdJKn.-nJ (7)
=d 12
P
= 2 yd D(Pl-pd)/kT
where F is the flow of molecules (number per unit time) to the surface of
14
-------
the particle. The surface concentration, n$, is determined by the surface
temperature and curvature. It is assumed that the condensation rate is
sufficiently slow for the latent heat of condensation to be dissipated
without changing droplet temperature.
For particles much smaller than the mean free path of the gas the rate
of condensation can be calculated from kinetic theory. The net molecular
flux at a surface of area ird2 is given by:
°(P1-Pd)ird«
F = *P (8)
(2irmkl) '
At one atmosphere, this expression applies to particles smaller than about
0.06 ym. Separate accommodation coefficients, ct, are often introduced for
the condensation, p,/(2-mnkT)l/2 and evaporation, pd/(2TrmkT) I/2, fluxes but
the values are assumed equal in (8). In general, these coefficients must be
determined experimentally. An approximate interpolation formula for the
entire range of mean free path has been proposed by Fuchs and Sutugin
!-V f ——2)
ll+1.71Kn + 1.333Kn J
F = 2*^(11,-n.) < i-^S Ty W
where the Knudsen number Kn = 2Vdp and & is the mean free path for collision
of the condensing species. WhenKn«l, (9) reduces to (7) for the continuum
range. When Kn»l, (9) is about 1.2 times (8) with a = l for rigid elastic
spheres.
When the Kelvin effect is important, the partial pressure driving force
for condensation takes the form
[
-T
(10)
i "pfcnSl
= PrPs exp l^g-
where ps is the vapor pressure over a flat surface or pool of liquid and d*
is the critical droplet diameter. It has been assumed that the nucleus ^
behaves like a pure drop of the condensing species. If the surface of the
particle is composed of a material different from that of the condensing
vapor, this result must be modified to account for surface wetting effects.
d* .2
Expanding the exponential of (10)
d*
Ap = p [S-l- g5- AnS - j / -r*- £nS)
- S P ^ P '
15
-------
where S = P-i/Ps- F°r small values of S-l, this takes the approximate form
Ap = -S. (S-l) (d - d*) (11)
P P
As an example, this result can be substituted in (7) for growth by diffusion
in the continuum range:
F = ps(s"1KVdp} (s-i «D
The rate of condensation is proportional to the difference between the particle
diameter and the critical particle diameter.
GROWTH LAWS
Growth laws are expressions for dv/dt or d(dp)/dt as functions of par-
ticle size and the appropriate chemical and physical properties of the system.
Such expressions are necessary for the calculation of changes in the size
distribution function with time as shown in the next section. When growth is
transport limited, the rate can be determined from the expressions derived in
the previous section. For the continuum range, the growth law based on (7)
is given by:
dv _ p m
dt ~ KT
The effect of the moving boundary of the growing particle is neglected.
For nuclei smaller than the mean free path of the gas, the growth law,
based on (8) is given by:
2
dv "V
dt ~
where the accommodation coefficient a has been set equal to unity.
Chemical reactions at particle surfaces are likely to be most important
near aerosol sources where the particle surfaces are fresh and their cata-
lytic activity high. (In the atmosphere, however, contamination probably
destroys the specific catalytic activity of aerosol surfaces.) Particles will
grow if the products of reaction accumulate at the surface. When reaction
rates are fast compared with transport, growth laws are of the same form as
the transport limited laws. When reaction rates are slow compared with trans-
port, the concentration of the reactive species in the gas near the surface
is practically the same as in the bulk of the gas, and the rate of conversion
is determined by the fraction of the collisions with the surface which are
effective :
16
-------
J 2
dv _ "Pl^p Vm
dt " (27rmKT)1/2
where a, the fraction of effective collisions, is usually much less than
unity.
Examples of growth laws including those limited by chemical reaction
in the particulate phase are summarized in Table 5. The dependence of
dv/dt varies from dl for diffusion in the continuum range to dp for droplet
phase chemical reaction. Different forms for the growth law can lead to
significantly different changes in the size distribution function with time
and to the distribution of chemical species with respect to size.
TABLE 5. GROWTH LAWS FOR GAS-TO-PARTICLE CONVERSION
Mechanism GT™th.
dv / dt
Diffusion (d »1) 2irDd v (P. - P.)
p p m 1 d
KT
Molecular vm(Pl - Pd}
bombardment (d «1) (2^mKT)1/2
Surface reaction aiTd v Pn
p mi r~^^-i\
(all sizes) , * 1/2 (0<<1)
Droplet phase
reaction 6P (/JM. v.):
(Friedlander, 1977)
MEASUREMENT OF GROWTH LAWS: ORGANIC AEROSOLS
The growth law for a polydisperse aerosol can be determined by measuring
the change in the size distribution function with time. In experiments
carried out by Heisler and Friedlander (1977), small amounts of certain
organic vapors, which served as aerosol precursors, were added to a sample
17
-------
of the normal atmospheric aerosol contained in an 80 m bag exposed to solar
radiation. Small concentrations of nitrogen oxides were also added. The
bag was made of Teflon which is almost transparent to solar radiation in the
uv range and relatively unreactive chemically. Photochemical reactions involv-
ing the nitrogen oxides lead to the formation of ozone and hydroxyl radicals
which reacted with the organic vapors to produce condensable species as dis-
cussed in a previous Section. The change with time of the size distribution
function was measured with a single particle optical counter.
The number of particles per unit volume larger than a given particle
00 O« Oj
size dp, C n (d )d(d ), is shown in Figure 3 for an experiment with
p
cyclohexene. Consider a horizontal line on the figure corresponding to con-
stant values of this integral. In the absence of homogeneous nucleation,
each such line corresponds to the growth with time of a particle of size given
initially by the curve for t=0. No particle can move across such a line
because the total number larger is conserved.
The growth rate, d(d )/dt = Ad /At can be obtained from adjacent dis-
tributions in Figure 3 as^a function of d and of time. The data were then
plotted with dv/dt as a function of particle diameter as shown in Figure 4.
For this set of data, an approximately linear relationship was found with an
intercept on the positive d axis.
As a reasonable model for explaining these results, it was assumed that
a single condensable species was formed in the gas as a result of chemical
reaction, or a small group of species with similar thermodynamic properties.
Molecules of these reaction products then diffused to the surfaces of exist-
ing aerosol particles. Hence a diffusion controlled growth law modified by
the Kelvin effect in the small saturation ratio approximation (11) correlated
the data as shown in Figure 4. The cutoff particle diameter probably resulted
from the Kelvin effect. For the run shown, the critical diameter was about
0.28 urn. The line is the result of a least-squares fit using (9) combined
with (11) in calculating the growth law. The curvature in the line results
from the form of the interpolation formula (9).
Aerosol volume distributions calculated from the data are shown in
Figure 5. Material accumulates in the size range near 0.6 pm which is
particularly efficient for light scattering. Small particles grow little
because of the Kelvin effect.
In experiments with cyclopentene, the data showed a pronounced bend
at a particle diameter larger than the lower cutoff diameter. The results
were correlated by Heisler and Friedlander (1976) using a growth law which
incorporated another condensing species.
18
-------
10
10"
10
a
u
10
I I
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
d - microns
P
Figure 3. Number concentrations or particles larger than a given
diameter vs diameter at various times in a smog chamber
experiment. Initial concentrations were 2.02 ppm
cyclohexene, 0.34 ppm NO and 0.17 ppm N02- The time
between measurements shown was made 12 minutes after the
addition of the reactants (Heisler and Friedlander- 1977)
19
-------
o
f*
N
tn
I I i I I T
I i I
<1
J_
J_
I
0.2 0.3 0.4 O.S 0.6 0.7 0.8 0.9 1.0 1.1
d , tnicrom
P
Figure 4. Particle growth rates between the fourth and fifth size
distribution measurements for the data of Fig. 3. The
solid line is a least-square fit of the diffusional growth
law, modified to include mean free path effects and the
Kelvin effect. The intercept on the size axis is the
average critical size, d* (Heisler and Friedlander, 1977).
20
-------
1500
£ looo
o
rH
<
500 -
1.0
2.0
d - micron*
P
Figure 5. Aerosol volume distributions for the data of Fig. 3. The
order of the measurements is indicated by the numbers.
Increases are from gas-to-particle conversion. The peak
is near the size range most efficient for light scatter-
ing (Heisler and Friedlander, 1977).
21
-------
SECTION 6
NEW PARTICLE FORMATION
EFFECT OF PRE-EXISTING AEROSOL
There is good evidence that in polluted atmospheres some new particle
formation takes place even though an aerosol is already present (McMurry,
1977) . In this section, the effect of the aerosol on the rate of new par-
ticle formation is examined. The aerosol will, of course, collect monomer
molecules of the condensable species generated by chemical reaction; in the
past, it has usually been assumed that rates of nucleation can be calculated
from Equation (3) derived for particle-free gases. However- as shown in
this section, the classical theory must be modified because of the scavenging
of clusters by the aerosol.
We consider first the case of particle formation in the absence of an
aerosol. Let n be the concentration of clusters containing g monomer
molecules. The time rate of change in ng as a result of particle growth
processes is given by (Frenkel, 1955; Friedlander, 1977):
3n
where Ig, the particle (droplet) current is the net rate of formation of
particles of size g from clusters of size g-1. The net rate of loss of
particles of size g to the range g+1 by growth processes is Ig+j- Collision
and coagulation among the clusters is neglected. In classical nucleation
theory, a steady state is assumed and 3-& = 0. Thus Ig = Ig+i = constant = I.
The result of a theoretical analysis gives Equation (3) for T.
If an aerosol composed of larger particles (dp > 0.05 ym, say) is already
present, (13) must be modified. Clusters deposit on the pre-existing aerosol
at a rate proportional to the aerosol surface area if the particles are
smaller than the mean free path of the carrier gas. (It is assumed that the
aerosol particles are uniformly dispersed on a scale comparable to that on
which the chemical reactions are taking place) . This introduces a new term
into the continuity relation, (13) :
22
-------
where k = Boltzmann's constant
T = absolute temperature
m = mass of cluster of size g
g
A = surface area of pre-existing aerosol per unit volume of gas
Making the steady state assumption, this equation becomes
KT
I - I . = n (^—) A (15)
g g+1 g^2trni ' k J
Changes in A with time resulting from gas-to-particle conversion are assumed
to be small. As a result of cluster deposition on the surface area A, the
particle current Ig is no longer independent of size but decreases with
increasing cluster size.
The particle current can be written as follows Frenkel (1955):
I = Bn -s , - a. n s (16)
g g-1 g-1 g g g }
where B the monomer flux is given by
and a the evaporative flux is given by
o
s
a = B ° ' •" exp
g s r
6 g
4av,
j
d~Kf
p
- inS
(17)
Also s = surface area of a cluster
g
a = surface tension
v = volume of monomer molecule
d = cluster diameter
P
S = saturation ratio
In the special case of negligible evaporative flux, (16) becomes
I = 0n ,s ..
g g-1 g-1
23
-------
Substitution in the steady state expression gives:
n ,s „ - n s =
n A
g
i _j , — ii -j — , / x;
g-1 g-1 g g gl/2T
Rearranging
n -
.
+ A
n s
g g g sgnl
1/2
If the surface area of the foreign aerosol is large, A/g s n » 1
and this expression becomes ^
n ,s -
g-1 g-1 „
n s 1/2
g g g
The cluster size distribution is given by
n,s, n ,s , n,s,
n_s. n s n s s_n. ' ' ' 1/2
22 gg gg 21 g s n
or
']
The rate of new particle formation is given by
7/6
eV [g!]
= 'V* • -
with the same constraint given in (19b) .
It is not necessary for the evaporative flux to vanish for this analysis
to hold at least approximately. Inspection of (16) indicates that the second
(evaporative) term will be much smaller than the first provided that
24
-------
n s « n ,s ,B
g g g g-1 g-1
or
n
n
n s
g g
n *s
g g
where the asterisk refers to the equilibrium distribution. This means that
the evaporative flux can be neglected provided that the cluster size distri-
bution is much steeper than the equilibrium distribution.
The largest value of <*g/3 occurs for a cluster of size g = 2. The value
of this ratio for the critical cluster size is unity; for larger particles,
the value falls below one, approaching S as g -»• °°.
The ratio a /g can also be expressed as follows using (17)
o
exp
UnS
(21a)
(21b)
2/3 s[(gVg)1/3-i]
(21c)
where the critical particle diameter is given by
4crv
A* - L
p KT£nS
For the case of the dimer, the values of a_/6 can be calculated from (21c)
with the following values assumed:
S = 10
then
g* = 8
rl,2/3 In0.59
LTT) 10 =2.45
25
-------
This is to be compared with the ratio of (18):
n ,s n
g-1 g-1 =
"
V.
As reasonable values for the Los Angeles smog aerosol, we take A = 2 x 10~
cm~land n = 1()8 cm~^, and a molecular diameter = 5 x 10"^ cm.
Then
wS x W
= 12.34
For this set of assumed values, the evaporative term will be small compared
with the condensation term.
These calculations indicate that when the surface area of the pre-
existing aerosol is large and the rate of formation of condensable species
is low, clusters are scavenged by the aerosol at a rapid rate. The slope of
the steady-state cluster distribution with respect to particle size is much
steeper than the slope of the equilibrium distribution. As a result, the
evaporative term in the expression for the particle current can be neglected
and a term for the "nucleation barrier" does not appear in the expression for
the particle current, (20). The rate of new particle formation is controlled
by the pre-existing aerosol. The particle current is much smaller than in the
absence of foreign particles; most of the condensable species deposits on the
aerosol and little condensing mass grows through the 10 to 100 A size range to
appear as new particles.
McMurry (1977) has studied the effect of an aerosol on the rate of new
particle formation in the S02-NOx-propylene system. Experiments were per-
formed in a teflon balloon reaction exposed to solar radiation. The concentra-
tion of the rate of formation of particles larger than 100 A with electrical
aerosol analyzer. He found approximate agreement between his experimental
results and a nucleation barrier-free theoretical calculation, as expected
from the above considerations.
26
-------
SECTION 7
DISTRIBUTION OF CHEMICAL SPECIES
WITH RESPECT TO PARTICLE SIZE
DISTRIBUTION OF SULFUR WITH RESPECT TO PARTICLE SIZE
The development of a low pressure impactor capable of fractionating
aerosol particles larger than about 200 A has been described by Hering,
Flagan and Friedlander (1977). The impactor has eight single jet stages
corresponding to a particle collected with an efficiency of 50% are 4.0,
2.0, 1.0, 0.5, 0.26, 0.11, 0.076, and 0.05 ym. The cutoff diameter for the
last two stages are approximate, based in part on calculation and in part
on an approximate calibration.
Particles larger than 0.5 urn are sampled at near atmospheric pressure
using the first four stages of the impactor. The last four stages operate
at pressures from 150 mm down to 8 mm Hg (absolute) to size segregate the
smaller particles. Material deposited at each stage can be analyzed for
sulfur containing compounds at the nanogram level by flash volatilization
followed by flame photometric analysis (Roberts and Friedlander, 1976).
Enough material for chemical analysis can be collected by sampling ambient
air with sulfate levels of 10 yg/m-^ for 15 minutes. The instrument has been
checked with ammonium sulfate and sulfuric acid aerosols by comparison with
an electrical mobility analyzer. Good agreement between the two instruments
was found indicating that the evaporation of these compounds in the impactor
was not important. A further check was carried out by sampling simultaneously,
in parallel an ammonium sulfate aerosol, with the impactor and a filter.
Excellent agreement was obtained.
Measurements have been made of the sulfur mass distribution with respect
to particle size with the low pressure impactor. In the absence of strong
photochemical conditions, the sulfur mass distribution usually peaks in the
size range around 0.5 ym (Figure 6). This is the size range which corresponds
to the wave length of the visible light and which is strongly light scatter-
ing. The sulfur mass distribution has also been measured with the low pressure
impactor during chamber experiments; small quantities of S02, propylene and
NO-N02 were added to unfiltered ambient air exposed to solar radiation. The
sulfur mass distribution peaked in the range between 0.1 and 0.2 ym,
(Figure 7), significantly smaller than the peak range for the ambient aerosol.
(Significant quantities of sulfur in the 0.1 to 0.2 ym range were also
measured in one measurement made on a day of strong photochemical activity.)
27
-------
N)
00
•n
H.
fD
4 __T
2 S 2. If •**
3 4 rr CT t
?* » ^ J
14
12
^
5 10
P "d O 3 O*
CL O O O 2
rt i- Hi ^
4 H' t/l
h- • 3 C ee\
i-> (O
i £££ co
r-, » P H 3 rt ~S
H. X-P 3- ^
OQ 0 4 ^'
<• o n o
O P t/l
v-1 4 H- (6
ps W MI f)
OQ M. O rt
P 5 4
3 3 3 rt
rt H- O
§3* P
I-- t- fp
W O C^ W
3 H-
&. CO f. N
(P M- M- CO
4 N rt
- (S> y H>
VD p
~J 3 >-« rt
•«JOS) O 3*
v-/ (B S (1)
L
r^ Q
o.
0> ft
J2
^^
<
^*
4
2
0
i
—
^^
—
9/13/76, 5«20-6'40 PM PDT
_Keck Lab., Pasadena, Ca.
™" V
Total Loading: 9.2
—
i
—
••••I^^^H
n^
i
• i
fJLq/m9
1 ^
1
••••••••J
_
^_
—
—
••^••••i
i
^^ ^^^j^^^
0.01 0.10 1.0 1C
PARTICLE DIAMETER, dp(/*m)
-------
CO
1000
800
eoo
• ~^_
~ 400
a.
•o
o 200
Comparative Size Distributions
PI6 #2
—EAA »2<36-l2'56
VT=353/Lim/cm3
—LPI I2'42-I2'44
0.01
O.I 1.0
Particle Diameter, dp(/xm)
10
Figure 7. The distribution of sulfur with respect to particle size for a balloon
chamber aerosol, measured with the low pressure impactor. SC>2 propylene
and NO-N02 were added to unfiltered ambient air and the mixture irradi-
ated. A peak in the sulfur distribution developed in the particle size
range between 0.1 and 0.2 urn (Hering, Flagan and Friedlander, 1977).
-------
AEROSOL GROWTH DYNAMICS
We consider the case of an aerosol uniformly distributed in a large
chamber with particles growing as a result of gas-to-particle conversion.
Coagulation and deposition on the walls of the chamber are neglected. This
analysis holds best for particles larger than about 0.1 ym in diameter, since
particle concentrations in this size range are relatively low and coagulation
rates are small.
It is convenient to introduce the continuous size distribution function
n (d ,t) defined by the relation
dN = ni(dp,t)d(dp)
(22)
where dN represents the concentration of particles in the size range between
dp and dp+d(d_) at time t. The change in n..(d ,t) with time in the chamber
is given by tne exression "
!!i
3t
3n,
3d
= 0
(23)
The distribution of the total aerosol volume with respect to particle
diameter is related to the particle size distribution by the expression
3V
log d
= 6 = £ An 10 d n.
6 pi
(24)
The area under the curve of y plotted as a function of logd is proportional
to the total particle volume. As a result of gas-to-particfe conversion, the
total aerosol volume increases with time. Each point on a curve of w vs_ logd
shifts to a new value determined by the growth law. The rate at which any ™
point on the volume distribution function changes with time is given by
£ •
-
where the particle growth rate, w, depends on the mechanism of gas-to-particle
conversion (Table 5) . By evaluating the right-hand-side of this expression,
it is possible to determine the particle size at which the value of w grows
most rapidly.
Substitution of (24) in (25) gives
3n
-rs
3d
i s
+ 4 w nd (26)
l
I
J
30
-------
If the particles are much smaller than the mean free path of the gas, and
they grow by collision with monomer their volumetric growth rate is propor-
tional to their surface area. The rate of increase in particle diameter with
time is constant in this case:
, ~ VrW .**.*? l*tt*i V. \£, I J
at
When w is constant, substitution in (23) gives:
3n 3n
IT + "*r = ° (28)
p
Substitution in (26) the result is
& = ~ £n 10 w d3 nl (29)
But the distribution of aerosol surface area with respect to particle size is
given by
8A » irjln 10 d3 n. (30)
aiogd u p "i
So
dt 331ogd
Thus the volume distribution tends to grow most rapidly at the particle size
Si A
corresponding to a peak in the area distribution function -ys — -r- • This is
observed, approximately, in chamber experiments with SC^, propylene and NO-N02
in air containing ambient aerosol particles (Figure 7J.
DEVIATION BETWEEN THEORY AND EXPERIMENT: SULFATES
Why then does the sulfur mass distribution of the ambient aerosol usually
peak at a particle size much larger than the diameter corresponding to the
peak in the area distribution function? Although we do not know the answer to
this question, there are several possible explanations: 1) The data of Whitby,
Husar and Liu (1972) indicate that coagulation can play a significant role in
reducing the number concentration of the Los Angeles aerosol between 2000 and
0400 hours (data of September 3, 1969). Over this time period the concentra-
tion dropped by a factor of almost 10 while the volume concentration of the
aerosol remained almost constant. Thus the mean particle volume increased by
a factor of about 2.15 over that time period. Part of this material remains
in the air shed and contributes to the sulfate present in the larger particle
sizes. The measurements were made at a fixed point in Pasadena so the effect
of advection was not taken into account.
31
-------
2) Certain condensable organic compounds generated by gas phase reac-
tions accumulate preferentially in the particle size range around 0.5 ym.
This was the case with two cyclic olefins, cyclopentene and cyclohexene, and
one diolefin, 1,7 octadiene, which served as aerosol precursors in experi-
ments reported by Heisler and Friedlander (1977). The experiments were-
carried out in a baloon reactor; the hydrocarbons and oxides of nitrogen were
added to unfiltered ambient air in the chamber and exposed to solar radiation.
Difunctional organic compounds produced in this way condense on existing
aerosol particles. The rates of particle growth could.be correlated by a
mechanism based on diffusion to the particles of the condensable gases with
a critical particle size near 0.2 ym below which growth did not occur (Kelvin
effect). This leads to the development of a peak in the volume distribution
in the particle size range, dp > 0.5 m (Heisler and Friedlander, 1977). The
area distribution would develop in a size range between 0.2 and 0.5 ym.
Simultaneous experiments with SOo were not conducted. If, however,
sulfuric acid were formed in such a mixture by the oxidation of SC>2, the
^804 monomer and small clusters might then be scavenged by the developing
peak in the area distribution at a larger particle size than observed in
chamber experiments without organic aerosols.
3) If the oxidation of S02 takes place in a droplet phase such as a
dCa)^
morning fog, the droplet growth rate —-TT*- ad (Friedlander, 1977) . In this
case it can be shown by modifying the analysis of the previous section that
the sulfur peak would tend to develop in the size range corresponding to the
peak in the volume distribution of the fog. When the fog partially evaporates,
the sulfate would be found in a larger particle size range.
DISTRIBUTION OF NITRATE WITH RESPECT TO PARTICLE SIZE (Moskowitz, 1977)
Measurements of aerosol nitrate were made at three locations in the
South Coast air basin; Pasadena, Hermosa Beach (30 miles SW of Pasadena, and
Chino (an agricultural area 30 miles SE of Pasadena) with the 8 stage, low
pressure impactor. Six samples were obtained over the period 7 AM to 7 PM
with each sample 1-2 hours in duration. The impactor stages were analyzed
using a chemiluminescent NO analyzer the same day.
A.
The data for the three locations for each run are shown in Figure 8. It
is seen that in midmoming, regardless of location, there is a significant
quantity of particulate matter in the submicron range. This is also apparent
in the late afternoon data of Chino and Pasadena.
An average diurnal nitrate size distribution for each location was
obtained by time averaging the nitrate levels present on like stages of the
impactor. The Pasadena distribution was bimodal with peaks in the two size
ranges 0.05-1 ym and 2-8 ym (8 ym is an arbitrary upper cutoff). The Chino
data, near an ammonia rich cattle feed area, showed a peak in the submicron
range although the distribution was weakly bimodal. In the coastal area the
predominant size range for nitrates was in the 2-8 ym region. Pasadena may
be throught of as possessing aerosol characteristics of both coastal and
agricultural regions.
32
-------
HERMOSA BEACH
10/4/76
PASADENA
10/5/76
CHINO
10/8/76
6:42-8:30
6:58-8:47
7:O5-8:3O
. rrrfl-rn
I I TT i r i i
I I I
8:43-10:31
9:15-10:29
8:42-10:12
I i i i
10:43-12:30
II I I i i I
10:37-12:12
II i I I i I
10:37-12:00
I if i i i i i
12:41-14:47
i i i i i i
(2:25-13:45
11 i i i i i i
!2:ia-l4:OO
r i i t i
I5:OO-I6:3O
it i i i i
15:14-16: IO
i * i t i
14:12-16:01
i ii i i i i i
16:42-18:28
ii i i i i i
I7:I5-J9:OI
i i r r i i i i i i i i i i » i
.02 .05 .11 .25.50 I 2 4 8 .02 .05.11 .25.50 I 2 4 t .02 .05 .11 .25.50 I 2 48
dt
Figure 8. Aerosol nitrate distributions with respect to particle
size for'different locations at various times of day.
33
-------
In an investigation by O'Brien et al. (1975), inorganic nitrate and other
particulate pollutants were measured in a number of California locations. It
was found that inorganic nitrate (NH4N03 and NaNQs) of a secondary or photo-
chemical origin contributed about 10% of the mass loading in the South Coast
air basin. The prevalent form of nitrate in the basin was NlfyNOs; the NHj
concentration was the highest in the Los Angeles air basin compared to other
California locations. The study also showed that the coastal Santa Barbara
sample contained 16% inorganic nitrate while the NH^ concentration was
essentially zero. This indicates that the reaction of N02 with NaCl is of
importance in coastal areas:
3N02 + H20 = 2HN03 + NO
HNO + NaCl = NaNO_ + HC1
«J O
The data indicate that the Pasadena aerosol consists of both small NH.NO- and
larger NaN03 particles.
Nitrogen (as nitrate) gas-particle distribution factors were calculated
.on
NO:
for the Pasadena data. The distribution factor f.T is defined as:
N
f = r
N N0x + NO"
where NOg = particulate nitrate concentration, as NOo, yg/m and NOX = gas
phase concentration as N02» converted from ppm to yg/m3, averaged over the
sampling period. The distribution factors ranged from 2.5 to 22.8% in the
same range reported by Grosjean and Friedlander (1975).
It is of interest to look at the amounts of NaNOj which would correspond
to the ambient levels of Na. Sodium was measured in the ACHEX experiment
(Hidy, et al. 1975) and the concentration was about 0.7 yg/m3.
This corresponds to a level of 1.89 yg/m3 of NOij. If one time averages the
levels of particles with diameter greater than 2.0 ym for each of the loca-
tions it is found that beach nitrate is 2.6 yg/m3, Pasadena 2.5 yg/m3, and
Chino 1.0 yg/m3. This would be consistent with the hypothesis that the con-
centration of Na should decrease as one moves inland, with decreasing levels
of large particles. The concentration of large particles in Pasadena is
consistent with the concentration predicted by ambient Na levels.
The results of the ACHEX study indicated that on a mass basis, the
nitrate component of the aerosol contributed less to light scattering than
the sulfate. The explanation ftfr this effect may be that most of the sulfate
exists in a highly efficient light scattering range (0.1-1.0 ym) whereas only
part of the nitrate is found in this range.
34
-------
REFERENCES
1. Calvert, J. G. (1973) "Interactions of Air Pollutants" in Proceedings of
the Conference on Health Effects of Air pollutants, U. S. Govt. Printing
Office Serial No. 93-15.
2. Frenkel, J. (1955) Kinetic Theory of Liquids, Dover.
3. Friedlander, S. K. (1977) Smoke, Dust and Haze: Fundamentals of Aerosol
Behavior, Wiley-Interscience, New York.
4. Friedlander, S.K. (1978) Atmos. Environ., 12, 187.
5. Fuchs, N. A. and Sutugin, A. G. (1971) "Highly-Dispersed Aerosols" in
Hidy, G. M. and Brock, J. R. (Eds.) Topics in Current Aerosol Research,
Pergamon Press.
6. Grosjean, D. and Friedlander, S. K. (1975) JAPCA 25, 1038.
7. Grosjean, D. and Friedlander, S. K. (1978) "Formation of Organic Aerosols
from Cyclic Olefins and Diolefins" based on work done on EPA Grant No.
R 802160. Submitted for publication.
8. Heisler, S. L. and Friedlander, S, K. *1977) Atmos. Environ., 11, 185.
9. Heist, R. H. and Reiss, H. (1971) J. Chem. Phys. 59, 665.
10. Hering, S. V., Flagan, R. C. and Friedlander, S. K. (1978) Environ. Sci.
Techno1. 12, 667.
11. Hidy, G. M. (Ed.) (1975) "Characterization of Aerosols in California"
(ACHEX). Final Report to Air Resources Board, State of California, ARB
Contract No. 358.
12. Hidy, G. M. and Friedlander, S. K. (1971) "The Nature of the Los Angeles
Aerosol" in Proceedings, 2nd Intl. Clean Air Congress, Englund, H. M. §
Berry, W. T. (Eds.) Academic Press, New York.
13. Katz, J. L., Scoppa, C. J., Kumar, N. G. and Mirabel, P. (1975), J. Chem.
Phys. 62^, 448.
14. Kiang, C. S. and Stauffer, D. (1973) Faraday Symposia of the Chemical
Society, No. 7, 29.
15. Mason, B. J. (1971) The Physics of Clouds, Clarendon Press, Oxford.
35
-------
16. McMurry, P. H., On the Relationship Between Aerosol Dynamics and the
Rate of Gas-to-Particle Conversion. Ph.D. Dissertation in Environmental
Engineering Science, California Institute of Technology. (1977).
17. Moskowitz, A. H. "Particle Size Distribution of Nitrate Aerosols in the
Los Angeles Air Basin," EPA-600/3-77-053, May 1977.
18. O'Brien, R. J., Crabtree, J. H., Holmes, J. R., Huggan, M. C. and
Bockian, A. H. (1975) Environ. Sci. and Technol. 9^ 577.
19. Reiss, H. (1950) J. Chem. Phys. 18^, 840.
20. Roberts, P. T. and Friedlander, S. K. (1975) Environmental Health
Perspectives 10, 103.
21. Roberts, P. T. and Friedlander, S. K. (1976) Environ. Sci. Technol. 10,
573.
22. Sander, S. P. and Seinfeld, J. H. (1976) Environ. Sci. Technol. 10^ 1114.
23. Whitby, K. T., Husar, R. B. and Liu, B. Y. H. (1972), J. Colloid
Interface Sci. 39, 177.
36
-------
APPENDIX A
PROJECT DISSERTATIONS AND PUBLISHED PAPERS
THESES
Ph.D.
Roberts, P. T.
Heisler, S. L.
McMurry, P. H.
MS
Moskowitz, A.
"Gas-to-Particle Conversion: Sulfur Dioxide in a Photo-
chemical ly Reactive System" (1975)
"Gas-to-Particle Conversion in Photochemical Smog: Growth
Laws and Mechanisms for Organics" (1975)
"On the Relationship Between Aerosol Dynamics on the Rate
of Gas-to-Particle Conversion" (1977)
"The Distribution of Aerosol Nitrate Compounds with
Respect to Particle Size: Vaporization Analysis with
the Low Pressure Impactor" (1977)
PUBLICATIONS
Friedlander, S. K. (1978) Atmos. Environ., 12, 187.
Grosjean, D. and Friedlander, S. K. (1975) JAPCA 25, 1038.
Grosjean, D. and Friedlander, S. K. (1978) "Formation of Organic Aerosols
from Cyclic Olefins and Diolefins" based on work done on EPA Grant No.
R802160. Submitted for publication.
Heisler, S. L. and Friedlander, S. K. (1977) Atmos. Environ. 215.
Bering, S. V., Flagan, R. C. and Friedlander, S. K. (1978) Environ. Sci.
Technol. 12, 667.
Moskowitz, A. H, "Particle Size Distribution of Nitrate Aerosols in the
Los Angeles Air Basin," EPA-600/3-77-053, May 1977.
Roberts, P. T. and Friedlander, S. K. (1975) Environmental Health Per-
spectives 10, 103.
Roberts, P. T. and Friedlander, S. K. (1976) Environ. Sci. Technol. 10,
573.
37
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
\. REPORT NO.
EPA-600/3- 79-052
2.
4. TITLE AND SUBTITLE
PHOTOCHEMICAL AEROSOL DYNAMICS
7. AUTHOR(S)
S.K. Friedlander
3. RECIPIENT'S ACCESSION>NO.
5. REPORT DATE
May 1979
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Chemical Engineering
California Institute of Technology
Pasadena, CA 91125
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27
10. PROGRAM ELEMENT NO.
1AA603A AC-16 (FY-78)
11. CONTRACT/GRANT NO.
R802160
13. TYPE OF REPORT ,AND PERIOD COVERED
- RTP, NC Final 4/73-3/78
14. SPONSORING AGENCY CODE
711 EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
New data are reported on (1) the rate of formation of condensable chemical
species by photochemical reactions, (2) the effect of the reaction products on the
particle size distribution and (3) the distribution of reaction products as a
function of particle size. Gas-to-particle conversion for cyclopentene, cyclohexene
and 1 , 7-octadiene , ranged from 5 to 39 percent of the initial gas-phase carbon con-
centrations. Size distribution data for cyclohexene were correlated by a diffusion
controlled growth law with a Kelvin cutoff diameter at about 0.25 urn.
In polluted atmospheres, some new particle formation takes place as a result
of homogeneous gas phase reactions even though an aerosol is already present. To
explain the results of laboratory studies of this phenomenon, classical nucleation
theory must be modified to take into account the scavenging of clusters by the
aerosol. Using a new low pressure impactor, the first measurements have been made
of the distributions of sulfate and nitrate with respect to particle size for
d_ < 0.25 um. In Pasadena, the data for sulfate often show a peak in the mass
distribution for 0.6 < dp < 1.0 ym; less often, a peak is observed near 0.1 um,
consistent with laboratory data for aerosols formed by homogeneous gas phase
reactions.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air pollution
*Aerosols
*Photochemical reactions
*Particle size distribution
18, DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b.lDENTIFIERS/OPEN ENDED TERMS
19. SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
c. COSATI Field/Group
13B
07D
07E
21. NO. OF PAGES
46
22. PRICE
EPA Form 2220-1 (9-73)
38
------- |